Document qkRkKd0Rkr1mMOnYyoGKOmvak

CHAPTER 1 C.4RCISOCENIClT~O' F BESZENE, TOLCENE AND NI'LENE: EPIDE>IIOLOGICAL AND ESPERI>lEST.-tL EVIDESCE Benzene was first isolated in 1825. and commerciail!~ recovered from coal-tardenied light oil in IS49 and from petroleum in 1941 ( [ A R C . 1981).By the late I950.j. most commercial benzene was petroleum-derived. Chronic numan exposure` to benzene may result in adverse hxmatolo_eicaleRects. characterized b! a \,ariet>-of blood dyscrasias. including leucopenia. rhrombocytopenia. panc)topenia and anaemia. Benzene is also considered to be a human carcinogen 2nd is well established as a cause of leukaemia tl.4RC. 1981). I t is recognized as being carcinogenic by the sovernments o f Finland. the Federal Republic of Germany. Italy. Japan. Siieden. Switzerland 2nd the I-`S..4. B? comparison uith benzene. there is no direct rpidemiological e\ idcnse concerning the c.;rrcinogeniciir of tulwne and xyleneb. and aril) limiied e\perimeniai t i idence from Jnimdl studies. E!'IDE\CE OF BEhZEXE C.4RCISOGE\ICITI' I \ HL'.\l.A\S -3- 4 McMlCHAEL The first case report suggesting an association between exposure to benzene and leukaemia was given by Delore and Borgomano (1918). A n acute lymphoblastic leukaemia was observed in a worker in France. who was exposed to benzene for five years. Since then. numerous reports describing leukaemia cases among persons exposed to benzene nave appeared. predominintlv from Italy. France and Turkey (Goldstein. 1983). Several pioneering epidemiological studies investigating the relationship between exposure to benzene and leukaemia. as-wellas other I> mpho-haematopoietic cancers. were published in the mid-1970's. Aksoy published a number of reports on the risk of leukaemia among Istanbul shoe makers exposed to benzene-containing solvents (Aksoy el a/.. 1971. 1974: Aksoy. 1980). He estimated t h a t the annual crude incidence rate of leukaemia among the shoe makers was 13.5 per 100 000. compared with an overall annual incidence rate in the general population of 6 per 100 000. Vigliani (1976). reviewing earlier studies (Vigliani gL Saita. 19641. estimated the increase in risk of leukaemia to be 70-fold for workers exposed to benzene in shoe factories in Italy. As with Aksoy's studies. the apparent levels of exposure to benzene were predominantly within the 200-500 ppm ranse. However. these exposure data uere too imprecise for further quantitative analysis. Thorpe ( 1974) reported an epidemiological study of leukaemia incidence among workers from eight Esso Perroleum atliliates in Europe. Employees "ere divided into two groups: those with a potential exposure IO 10'0 or more benzene over a five-year period and those not exposed or only occasionally exposed. Among the exposed. 8 leukaemia deaths here observed. compared to the expected 6.6. yielding an agestandardized mortality ratio (SbIR)of 121 (no1 statistically significant 1. Among the non-exposed. the number of deaths observed n a s I O and the expected figure was 16.7. the corresponding S M R being 60 I not statisricall~siyificant t. These tigures indicate that the employees exposed to benzene experienced an approximatel:. two-fold risk of leukaemia u hen compared to non-exposed employees. No benzene exposure levels were available from the study. E ~ / ( ~ ~ ~ t ~\fl~tc/i~c'\( i~'Jllllh(/ll~lf ~l / ~l l ~~/ t l l.l l"i l l Il l ' ( '~ t ~ t ! t t l ~ if tiIt 't .. /~. ~ k101. Ic,lrX t / i ' t ) 1 \ ~ 1 The results ol' .' L.S. k a t i o n a l Instituie o f Occupational Sat'et! and Health r\IOSH) stud! ot ,1S H h i l e male isorkers e . \ p o d to benzene during the manufclcture ol'rubbttr h!drochloride have been publi>hedo n three occasions i Infante c t ( I / . . 1977. 1979: R l n A : O I d..1981 ). The first I N O paper<uere based on the results w i t h ' 5 O complete t'ollo\s-up o i \ i t a 1 StatUh. uhile the R i n A y ('I i d . i IWI ) paper \\as based o n d~ltdk i t h 9 Y 0 0 complete t'ollots-up. The findings &ere esheniiall! the same i n each of the three papsrs: there ttas a three-1'01d increasc in the risk ot' death from I! mpho-haemdtopoirtic cancers and an Jlmosr .;i.x-fL>ld increase f o r leukaemia specific all^ f IeuLaemia deaths i w \ i i s I .25 txpected). Ainong tsorkrrs smploJ.ed for more t h a n 5 )ears. 5 had d i d from leukaemia. compared to (J.2-7 expected. indicatins J 21-fold increase rn ri5L Rinsk? til. ( I9XI 1 & \ o t r J consider3ble discussion I O thc industrial hygiene measurements aLailahle for one of rhe t t k o 5tuc.i: Iocaiims in Ohio. HoueLrr. the investigiitors made no link betuern the d\.:iiLible ben7rne n~easurcrnentsand the CARCINOGENICITY OF BENZENE. TOLUENE AND XYLENE Table 1 Estimate of excess r i s k of leukaemca dealhs per 1 OOo workers exposed to benzene' 5 45 a 44-152 30 30-104 15 3 5-54 5 5-18 1 1-4 .cram while et at 179821 *RIPSRVer d l '9811 'On or a1 ,'978l ' 4 5 veara = occunaiacnal lifelime 48- 1 36 32-93 1648 5-16 1-3 5-16 3-1 1 1 5-5 05 2 01 4 4 ~~ 5-15 2-10 2-5 05 2 0 1433 employment histories of the indh.idua1 cohort members. (Subsequentlv. Crump and Allen (1984) and Rinskc pf ul. f19P5) did link these data in unpublished risk assessments. Cumularjve benzene exposures of individual uorkers in this cohort itere estimared from !ob-exposure matrices. This approach indicated a heightened. 