Document vVGpdg5dLg9bdVozj47ZE54bb

American Journal of Industrial Medicine 5239-249 (1984) v An Evaluation of the Associations of Leukemia and Rubber Industry Solvent Exposures . Harvey Checkoway, mol Timothy Wilcosky, Pho, Pamela Wolf, wH, and Herman Tyroler, MO Excessive leukemia m o d i t y has appeared codstentiy in epidemiological srudics of British and U.S. rubber industry workers. Attempts to identify causative factors have f- on exposure to benzene and other soIvuw. Interpretations of findings from these studies have often been influenced by expectations of a benzcnelnonlymphocytic leukemia association, Seen from previous wort in other settings. However. data from the rubberindustry studies have nos becn consistentwith this expectation. as lymphocytic and nonlymphocytic leukemia have shown similar mortality excuse. Diua from a small casecontrol study of Iympbocytic leukemia am presuucd to illusu;ut an approach tha consiJers multiple solvent exposures. The assock&nr with lymphocytic leukemia risk observed for a number of solvems, most rocably carboa tetrachloride and carbon disulfide, were stronger than those dctectd for benzene, INTRODUCTION Background Leukemia mortaiity excesses have been prominent among the findings from numerous epidemiological srudies conducted in the British and U.S. rubber industries. Recognition of the extensive use of organic solvents in rubber and tin manufacturing stimulated a priori suspicions that there might exist lcukcmiia hazards in the industry. The welldocumented causal association between benzene exposures and leukemia Seen in the rotogravurc and shoe industries [ v i g i i i and %ita, 1%; Ahoy and Erdem. 19781served as a guide in this regard The relationship between leukemia risk and solvent utposures in the rubber industry has k e n . a! most, indirtctly inferred. The main reason has been that these investigations were retrospective designs-a common, often necessary feature of epidemiological studies of m e , chronic diseases-with either nonexistent or . - unsystematically obtained environmental exposuredata. Nonetheless, the consistently Occupational Health Studies Gmup. University of Noah Cardiq chppd Hill, NC(H.C.. T.W.. H.T.). Division of Biostatisticsand Epidemiology. Georgetown UNvcnicy School of Medicine. Washington. Dc (P.W.). Address reprint rcqucsu IO Harvey Checkoway. University of Nonh Camlina. Occupa1bm1Health Studies Gmup. NCNB Plaza. Suite 32. Chapki Hili. NC 27514. Accupted for publication Junc 14. 1983. 0 1984 Alan R. Li,Inc. 2jn Chcckoway observed patterns of elevated leukemia mortality rates, accompanied by apparent localization of risks to workers with the greatest potential for past solvent exposures. prompted the International Agency for Research on Cancer to conclude in a recent monograph on the rubber industry [IARC, 1982aj that therc is sufficient evidcncc to consider the association as causal. In this paper we review the studies of leukemia and solvent exposures in the British and U.S. rubber industirs. particular attention is paid to the specificity of this relationship, in terms of leukemia cell type. and with regard to the various solvents used in the industry. Much of the interpretation of the findings from previous research has been influenced by how well or poorly the results accord with expectations derived from studies of benzeneexposed workers in other occupational settings. In particular, the presence or absence of a bemendacute nonlymphocytic leukemia association has been the standard for causality, despite some conflicting evidence. The implications of this approach are discwed. In addition, we present findings from a small-scale casecontrol study that are suggestiveof etioiogical associations between lymphocytic leukemia and a number of rubber industry solvents other than benzene. most notably carbon tetrachloride and urbon disulfide. A detailed summary of the findings from the British and U.S. studies of leukemia among rubber workers is given in Table i. Selected important findings and conciusions drawn from these investigations are discussed below. British Rubber Industry Studies - Two companion series of epidemiologic studies have been undertaken by the Employment M e d i d Advisory Service (Fox et al. 1974; Fox and Collier, 1976; Baxter and Werner, 19801 and the British Rubber Manufacturers' Association [veys. 1981; Parkes et aI, 19821. Modest excesses of leukemia have beenobserved generally. The most direct evidence pertaining to solvent exposures was a marked excess of acute myeloid leukemia in one tire plant (SMR=714) when plastic film manufacnrr- ing, entailing benzene exposure, had occurred in past years [Baxter and Werner, . 