45-fold. increase in risk of leukaemia in thz highest cumulative exposure categor). I Rinsky c-'[ 'If.. 1985). See ais0 .Appendix. p. 18.) Otr Y I 111. ( 1975) reported the morralit! of a cohort o i 594 DONChemical ILorkers exposed to benzene in the production oi chlorobcnzol. alk>l benzene and ethvl celluloje from 1938 to 1970. Their mortality experience was observed from 19iO throueh 1977. With expected deaths based on U S . Nhite male mortalirv. no statistically significant increases mere found in any cause o f death for the cohort 3s a whole. even m hen mortili[y u'as anal! sed by estlmatcd cumulari\e benzene dose (ppm-months)2nd latenc! Houeier. !.urther revie\ of the three cases n i Ieukaemla shoued J i l three to be myelocytic. \ \ i t h tun being classified a j acure. Ea>c'd on incidence ddta reporred in the L . S . Third yational Cdncer Surtm. [he s\>ecied incidence of m\.eloc?tic Ieukuemi3 it as only 0.3 cases. Because the Occupational Sat'rr! Jnd Health Administrarion in the L.S..\. 15 l -e ~ .~ i i ~ required to consider quantitative risk d.\essments Nhenrver possible from 'ii aildbic data. such a n assessment has: been a t t m p i t d (White d l . . 1382) usln? [he a b o ~ t mentioned tuo cohort studies. L'ncertarntieh regarding the Ietels J n j length?, O( expmure to beniene iicrc incorpnrJ[ed into the anal>sis and rhc results ;\re shoun in Table I B,ised on the onc-hi[ model. Jnd 3ubject to the tbpical unccriainiiej nl' ~ p i J e m i o i o ~ i ~datla. the :ih>e5sment ind~caissthar d H I v k i n g liie[ime s\po.;urr 10 benzene a[ a n tuposure k L t . 1 of' 10 ppm poses ;I subsranri.rl C Y i t j S risk of'dcath t'rom leuh emia &'hire cr I 19s:) wlt.ctd thc one-hit. or I i n t a r . mocic.1 because rhc blolo,+-~i mechani5tn ot' henmx-iiiduced leukaemia 1s noi understood \\ell enough i o iu.4[ij\ 3 more comp1ic:itc.d model. <ind boxuse ini'orm;ltion from cpidemiolo~ic;tI, I U ~ I ~ . C regardins the reldtiis risks o t ' l e u k a m l a difl'erent I e ~ e l sofexpobure 10 ben7cnr ild, 6 McMlCHAEL Table 2. Analyses of relative risk and excess risk of death from leukaemia per 1 OOO workers exposed to benzene lor an occupational lifetime* Authors Ott ef a / (1978) Infante et a/ (19771 Rinsky et dl (1981) Vigliani (19761 Aksoy ('9771 375 56 21 20 25 1-30 ppm for 8-9 years 10-100 ppm. 8 5 years average 10-100 ppm. 2 5 years exposure 200-500 pgm 9 years average 150-210 pprn 9 7 years average 72 170 140 475 534 26720 17-170 14-140 95-238 254-356 2672 17- 17 14- 14 10- 24 25-36 ~~ 'Table aaaorea lrom Infante ana W h m 119851 Calcuiatlonr In tne 1.11 tnree coumns were oe:T(ormea vamp the method aoolma by IARC 1982) IO data from inlante ef /I. I19771 ana Rinshy et af il91111 inadequate for testing more complex models. Furthermore. as Infante and White ( 1985I have subsequently noted. several indications of benzene toxicity to bone marrow and peripheral blood cells have demonstrated a linear dose-response relationship. This linear. non-threshold model assumes that ever? increment o i dose is accompanied by a commensurate increment in the e x c s s cancer risk. The use of thls toxicologically plausible model allows extrapolation 01 risks from relatively hish dose levels. hhere cancer responses can be measured. to relatively low dose levels. where such risks are too small to be measured directly through animal or epidemiological studies ( W H O . 198I ). The IARC (1981)also conducted a quantitative cancer risk assessment for benzene 3nd Irukaemia usins the results ot'four studies (C'isliani. 19'6: 4ksoy. 197': Orr er ai.. 1978: R i n s k y Y I d..1981 I IARC assumed a linear dose-response relationship. k i t h cumulative benzene dose predicting the relative risk of developing leukaemia. I t also made slightly different ;ijsumptions from those 01' White c f d.( 1982) regarding the length o f exposure for the NlOSH (Rinsk?) and DOHr Ott I studies. For purposes of comparison with the estimations of White er ol. (1982). a summar! of the IXRC ( 1982)assessment of b e n x n e risk. uith extrapolation to l e ~ e l sof I O ppm 3nd 1 ppm is shclun in Table 2. The calculations in the last three columns uere performed by Infante and White (1985). using the method applied by the IARC (1982) to rhe NlOSH stud! and taking account of the IARC's opinion that " risk calculations (should) reflect the degree of uncerrainry in the estimates of dose rate . . . by ciring upper and loher bounds of such estimates" (IARC. 1981).The estlrnated excess risks oileukacmia associated uirh uorking lifetime (45!ears, ofesposure to I O pprn and I ppm of benzene are similar in Tables 1 and 2 . Recognizing the need for further quantitative ini.estigations into the relationship between benzene exposure and 1) mphatic and haematopoietic cancers. Wong 1983) CARCINOGENICITY OF BENZENE. TOLUENE A N 0 XYLENE 7 carried out a l a q e industry-wide mortality study of U.S. chemical workers occupationally exposed to benzene. The population in this historical cohort study consisted o f 4 602 male chemical workers from seven plants. who were occupationally exposed to benzene for at least 6 months between 1946 and 1975. and a comparison eroup ot' 3 074 male chemical workers from the same plants who weie'employed for it least 6 months durins the same period of time. but who here never occuparionally zxposed to benzene. Exposure to benzene was divided into two categories: continuous exposure and intermittent exposure. Continuously-exposed workers Bere classibed according to both an eight-hour rime-ueighted aberage ( T W A )and a peak evposure level. whereas intermittently-exposed workers were characterized by a peak exposure level only. The eight-hour TW.4.s Here estimated by compan! industrial h!gienists. using uniform criteria: nevertheless. i t "as not possible to quantify accuratelv the exposure of workers to benzene. The major findings of this study were as follows. When compared to the L.S. population. the SMR's in the exposed group for all IFmphatic m d haematopoietic cancer combined. for leukaemia. and for non-Hodgkin's lymphoma (lymphosarcoma. reticulosarcoma and other lymphoma I were slightly. but nor signiricantly. higher than the national norm. When the group with no occupational exposure was used for comparison. the exposed group (continuous and intermitrent)experienced a 1)mpho- haematopoietic cancer relative risk ( R R )of 2.99 (borderline statistical significancei. This excess u a s primarily due lo seven leukaemia deaths in the exposed group. WSUX none in the comparison group. For the continuously exposed group. the R R for Iympho-haematopoietic cancer \\as 3.20 rp < 0.051. The a s s o c i ~ ~ ~bi ostnu-een exposure IO benzene. and. specificall!. leukaemia was of borderline significance H hen those exposed continuously and those exposed intcnnittcntly bere grouped togerher. but the association berueen continuous exposure alone and leukaemia was statistically significant ( p< 0 05).None of the seven leukaemia deaths h a s of the x u t e myelogenous cell type. The R R for non-Hodgkin's 1)mphoma among H hite males in the continuous]? exposed group uas 5.02. but b a s not significant ~p = 0.12). Fu rt herm ore. the da ta demonst r3t ed st aI i st i ca11) L,I gn i ri c;in I d o w- rcspo n se r e l ~ t i o n s h i pb~erueen the estimated cumulati\ e exposure to benzene 2 n d mortaiit! t'rom all I!nipho-haematopoietic cancers combined and !'rom leukaemia. Those chemical workers H i t h a curnulati\e c:\po,ure of at least -20 ppm-months e\perirnced ~n RR of 7.93 for I!mphatic a n d haematopoirttrc cancer. uhen comp.ired to the Hnrksrs with no occupational exposure. Howe\.er. as M'ong ( ISS?, dcknnuledges. the limitations in exposure estimates preclude useful dose-responw anal?315 dt ION bels dt' exposure. In a more recent rlttempr to drri\e quantitatiLe risk assessmenis of' leukaemia in zubrects exposed to benzene. Crunip 2nd . 4 l h 19841 used the data from 011c'i 01. t Ic)?S). Rinsk? t'r ui. 1 l '9x1) d n d l i L m g ( 1 9 8 3 . ;ilons \{ith animal e\periinental data from inhalation studies. The estimdtes of Irukacmia risk derited b! Crump and .Allen ( 1384) from dnimal inhalation yludies (Tdblc 7 ) . uhile q u i t e iansistent. arc 311 c ~ n s i d e r a b l yless than those made from human data table 2 ) . .Acceptins the dsjumptions underlyins these rstirnates. this indicates that rodents ma! be I e b j jusseptible than humans to benzene-induced leukaemia. 8 McMlCHAEL Table 3. Estimates' of Iilelime human risk per 1000 exposed persons' from animal inhalaiton studies' 011. w1 All squamous-cell carcinomas male rats. NTP (19841 All squamous-cell carcinomas male mice N l P ( 1904) Lebkaemia both sexes ra!s Maltoni et a / ( 1983) Leukaemia male mice Goldsteln er dl 11982) 9anzane m r concentration 'DWW 1 _- 10 1 10 1 10 1 10 RIU 0 77 7 65 2 01 19 90 0 15 1 46 0 04 0 38 'Assummq risk to anmais ts :ne same IS risk io numans Ynen aose :I measurea In mqikg DOdy WcqnUdJv 'Lenqin ol human eioasure 10 aenrene = r0 yean "a01e bnsea on Crumo ana Allen 1!904\ Crump and Allen 1984) consider that their estimates from human studies. based upon cumulative or weizhted exposure. are the most reliable. The attendant estimates of additional leukaemia cases per 1 000 workers exposed to I O ppm of benzene ranze from 0.5 to 2.6 cases from one year of exposure beginning at age 20. to 15 to 88 cases from 40 years of exposure. The "uindou exposure" variable (defined as a jhorterterm exposure occurring at the putative. latency-related. cancer induction time) tends to give lower estimates than either weighted or cumulative exposure. However. C r u m p and Allen 1984) judge that the estimares based on either cumulative exposure o r ueiehted exposure are preferable 10 those based on window exposure. &on-linear models would predict about the same results 3s linear models at high doses ( I O ppm for 40 years). X person exposed to I O ppm for 40 Fears would attain a cumulative exposure of 400 ppm-yxrs. uhich. in the Rinsky e f oi.( 1981 cohort. would place him in a c a t e ~ o r yfor uhich a signihcanl elcess af leukaemia occurred. Indeed. the subsequent re-anal)sis b!, Rlnsk! c'r d.( 1985, gdve an estimated 45-fold increase in risk of leukxrnia for that c a t e z o p of exposure. Thus. no extrapoiation to low doses is required to estimate risk from lifetime occupational exposure to 10 ppm. However. the assumption orlineurity pla>sa more imporrant role at lower eKposure levels. and the estimaies o f risk are more uncertain. This 13imporrant. since. under the benzene exposure limit of I O ppm prevailing in the U.S..4..the risks csrimated from continuous exposure to I O ppm represent a theoretical maximum. 