19801. In contrast, Parkes et al(19821 discounted the likelihood of a causal association between leukemia mortality and solvents because the observed excesses occurred in years when benzene was mi longer used, and therc was M) unusual leukemia cell type distribution. I U.S. Rubber Industry Studies Epidemiologic research in the U.S. rubber industry has revealed faidy consis- tent patterns of elevated I-dity rates (see Table I). While there has been considtrabl>anability the reported relative risks, most have k n in the nnge of 1.5 to 3.0,and lymphocytic leukemia excuses generally have been as prominent as those from the noniymphocytic leukemias. Cell tvpe distinctions were not made in many of the studies. howcver. Etiologicrrl ;Lsmciation. with solvents havc hccn cxamind by .some invc5tigators who uscd work history infomiation to indicate the likeiihoori for past exposure. Sevenl of these studies wamnt comment. Researchers from the National Institute for Occupatiod Safety and Health reported a Standardized Mortality Ratio (SMR) of 560 for leukemia among 748 workers engaged in rubber hydrochloride (Pliofilrn) manufacturing in the 1940s 4 [ C e F 1, k 'c\ c e. D e: fr tf ifl [I th re la (1 hi in. an to co W( en co &. or( lec COI en1 SFK the in\ PE his thi? cot onl risk WO : -0t :nt 2s. ant to he his !Its i.ch Ins In .iia ...e. Ire of :nd of ,nd ..vS Ily. 0f ur;er, ion I in .'Pe iis:cn : of as : in 0rs ire. 11th 148 40s e Leukemiaand Rubber Industry Solvents 241 * (Infante et al, 1977; Rinsky et al, 19811. Benzene exposures were known to have occurred in the Pliofilm departments studied. An argument for a potent leukemogenic effect of benzene was offered on the basis of the estimated low exposure levels, presuqd to be in compliance with existing standards,and becauseall seven observed leukemia deaths were either myelogenous or monocytic [Rinsky et al, 1981). Andjelkovich et al [1976, 19771 previously had found no leukemia excess among Pliofilm workers. McMichael et al][19741 r (SmTQtor workers aged 40-64 in o wt. Leukemia risk was related to work experience in a number ot Jobs where there existed potential for solvent exposures [McMichael et ai, 1976bl. An apparent dose-effect relationship W e e n solvent exposures and lymphocytic leukemia was detected in a case-conuol study of 15 deaths from several plants [McMichael et al, 19751. Specific solvents were not identified in these studies, as the exposures were considered to be mixtures of various aromatic and aliphatic compounds. In a larger cascconml study of Ieakemia death in four companies, Wolf et al [I9811 confirmed the association of lymphocytic leukemia with solvent exposures in the company reported on earlier by McMichael et a1 [19751 but did not find similar results for the other three companies. Also, myeloid leukemia was apparently unrelated to solvent exposures. In the casecontrol studies &ported by McMichacl et al[1!375] and Wolf CI al (1981], solvent exposures were inferred from occupational titles recorded on work history records, where the occupational titles were functionally homogeneous groupings of jobs, similar with respect to materials handled and machinery used [Gamble and Spirtas, 19761. Occupational titles were useful for distinguishing jobs according to generalired solvent exposure potential; however, cxposures to individual solvents could not be specified. Consequently, Arp et al I19831 devised a method to estimate worker exposures to the specific solvents encountered in the company that experienced the greatest lymphocytic leukemia excess [McMichacl et al, 19741. From company documents dating back to the 19209, Arp created historically validated charts of annual authorirod solvent usage by process area for all solvents used in order to study the association between solvents, especially benzene, and lymphocytic leukemia. The cascs and convols in Arp's sNdy were the same subjccts from this company that were included in the larger case-contml study wolf et al, 19811. The entries from work history cccords and the solvent charts indicated whether workers spent time in a process area when a solvent was authorized for use. Descriptions of the job titles listed on the work histories permitted differentiation of the jobs that involved routine handliig of solvents (priinary exposure) from other jobs in the process which pnsumably involved lesser exposure to solvents (scamdary exposure). Table II shows data from Arp's study [Arp et al, 1983) ofthe solvent exposure . history comparisons of 15 I>mphocytic leukemia cases with 30 matched controls. In this analysis, workers with potential solvent exposum of at least I year were considered exposed. Although workers with secondary solvent exposures experienced only a slight excess of lymphocytic leukemia, the data indicated a 4.5 times greater risk among workers who routinely handled solvents (primary exposure). It is note- wonhy that the cxposure odds ratio for the benzene categories and the aggregate 'other solvent" categories were almost identical. . I . 242 Checkoway -I I u' zVL pE 3 E f' 5 e a -Y z" c t dt= i cu iz -. . L - u> 0 '0 ic a - Leukemia and Rubber Industry Solvents 243 UUaUn ..u-u 1: Y oo cy Y cz 5 s g.- u sB .Y u P e- e- 244 Checkoway n n II n P n n n cec~ce~~cr~~a 00000000 4 UH Uu uV 0L L m u e :: 7 L c0 4 Leukemia and Rubber Industry Solvents 245 TABLE 11. Solvent Exposure and Lymphocytic Leukemia Relative Rslc F-\t i t n ~ i d Exposure utcgory Number of cases cxpoxd OJJS d0 Chi square Primary benzene k o n d a r y txnzcne Primary othrr solvent Secondary other solvent *Data from Arp d al[1983]. - 3 2 I1 8 4.5 I .M I .5 0.13 4.5 3.M 1.6 0.4 I N E W RESULTS FROM A STUDY BASED ON MULTIPLE SOLVENT EXPOSURES Methods and Results The casecontrol analysis of lymphocytic leukemia reported by Arp et al (19831 was extended to consider risks associated with 24 specific solvents. As before, solvents exposure chans for process areas were reconsvucted from historid plant records. The determination of workers' exposure histories differed from the method used previously in that exposure codes for each solvent were assigned to the department rather than to the individual job title entries on the work history records. This simplified method of exposure coding enabled the use of a computer algorithm, and it proved to be more cost- and time-eficient for examining risks related to multiple agents than did the manual coding process used in the previous analysis, which was restricted to an evaluation of only two classcs of txposure, benzcne and ail other solvents grouped together. The only information sacrificed was rhe judgmental dis- tinction between primary and secondary exposures. The cases in the present analysis were 11 male, hourly workers whose underly- ing cause of death was lymphocytic leukemia(ICD8th Revision Code 204) who were identified from the cohon reported on by McMichael tt al[1975]. These cases were also incfuded in the studies by WoIf et aI (19811 and Arp et al(1983). Controls were a 20% age-stratified random sample of the cohon [ M c M i c i d et al, 1976b). A total of 1,350 controls (1.280 whites and 70 blacks) was included. Age-adjusted relative risk estimates, as approximated by the odds ratio, and associated with chi-square tests of statistical significancewere computed according to ,proctdures described by Mantel and Haenszd (19591. Worlrcrs were considered exposed to a particular solvent if their cumulative duration of work in a solventrelated department e x d e d 12 months. Table III shows the odds ratios associated with each of the 24 solvcnrs. As mentioned above, no distinction was made between primary and secondary exposun to benzene, which