3s i t IS unlikcl) [ha[ a n y current worker'j exposure will approximate this 11'the I O ppm standard is srrictl> enforced. Crump and Allen ( 1933)point out that their estimates of risk from human data were limited to leukaemia. as this appeared to be the oni: r:pe ofcancer conclusively linked to exposure to benzene in humans. This approach could underestimate cancer CARClNOGENlClTY OF BENZENE. TOLUENE AND XYLENE 9 risk if. in fact. benzene causes other cancers. as it does in rodents. The recent epidemiology literature (Decoufle et ai.. 1983; Rinsky rt ui., 19x5) suggests that benzene exposure may also cause multiple myeloma in humans. Other less yuantitarive or less spc~lfirstudies -- In a cohort study of 35000 workers at 8 petroleum refineries in the U.K.. the leukaemia S M R %as found to be 94 (Rushton & .Alderson. 1981a). Using data from this cohort study. 36 cases with leukaeinia mentioned on the death certificate "ere the subject of a subsequent case-control \tudy (Rushton & Alderson. 1981b). Benzene measurements were not available. blit each worker was assigned. based on work histories. to either low. medium or high levels of exposure to benzene. A t\vo-fold increase in risk was detected for uorkeri with medium or high exposure. re1atiL.e to those uith low exposure to benzene, when length of service u'as taken into account. This tuo-fold increase was of borderline statistical significance ( p = 0.05).There was no evidence that the risk increased with length of service. Decoufe PI al. ( 1983) conducted a historical cohort mortality study of 259 male employees of a chemical plant (converted from a petroleum refinery) t+ here benzene was used in large quantities. Unfortunately. no quantitative measures of exposure were available from the plant. Four deaths from Iynpho-haematopoietic cancers were observed. when only 1.06 deaths uere expecred: the SMR of 377 u3s statistically significant. Three of these four deaths uere frL)m leukaemia: one was chronic lymphocytic. one acute monocytic and one acute myelomonoc>.tic. This last leukaemia case also had a history of multiple myeloma. Since the fourth I>mphohaem3topoietic cancer death &as due to multiple m>eloma. the authors suggested a possible link between benzene rind multiple myeloma. A number of epidemiological studies ha\.e found significant mortality e.xcess from leukaemia among workers exposed to solvents in the rubber industrl,. Houever. the relationship between this observed leukaemia excess and. specificall>. benzene is less clear. I n one cohort. benzene was not the predominant solvent used and the authors were hesitant in ascribing IO benzene the sole responsibility for the excess ofleukaemia (Monson & Sakano. 1976). I n another cohort nirubber uorkers. the e.\c'sss risk NJ\ limited 10 lymphocytic leukaemia and ~ n nsot dssociat?d u x h either monoc!.tic or m!eloi!iic leukaemia iJlc.llichael cr d..1975. 19-6: Lt'olf P I id., I9YI ). HoueLer. workers in this cohort acre exposed to other soltents ;IS uell as benzene. ;Ilihough the Litter ot'ten occurs as 3 contaminant of oilier soltents 1c.g.. petrolcum naphrha). To further in\esligale in this latter cohori the relationship betueen the obsert.ed excess ot'l>mphocytic leukaemia and esposure to \arious solvents. .Arp of til. 1983) examined the soltent exposure histories of I5 cases o i Iyrnphcqric Iruii,iemla and 70 matched controls H i t h i n this second cohort. A n indi\.idual HXS classitid as exposed to a soltent if the cumulrrtise length oie.cposure exceeded 12 monrhs. The relatiktt risk estimates s e r e elevated for exposure to ben/ene 1 R R = -1 501 3 r d io other solkenis I R R = -1 50).Seither tinding reached statistical signiticance at the 0.05 level. .A much more detailed dnalysis o i this same population as reported b> h'11c'oskt~'i ui.I I9Y-l). Detdils 01' estimated indikidual exposure to each of' 3 solLents irere obtained. L! mphosarcorna and 1) rnphatic leukaemia uere strongly ~ssociarcdu ith exposure to 10 McMlCHAEL carbon tetrachloride and carbon disulphide. whereas their association with benzene and xylene was of marginal statistical significance only. Several epidemiological studies of petroleum refinery workers who are potentially exposed to petroleum products. including benzene. have also found an excess of lympho-haematopoietic cancers (Tabershaw,CaoperAssociates. 1974: Thomas er al.. 1982; H a n k er al.. 1982). However. none of these refinery studies provided any direct link between the observed mortality excess in Ivmpho-haematopoietic cancer and benzene exposure. Eraluurion of the epidemiologicai eridewe On the basis of these studies there is substantial ~pidcmiologicalevidence that persons exposed to benzene (or an associated impurit: 1 experience an elevated risk of certain types of lymphatic and haematopoieric cancer. parricularly leukaemia. Reports providing such epidemiological evidence include those of Aksoy er al. ( 1974). Ott ef al. (19'8). Vianna and Polan (1979). Aksoy (1980). Rinsky et al. (1981). Rushton and Alderson (1981b). Decoutle er ai. (1983) and Wong (1983). The most frequently reported f o m of leukaemia associated with benzene exposure is the acute myelocytic form. However. research on rubber workers (McMichael el ul., 1975. 1976) indicates that. in chronic low-level exposure situations. lymphocytic leukaemia may predominate. Perhaps this is a more likely manifestation of non-acute bone marrow insult. (Chronic forms of leukaemia are also less likely to be reported on death certificates as the underlying cause of death.) On the other hand. there are reports indicating no increased risk of lymphohacmatopoieric cancer 3s ;i result of benzene exposure. h recent cohort study (Tsai vi al.. 1983) of 454 male refiner? workers exposed to benzene during 1952 to 1978 at a Texas refinery did not find any Iympho-haematopoieric cancer deaths: however. the cohort size was small and the statistical power to detect a modest risk in either lympho-haematopoietic cancer or leukaemia was limited. The case-control studies of I>,mphocyticleukaemia among rubber workers (Arpe! ai.. 1983: Wilcosky ei of..1984) indicated that the lymphocytic leukaemia excess in the rubber industry came primarily from carbon tetrachloride and carbon disulfide. rather than from .x>'leneor. even less Iikel!. irom benzene. The large mortality study oiThorpe I 19'4) o i 38 OOO petroleum uorkers potentially exposed to benzene has often k e n cited as shoHlng no evidence of significant leukaemia excess as a result of exposure to benzene. This interpretatikm. however. ignores the comparison of the exposed workers with the control uorkcrs (from the sdme t'acilities. but not exposed to benzene): the exposed group had a twofold risk for leukaemia relatite to the non-exposed group. There are a number ~ 3 foutstanding issues reprding the relationship between exposure 10 benzene and increascd risk o i I~mpho-haematopoleticcancers. First. none of the published studies pro\.ides adequate exposure inforniation for satisfactory dose-response analysis. Second. the choice of an appropriate comparison IS an important issue. In a number oi studies. anal>,sesbased on comparisons xith the general population fulled to detect an increased risk for I\..mpho-haematopoietic cancer or leukaemirl. In the Thorpe I 1974) study in Europe. the indusrry-wide study of refiner), Horkers in the C.5.4. (Tabershaw Cooper .4ssociates. 19-41 m d the CARClNOGENlClM OF BENZENE. TOLUENE AND XYLENE 11 Table 4 Studies Of Chromosomal breakage in workers exposed lo "low ' atmospheric concentrations of benzene Aumor Picciano. (1979) Sarto ef a/ 11984) Love1 01 exwsure lo tmnzenc i T W A i -- Controls 1 Ppm 1-2 5 ppm 2 5-10 ppm 3% (workers) 22% 2646 33 Do' Range of snort-term samples = 0 2-12 4 ppm 1 2% (cells)' 2 7% 'Benzene measured In B-h WP.s chromosome breakage = creams dlcenircs iransiocaimns anc e x c i a o g e qures * * D z *. of workers wain chromosome o'eaaaqe '9enrene measurea in 5- an6 a m i n samoies cohort study of eight oil refineries in Britain ( Rush ton & Alderson. 198la 1. the initial calculations of the leukaemia S M R . based on a comparison uith [he _general population. revealed no significant excess. However. subsequent analyses that rook account of variations in benzene exposure between categories of workers reLealed positive associations with leukaemia. Third. although some studies (Aksoy. 1980: Ott et ai.. 1978: Rinsky pf ai.. 1981) indicate that acute myelogenous leukaemia is the type frequently associated with benzene exposure. other reports show associations uith different leukaemia cell tFpes (Tabershau Cooper Associstes. 1974: .LlcMichaeler d . , 1976: Linos et ai.. 1980: Schottenfeld e! d . .198I ). The studies of Decoutli p r uf. 1983) and Rinskl; e1 ai. (1985) have also suggested an increased risk o i multiple m>eloma. .Urrrug~nic.ii!m. d cAroninsomal rfik.rs in hwtianJ Numerous studies have been carried out on the chromosomes of bone-msrrou cells and peripheral I>rnphoc>res from people known lo have been exposed to benzene (Dean. 1978; Picciano. 1979; Watanahe er al.. 1980: Sarro t'r ~ 1 . .1984). The p o p u l ~ t i o n sincluded in these sludies Fdll into tu0 general categories: 1 I patients uiih either a current or a past hi3iorv of benzene-induced blood d!.scr3jiaj i'-benzene h;lemopathies"L often associated uith e\iensi\e exposure 10 benzene: and I 2 1unrkers ti ith k n o n n current o r past e.xposurr [ obenzene. but H i t h no ob\ ious clinic31 rtTeclj. In some ot'thesr studies. signincant increases in chroiiiosomal aberrations nit\e been >ern: in some cases these h a w persisted for years after cessation ot'esposurr (1.4RC. I982 1. The tindings trom tu0 studies are summarized in Table 4.Picciano 19-91 reported that 5 2 uorkers exposed to average benzene conceniratrons belou 10 ppm demonstrated 3 stat1stic3ll? significant t u 0- to three-fold excess of chromosomal darndgr.. compared to that absrrled among 44 unexposed aorkers. Furthermore. the percentage 01' uorkers cxhthiing both chromosome breahh and marker chrornos~~rne~ U ~ 2S0. 35 and 31"" !-or uorkers categorized 33 n p o s e d to al'crsge trcn7cne concentrations belou 1. 1.0-2.5 Jnd 2.5 - 10.0 ppni. req-wc[i\-el!. Lcbs than 3 O 0 GI' rhz unexposed u orksrs exhibited both chromosome breaks and mJrker chromosome5 Yore rccenrl?. Sarto el uI.i1984) hake rcportcd wpportiw. albeit *taker. s\idenie of chromosomal breaks in Horkers esposed to benzene. 12 McMlCHAEL Other cancers in humans Alihou_eh induction of types of cancers other than of the lympho-haematopoietic system has been suspected. these have not been adequately evaluated from epidemiological studies of workers exposed to benzene. The studies have not had sufficient sratisticai power to detect low exctss relative risks. However. since 1978. experimental studies have demonstrated that benzene administered either bv oral gavage or by inhalation induces cancer at multiple sites in experimental animal;. This research is reviewed further on. No cpidemiological studies ot' differentiated exposure to toluene or xylene have been reported. It is possible that some or all o f t h e aboi.e-mentioned increased cancer risk associated with non-specific exposure to solvents is attributable to these. or other. non-benzene solvents. However. evidence of any such effect will be difficult to obtain. particularly since benzene occurs as an impurity in toluene and xylene. ANI\lAL EXPERIMENTAL STUDIES OF BENZENE Intragastric administration of benzene dissolved in pure olive oil to female rats induced ZJmbal gland carcinomas. mammary gland c3rcinomas and leukaemia. With the exception of leukaemias in the high-dose group. no such tumours occurred in male rats (Maltoni & Scarnato. 