accounts for the lower rdative risk Cstimate; however, the odds ratio (2.5) is within the range of the estimates f d in the pmious analysis (see m,Table n). Among the solvents listed in Table carbon terrachloride(OR=14.8) and cubon disulfide (OR3.8.7) showed the strongest aSSOCiations with lymphocytic 1kcmia. Several other solvents w e n also related to leukrrni mortalii, and these d u p l e aSSOCiations probably reflect simultaneous utposurcs to diffmat solvents used cotacumntly in the industry. Neither carbon tetrachloride nor carbon disulfide has been shown previously to be leukemogenic, and it is possible chat the present furdings arc spurious. Nevertheless, the high odds ratios forthese solventssuggest that, even if they an not etiological ..I -....D-.3 .=? *y3 ,:,;ZJ %hi .:aI*i ca-4 -.--j ?a, *: I-*-, ''-&.."1< I$; ,In"*::=a&' i- )* :it , il t 1!ii . ]! I I. 'I I/ f I 1,I* I x I, .. .!;*. ..! .x i 246 Checkoway TABLE 111. Casecontrol Analysis of Lymphocytic Leukemia Risk Associated With Exposures to 24 solvcnts Accconc . Aqua ammonia BemCnC -carbondisulfide Carbon tetrachloride Dipcfim Ethanol Ethyl Gasoline 3 5. I 1 2.4 4 2.5 ,7 8.7 8 14.8 I I .o 4 2.0 3 5.1 9 4.9 !-. HCXm ~Propa~l MUhaJlOl Medrylcne chloride M i d spirits pefihlomcthylery Phenol Solvent A' special m p h h Toluene Trkhloroatwe Trichlomtthylene VMandPrr;rphthas Xylmcs 2 0.9 7 3.8 6 1.8 -1 8.3 0 1 I.5 I 0.7 1 0.5 7 27 8 2.7 -2 3.0 0 2 0.8 3 2.9 4 3.2 .Ropriaay mixfunof tducrre and other solvent$. 'p < 0.05. "p < 0.001. 6.x'* 0.7.1 I .59 12.8218.000.00 1: 14 6.33. 3.10 .. 0.0 1 4-32. 0.84 5 3I 0.32 0.15 0.09 0.40 2.14 I .62 1.81 0.52 0.07 2.68 3.35 agents themseives, they arc probably closely associated with the actual causative txposurt(s). The exploratory nature of this analysis and the small sample size preclude firm conclusions about specific solvent-leukemia associations. However, these findings illustrate c l d y the danger of focusing exclusively on a single exposure, ie, benzene, in a dticxposurc mimnment, DISCUSS1ON The examinations of leukemia risk factors in the British and U.S. rubber industries have pnmeded Sequentially from the identification of mortality excesses, nlative to the generally unexposed national andngionai populations, to more inten- 2-.- I-----I---.- e --y related work environments and exposures within the tndustry. Leukemia mortality excesses of'varying magnitudes have been observed in most studies. The use of soivents in various proctsscs of rubber and tin manufactur- ing (eg. tire building) and the well-recognized leukemogenic potential of benzene suggcst an obvious etiological association: howcvcr, thc findings from m n y investi- gationsof mbbcr worker populations do not fit especially well with a bemcne-related cffaz. In this regard it is noteworthy that the animal data on WtperimentaI benzene car, 197 bee pol- [D;: plai ben of c I92 eve' acuI ther [De fror ha tors Pari mo? The evic reh exP leu L UnC* ana! wcr Secl hf0 be tecfI conindit the and iden a co to b. leuh m. thc mu1I necc f.t!ukrmi? aad Rubber Industry Solvents 247 carcinogenesis also have been conflicting [wad et al, 1975; h4altoni and Scarnato, 1979; Snyder et al. 1980; IARC. 1982b3, and the reported mutagenicity results have been largely negative [Dean,1978; IARC, 1982b]. Contradictory results in human populations are therefore not surprising. Benzene was used in the industry as a g e n d purpose solvent in the 1920s [Davis, 19291, but following repons of toxicity [Hunter, 19391, benzene was sup- planted by toluene and other aromatic and aliphatic solvents (Mellan, 19571. W e benzene exposure still occurs in the industry, it is present primarily as a concaminant of other substances Wan Ert et al, 19801, and ~posurcsarc much lower than in the 1920s; thus it is quite Iikeiy that onty mail proportions ofthe cohorts studied were ever exposed to benzene in concentrations knownto induce leukemia. _ T h e n g a t y s a l associations for benztne have k n wirh the acute myeloid types ot leukemia, -;- lcukcmii [Goldnein, 19771, yet are also case reports of lymphocytic leukemia as a -la ofbenzene poisoning [Delorc