19791. Experimental srudies of lifetime benzene inhalation at 300 ppm im mice show the widespread occurrence of anaemia. lymphocytopenia. neutrophilia and bone-rnarroN hypo- or hyperplasia (Snyder et af..1980). Statistically sisnificant increases in the incidence of Iympho-haemaiopoietic tumours (lymphocytic lymphoma. plasmacytoma and myelogenous leukaemia) also occurred {Snyder e[ of.. 1978a. 1980). However. no such tumour occurred in similarly txposed (malei rats (Snyder et ai.. 197Sb~ Subsequenl studies b) %t.iltoni and co-workers h a w shown that benzene is 3 potenr carcinogen in rats of both sexes. uhether ingesttd or inhaled: benzene produces different types of tumours in different organs and IS lherefore ;L "multipotential carcinogen" [Maltoni PI d..1985). Since all the benzene inhalation studies reported to h t e entail concentrations in the 200-300 ppm range. there is no direct animal experinicntal evidence relating to low-concentration exposures that now t>.pifyh u m a n occupational exposures. Long-term carcino_eenicity bioassa>s clt 50. 25. IO. 5 and I ppm may therefore be helpful for scientific risk assessment in humans. A n earlier comprehensive re\ien o f r h e experimental data (IARC. 19Y2) concluded that benzene does not induce specific gene mutrltions in bacterial test s>srems. in CARClNOGENlClTY OF BENZENE. TOLUENE AND XYLENE 13 Drosophih melunogasrer. or in mammalian cells. However. there was evidence that benzene induces cytogenetic abnormalities (chromosomal aberrations in bone- marrow cells) (IARC. 1987). Furthermore. the micronucleus test in mice and rats has been consistently positive. and numerous studies have shown that exposure of experimental animals to benzene in vivo leads to the induction 'of chromosomal aberrations in bone-marrow cells (IARC. 1982). While Van Raaite cf ul. (1984) have argued that only peak exposure to benzene above 100 ppm can cause leukaemia or bone marrow mxicity. the results of recent rpidemiolo_eical and experimental research suggrst otheruise. Several recent animal experimental studies have demonstrated adverse etTects on chromosomes and bone inarrow' as a result of low-level exposure to benzene. I t has recently been demonstrated by Tice er (11. (unpublished data quoted by Int-ante & White. 1989 that exposure to benzene at a concentration of 10 pprn in air. 6 h dav for 9 days. induces a significant increase in micronuclei (.a measure of bonemarrow chromosomal damage) in peripheral blood cells of mice. Mice exposed to 2s ppm benzene for a total of only 4 h had a two-fold increase in sister chromatid e.xchanges ISCE) in b o n e m a r r o w cells. The etTect on mouse bone-marrow cells increased linearly with increasing levels of 4-h benzene exposure. indicating a linear dose-response. Previousl?.. Kligerman had reported that mice exposed 10 10 ppm benzene by inhalarion for only 6 h demonstrated a significant increase in SCE in peripheral blood lymphocytes and in micronuclei in bone-marrow rryrhroc\ res (Kligerman er nl.. 19831. Gad-El-Karim ef al. 1983) have reported a dose-response increase in micronuclei in bone-marrou erythroc>tes of mice adminisrered 3 total of tuo oral doses of benzene at concentrations as low as 8.8 rng kg body weight. This dosage w.ould be equivalent to t-o 8-h atmospheric exposures of Jpproximately 6 ppm. Collectively. these studies demonstrate chromosomal damage in bone-marrou cells a s a result of the equivalent of one to several working day exposure to 10 ppm of benzene. They also demonstrate a linear dose-response for benzene exposure 3nd chromosomal damage ta both marrow and peripheral blood cells. Since most leukaemids and related disorders in man seem IO Invoite stem-cell anomalies. the response of harmatopotetic \[em cells u;is intrstigated by Bdarson uf.( 1981)w ho exposed mice hy inhalation to 10 ppm benzene for 6 h d a \ . 5 d ~ y ~s e e k for -??. 66 and 175 da!,s Progenitor cells from rhe esposed mice shoived markedlv reduced Abilities to i'orm colonies. compared to cells from control mice. In another recent stud?. shori-term exposure t6 h day for b days) to I O ppm henzcne in air signincdnrly depressed miiogcn-Induced hiastogenesis ifboth B- jnd T-lymphoc?tes in mice ~ R o s e nc'f ( I / . . 1981). The authors concluded that benzene sxposure ai or near 10 ppm may aHec~immune f'unction. Thus. less than one ucek of exposure to 10 pprn 01 benzene is assocured u i t h chromosomal damage 10 bone-marrow cells. sryvtiiant depression of rhr. bone inJrrou and disturbances o f immune $!stem I'unction. Since most leukarrnids Jnd reiared disorders in man seem lo in\ ol1.e sttm-cell .Jbnormalities and iiiimune S\ >tern debcirncres. rhese animal ssperinwntal findings ma!' be highlL relevant 10 l i u k x m i a and exposure to benzene in humans. The tindinzs proLide fuirhrr support f'dr the conclusion that L\II elevated risk of A bonc-marro\i prolifcraii\c cancer I4 McMlCHAEL (leukaemia) could be expected as a result of exposure to 10 ppm benzene (Infante & While. 1985). - -ANIMAL EXPERIMENTAL STL'DIES OF TOLLEYE ASD XYLENE In a series of experiments. primarily directed at studying benzene carcinogenicity in rats. Maltoni CI d.(1985) have also studied the effects of other benzene-correlated aromatics (toluene. xylene and ethyl-benzene). A t high concentrations. toluene and xvlene (and. to a lesser extent. ethyl-benzene) caused an increase in the total number 01-malignant turnours. The increase in tumour yield was approximately one-third that induced by an equal dose of benzene. N o orher experimental data on this question appear to have been published. SL'XI h1.4R Y Ben:ene The evidence for carcinogenicity of benzene in humans was evaluated by the IARC in 1982 as follows: "It is established that human exposure to commercial benzene or bentenecontaining mixtures can cause damage to the haematopoietic system. includine pancytopenia. The relationship between benzene exposure and the development 07 acute myelo_eenous leukaemia has been established in epidemiological studies. "Reports linking exposure to benzene u ith other malignancies were considered to be inadequare for evaluation. "There is sutfrrienr evidence that benzene is carcinogenic to man." This evaluation now warrants some elaboration and updating. While the epidemiological evidence concerning benzene carcinogenicity is strongest for acute myelocytic leukaemia. there is some limited evidence of increased risks of chronic myeloid and chronic l!mphocyric leukaemia. In addition. recent studies have suegested an increased risk of multiple myeloma. while others indicate a dose-related increase for total lymphatic and hacmatopoietic neoplasms. Corroborative evidence for such ;L generalized effect comes from experimental studies showin? that exposure to benzene dcprssses all Iympho-haematopoietic sell lines. While onl! limited evidence of benzene cclrcino_eenicity in experimental animals exists. the receni findings of the Sational Toxicology Program ( N T P . 1984) in the L'.S..-4.and M ~ l i o n i01ui. (1985) strongl) indicate that benzene is an experimental carci nogen. Toircc*niu~nd I I iotw While no direcr humdn etidence i s JLailable. there is recent etidence of carcinogenicit! of toluene m d x> lene at high concentrations in experimental animals I t should ~ 1 5 0be nored t h d t J n ? future r p i d e m i o l o ~ i c ~olb w \ a r i o n s of cancer rishs ,issociared w i t h toluene or \!lene would h j t e to take ,iccount ot'the suspected effects of benzene impurities CARCINOGENICITY OF BENZENE. TOLUENE AND XYLENE REFERENCES 15 Aksoy. M.(1977)Leukemia in workers due to occup3tional exposure to benzene. . Y m . - -lsranhul Contrib. din. Sci..I t , 3- I3 ,Aksoy. M . ( 1980) Different types of rnalipancies due t o occupational exposure to benzene: a review of recent observations in Turke:,.. Environ. Rrs.. 23, 101-190 Aksoy. M..Dincol. I(..Erdem. S . & Dincol. G . (1971) Acute leukemia due to chronic exposure IO benzene. .4w. J. .\.fed..52. 160-166 Aksoy. M.. Erdem. S. & Dincol. G. (1974) Leukemia In shoe-norkers exposed chronically to benzene. Blood. 44,837-841 Arp. E.W.. Wolf. P.H.& Checkoway. H. (1983) L>rnphoqtlc leukaemia and exposures to benzene and other solvents in the rubber Industry. J . O c m p .\fed.. 25, 598-602 Baarson. K . A . . Snyder. C.A. & Albert. R . E . (1984) Repeated exposure o f C57B1 mice to inhaled benzene at 10 ppm markedly depressed eryrhropoietic colony formation. To.\-irol. L e f t . .20, 337-342 Crump. K . S. & Allen. B.C.( 1984) Quanrirarive Esiitriarrs o l Risk o i L w k a e m i a from Ocrupurional Exposure to B m m e . Unpublished report prepared for U.S. Occupational Safety and Health .AJ m i nistra t ion Dean. B.1. (1975)Genetic toxicolog>.of benzene. toluene. xylene and phenols. ,\f1dfur. Res.. 47, 75-97 Decoufle. P.. Blamer. W. & Blair. A . (1983) Mortality among chemical Horkers exposed to benzene and other asenrs. Etrriron. Res.. 30. 16-25 Delore. P. & Borgomano. C. (1928)Acute leukemia folloiiing benzene poisoning. J. .Wed. L?*on.9, 227-233 Gad-El-Karim. M . M . . Harper. B.L. & Legator. M . S . (1981) Modifications in the rnyeloclastogenic effect of benzene in mice u i t h toluene. phenobarbital. 3methylcholanthene. Aroclor I251 and S K F - 5 3 A . .\Iididr. Re.r.. 135, 3 - 2 4 3 Goldstein. B. 11983) Benzene is still u i t h us. .-Itit. J . I t i d . .Ud. 4 , 585-587 Goldstein. B. D.. Snvder. C.A..Lxkin. S . . Bromberg. I.. .Albert. R . E . & Nelson. 5 I 19s:) htyelogcnous leukaemia in rodents inhalin: b e r u n e . T O W O / .Lt,tr , 13, 169- 173 Hanis. \.\I.. Holmes. T. 34.Shallenberger. L.G. 6r Jones. K .E. ( 1981) Epidemioiog! stud! o i refiner! and chemical plant uorkers. J . 0cc.up. V d . .24. 107-: 12 I . A R C ( 1982) 1.4 RC .\jOt?f),CP'tJph.Y 0 1 1 lht' &I'N/I/OI/Ot! Ot ! / I [ * C [ J r ~ ' / l i O ~ r ' tRl /i<s'k of Chtwiicds to H i i t w r i s . 1-01. 29. Sotrie ltidii.sir/(il Ciicwiic uis ( i t i d D.i-c~srrrtfSL.! on. International ,Agency for Research on Cancer. pp. 9.: ! 4 -.Infante. P 8: White. X l . (1985) Projections ot' Ieukciria risk 3ssociated u i t h occupational exposure to benzene. ,-It??. J. Itid .\l~d., 4 0 3 1 1 ? Inl'ante. P.F.. Rind!. R..L\.. Wagoner. J . K . & I'ouns. R . J . I 19771 Ltuktrnia in benzene uorksrs. Luti(ei. 1. 76- 78 Infante. P F.. Rind!. R..A.. Wagoner. J . K . 