and Borgomano, 1929; Hunter, 1939; Vigliani, 1975). Excessive mortality from lymphocytic leukemia has been & greatas that fromthe nonlymphocyticvariants in a number of rubber industry cohorts, a tcsult that has prompted seven1 investiga- tors to discount an etiologic relationship with solvents [Monson and Fine, 1978; Parka d al, 1982). The NIOSH study of Pliofilm workers (Rinsky et ai, 19811 perhaps offers the most convincing demonsPation of a benzare/nanymphocutic lcukunia association. The importance placed on the fivefold mmaiity excess dctcctd in this snrdy is e x e x ? . Even in view of the magniadc of the reported risk for noniymphocyuc leukemiaassociated with benzene exposure, the specificity ofthe asxiation remains uncertain, as exposures to other rubber industry'solvents were nat evalutated in the analysis. Findings from a reanalysis of the case-controidata of Arp et al(l9831, that were presented in this paper, demonstrate the extent to which intcrprctations of seemingly causal relationships can change in light of newly considered expostm information. Additional research into the etiology of leukunia in the rubber industry should be guided by several considerations made apprvent from prcvbus work. It .should be recognized that the presumed c l i ofcausative agents, solvents, is a mixedgroup of compounds, of which betuene is gmbally a minor quamitative component. The indication therefore is to examine possible etiological rdatiotlshipswith othersolvents. As shown from the data presented in Tabie UI, muncroussolvents were used in the industry. Momver, usage has varied qualitativdy and quancitativeiy over time, and there has been conside&le temporal ovdap of atpoaurcs. consequenty, the identificationofthe spa5ficagcnt(s) responsible for leukemia exctsses is likely to be a complicated process requiring investigatorjudgcmcnt asto which solvents arc likely to be leukemogenic, based on previous experimentaland human observationaldata. While the foregoing discussion suggests the llftd to investigate the potential leukemogenic effects of solvenu other than benzene, we do not imply that continued -h into benzene toxicity in the rubber industry should be abandoned. Rather, the scope of the research should be expanded to include assessment of effects of multiple exposures, and the interpretations of the findings should not be evaluated necessarily with respcct to a benzene/nonlymphoqtic leukemia association. 248 Checkoway ACKNOWLEDGMENTS lhis work was supported by the United Rubber Workers Union, the Fircstone Tire and Rubber Company, the General Tire and Rubber Company, thc Goodyear 9 Tire and Rubber Company. and Uniroyal, Inc. Thc authors are gntcful to Ms. Pamela Hookcr for prepring this manuscript. REFERENCES Alrra)r M. Erdem S (1978): Follow-up study on the mortality and dcvefopment of ~ e u k ~in u44 puytopcnic pati& with chronic exposure io be-. Blood S2:28S-292. A n d j d k d D. Tadbe J. Symons M (1976): M o d i t y Urpaiarc of a cohort of Nbkr ~~~, 1964-73. J Ocarp Med 18387-394. Ancljcbvid~D. Tauibe J. Symons M. Williams T (1977): Momliry of rubber WO&K with rcfemcc to work cxpericncc. J Orrup Med 19397405. Arp EW. Jr.. Wolf PH. Checkoway H (1983): Lymphocytic leukemia and expauva LO benzene and ahasdvats in the rubkr indusuy. J Oceup Mcd25:59&602 Baxtcr RJ. Wana JB (1980): 'Morality in the Rubber Idustry." London: Her Majq's SluioaayOmCC Davis P (1929): Toxic substurns in the N W C ~industry, part 2: Bauol. The Rubber Age. 13:367-368. Dcrn BD (1978): Genaif t o x i c ~olf k~ n z ~ nt~o,l u e ~~. y l t m arnd phtnols. Mum Res 47~75-97. Ddat P. Borgonwo C (1929): Acute leukaania foilowing knzart poisoning: On the toxic origin of drcutc ieuhcmias and their &tion 10 serious anaunk (Fr.). J Med L p 9227-233. DdLdl E h i k C. Lcwis C. Monton RR (1981): Monnlity and cancer morbidity among workers in the fubbcrindusay. Am J I d Med2209-216. D d d E. Momoa RR (1981): Mortality among rubber wafrar III. cnusC-spsific momlity. 1% IQ18. J Omrp Med23:677-684. Ddtdl E. Momon RR (198t): Mortality among tubber workem. V. Racasing work&. J Ocarp Mcd 24539-54s. ( Fox AI,Collier PF (1976): A survey of ocarplhod ~ c cinf thc rubber and c a b l d n g industries: Amlysk of daths occurring in 1972-74. Br J I d Med33249-264. Fox a. Lindars DC. Owm R (1974): A survey of 0ccupatiOrUi cancer in the rubber a d cablaruking indusucs: Ruulu of five-year analysis. 1%7-71. Br J Ind Med 31:140-U1. GambkepIF.l.deSmpt.idrotsgyR. (1976): Job classificationmd J Occup Med 183994W. utilintion of complete vmrk histories in axu pac iod Gddncio BD (Im:Hamra6ricity in humam. In Lvkin S. Goldstein BD (ah):"8mztneToxiciw A Critial EvalurciOn." Washington: Hempshire Publizhi Corp.. pp 69-105. H m (1939): Chmnic utposu~IO knraK (anror). II. T k diniczl de.J Ind Hyg 21~331- 354. IARC (19821): 'IARC Momgraphr on the Eduation of corcinogcnic Risk of Chemicalsto Man. Vol 28. Thc Rubkr Indusrry." Lyon: IntcnraliOnsi Agency for R d on Curs. pp 183-230. [ARC (1982b): "URC Monogrpph, on the Evaluaion of cnrcinogenic Risk to Humurs. Vd 29. Somc rdusthi- and Dyatuffs." Lpn: Inarmioml Agency for Resarch on Cancer. pp 391-398. lahncc PF. R h k y RA. WIgOnr JK. Young RJ (IQn): ' in +uene workas. L;incec 276- 78. M.honic.sarnru,c(1979):Fi~uperiddemoaanclonofthc CJlrinogenic effects of kruaw: b#uspyr 011 Spng~c-lh- r;illby dd d i . Mcd L;w %3S2-351. M.aa&a TF (197s): E p i d a n i i i imtstigation of occopoliaarl cancers in the ~ b kinrdustry. In LerirrroaC (ed):T h e New Multinational Halth HPtudr" Gam:ICF. pp 80-136. . h h a m TF. Ciacco A. E l - A m AA (I%@: An approd to rhe Nbkr indusuy: A study buedon &pmnaml expericncc J Ocrvp Med 10213-232. Maad N. Himw.cI W (19S9): Statistical q a x s of h e analysis of data from mro*pcaive Sndi of * distw.J Nacl cvlccr Inst 223719-748. McMictwi AJ. Andjdkovich DA. Tyroler HA (1976a): Cancer modity among rubber workers: An epidcmiolOgi~d y . Ann NY A d sd 271: IU-137. McM McM McM I Mclla Mom Mom Mow ParLC Rinsk! Snydc Van E "eys( Viglim Vigliar Ward. Wolf I I le dr 44 rs. ICe .nd i'S 4. of in led s: '"g *d ly: i Idol Inc PP '6- IY: In A of 4n Leukemiaand Rubber industry Solvents 249 McMichael Al. S p i n s R. Gamble JF.Touscy PM (1976b): Momlity Jmong r u b b w~rlrers:Relacionships IO specific jobs. 1Occup Mcd 18: 178-185. McMiclycl AJ. Spinv R. Kuppcr LL (1974): An epidemiologic &iy of monality within a cohort of rubber workcn. 1964-72. J Occup Med 1 6 4 5 8 4 . McMichael AJ. Spirras R. Kupper LL. Gamble3F (1975): Soivemuposurr and l e u h i a among rubber woclcers: An cpdcmioIogic study. J Occup Mcd 113234-239. Mellan C(1957): " W b o o k of Solvenu. Vol 1. Pure Hydrocyboru," New Yo*. Reinhold. Momon RR. Fine U (1978): Cancer mortality and morbidity among ~ b bWaO ~ C K . J Natl Cancer IN1 61:1047-1053. Monson RR. N a h o KK (197Q): Monality among rubkr worktrs. 11. 0th employees. Am J Epidcmml 103:297-303. Momon RR. Nakam KiC (1976b): M o d i t y among rubber workers. I. White nuk unionanployets in Akron. Ohio. Am J Epidcmiol 103:284-2%. pprktl HG. Veys CA. WatcrholrreJAH. PC~CAK(1982): Canca monrlity in I&British rubber industry. Br J Ind Med 39209-220. R~NQRA. Young RI. Smith AB (1981): Lcukaru. lin baucneworlren. Am J Ind Med 2217-245. Say& CA. Goldstein ED. Scllakumcn AR. Bmmkrg 1, h k i n S. Amen RA (1980): The inhalation toxicology of benzene: Incidence of hemstopoietic acop(auns ud hematomxicity in AlWJ and C57BU6J mice. Toxic01 AppI phumacolW.323-331. Van En MD, Arp EW.Hams RL. Symons MI. W i l l i Th4 (1980): Worker uposuna to chemical agenu in IIUmanufaawe of rubber t k Solvent vapor d i e s . Am I d Hyg Assoc J 41:212219. Veys CA (1981): Bladder cancer in rubber workers: 7he story revimd and updated plvt Rubbu Rocest Appl I :207-212. Vigliani EC (1975): Bauene and kulrun.u In Levinson C (ed): Ihe New MulciMtional Health Hwrdt." Geneva: ICF. pp 1%-201. VigliUri EC. SaitlG (1%): Bmrmeanl Ic~kUnt.N Engl I Med 271~827-816. Wud JM. Weisburger JH. Yamamoco RS (1975): Long-tcnn effects of benzene in CS7-BU6N mice. Arch Envimn Health 3022-2% Wolf PH. AndjeIkovidr D. Smith A. T p l a H (1981): A CUeCOWCOl study of leukemia in the U.S. rubber indusuy. J Occup Mcd 23: 103-108. ' ;1'; ; iI!8Jr "iJ$ :, -! \l 54 '1 1 4*'; 31. *;l L,Ei p ?t