6r I'oun:. R J . (19791 Leukcmi3 in benzene u orkers. J E!ii,iroti. P u t / i o / . To.~ic.ol2.., :5 1-25Kligerrnan. .A. D.. Erexson. G. L. & N'ilmer. 1.L. ( 19831 De\slopinent of rodent peripheral blood !I mphocyre culture systems to Je[cct c!-ropwtic damage in \]YO. 16 McMlCHAEL (Abstract) lnt. S!vnp. Sisrer Chrotnarid Eschon,qe. Brookhaven Natl. Lab.. Dec. I983 Linos. A.. Kyle. R.A.. O'Fallon. W . M . & Kurland. L.T. (1980) A case-control study of occupational exposures and leukemia.-fw. J . Epideniiol., 9 , 131-135 Maltoni. C. & Scarnato. C. (1979) First experimental demonstration of the carcinogenic effects of benzene. .\fed. Lur., 5 , 352-357 Maltoni. C.. Conti. B..Cotti. G. & Belpoggi. F. (1983)A multipotential carcinogen. Results of long-term bioassays performed a t the Bologna Institute of Oncology. .-lj>i. J . h d . .\led.. 4, 589-630 Maltoni. C.. Conti. B.. Cotti. G. & Rclpoggi. F. (1985) Experimental studies on benzene carcino_eenIcityat the Bologna Institute of Oncology: current results and ongoing research. .hJ.. Ind. Med., 7 , 4 1 5 4 4 6 McMictiliel. A.J.. Spirtas. R.. Kupper. L.L. 8: Gamble. J . E . (1975) Solvent exposure among rubber workers: 3n epidemiological stud!. J . Occup. ;\led.. 17, 233-239 McMichael. A.J.. Spirtas. R.. Gamble. J.R. & Tousey. P.M. (1976) Mortality among rubber-workers: relationship IO specific jobs. J . Occicp. Med., 18, 178-185 Monson. R. R . & Nahmo. K . K . (1976) Mortality among rubber workers. I . White male union employes in Akron. Ohio. .-fmJ.. Epideniiof..J03, 284-296 NTP ( 1984) National Toxicology Program. Tosicoiogj*and carcinogenesis srudies qf henzerie iri F34J .V rais utid B6C3F mice ipui,age studies/. NTP-53-07?. NIH Publication 34-2545. C.S. Department o i Health and Human Services Ott. M . G.. Tounsend. J.C..Fishbeck. W . A .& Langner. R.X.( 1978)Mortality among individuals occupationally exposed to benzene. .irch. Enviroti. Heafrh. 33, 3-9 Picciano. D. ( 1979)Cytogenetic study of workers exposed to benzene. Enriron. Rrs.. 19, 33-38 Rinsky. R . A . . Young. R.J. & Smith. A . B . f 1981) Leukemia in benzene workers. .4m. J. lnd. ,Wed.. 2. 21 7-45 Rinsky. R..\..Smith. A . B.. Hornung. R.. Filloon. T.G.. Young. R.J.. Okun. A.H. & Ldndrigan. P.J . ( 1985) Bcw-cwe uiid k~ukeiiiru.uti eprdcmiologic risk ussessnienr. Unpublished report. U.S. Department ot' Health and Human Services. Public Health Serlicc. Rosen. M.G.. Sn>dsr. C..A. & .Albert. R . E . r 1 9 Y l l Depressions in B- and TI>mphocyte mitogen-induced blas[opiesis in mice s.xposetd to low concenirdtions of b e n m i e . To v i ( 01. L c t t . , 20. 717--%Y Rushton. L. & .4ldrrson. 31 R. 11981a) A n epidmiological survey of eight oil retineries i n B r i u i n . Br. J . I d . .\/(Y/.,38, 2 5 - 2 3 Rushton. L. & .Alderson. M . R . 11981b) X case-conirol stud!, io investigate the associalion betibeen ciposure IO benzene and deaths from Ieukcmia in oil refiner?. workers. Br. J . Cumw. 43, 7 - 4 4 Sirto. F.. Cominuto. I . . Pinion. A . V . . Brovedani. P . G . . \lcrler. E.. Peruzzl. 31.. Bianchi. V . & Levis. A.G. I1381).1qiogeneiic stud! oi \rorkers exposed 10 Ion ioncentraiions 01' benzene. C'irrriioec~rrcJsrs5.. 82'-5!2 Schottentild. D.. U'arshauer. 51. E.. ZJuber. .A.G..Jlctiklr. J . G . S: Hart. B. R . I 198 1 I ,4 prospective siudl; ut' morbidity a n d morrality i n petroleum industr? empio!ees in the L'nited Sbteb - J preliminary repori. In: Qiiotitrficurron of- Ckcuporronui Ccrnccr. . B1rtihirr.i. Report Y '. Cold Spring Harbor. S Y . CSH Press. pp. 147-265 CARCINOGENICITY OF BENZENE. TOLUENE AND XYLENE Snyder. CX.. Goldstein. B. D.. Sellakumar. A.. Albert. R. E. & Laskin. S. (1978a) The toxicity of inhaled benzene (Abstract No. 106) To.Yicd. .4ppf. Phurmucoi.. 45, 265 Snyder. C.A.. Goldstein. B. D., Sellakumar. A.. Wolman. S.R.. Bromberg. I.. Erlichman. M . N . & Laskin. S. (19786) Hematotoxicity of inh'ded benzene to Sprague-Dawle); rats and A K R mice at 300 ppm. J . To.ricol. Emiron. Hculrh. 4, 605-6 18 Snyder. C..4.. Goldstein. B. D.. Sellakumar. A . R . . Bromberg. I . . Laskin. S . & Albert. R.E. ( 1 980) The inhalation toxicology of benzene: incidence in hematopoietic neoplasms and hematotoxicity in A K R , J and C F B L 65 mice. To.uicoi. , . ~ p p / . Phumiac~ol.5. 4, 373-3 1 Tabershaw Cooper Associates ( 1974) ..It t i w i d i i ! . sfuc(r. o / ' p c i r o l r u tr~c~fliier?, \c.orkers. The American Petroleum Instilute. Berkeley. California Thomas. T.L.. Waxweiler. R.J.. Jloure-Erasso. R.. Itaya. S. & Fraumeni. J.F.. Jr. ( 1982) Mortality patterns among worker.; in three Texas oil refinerits. J . O ~ c u p . .\.fed..24. 135-1-Il Thorpe. J . J . (19741 Epidemiologic survey of leukemia in persons polentially exposed to benzene. J . Occirp. Jlcld.. 16, 375-382. Tsai. S.P.. Wen. C.P.. Wziss. N.S.. U'ong. 0.. McCellan. W . A . & Gibson. R . L . ( 1983) Retrospective mortality and medical surveillance studies of workers in benzene areas of refineries. J . O c w p . .Wed.. 25, 655-691 Van Raalte. H.G.S.G.rasso P. & Ir\.int. D. (19841 Tackling a very difficult problem. Letter to the Editor. Risk .-lnol..#. 1-8 Vianna. N .J . & Polan. A . ( 1979) Lymphomas and occupaiional benzene exposure. Lutic.cr. 1394-1395 Vigliani. E.C. ( 1976) Leukemia associated wirh benzene exposure. ..Inti. .V. Y . .-lcad. Sc.i.. 2-1, 143-51 Vigliani. E.C. & Saiia. G. (1964) Benjlene and leukemia. .Yew Eug. J . .\led.. 2-1. 371-876 Watanabe. T.. Endo. .A.. Kato. Y . . S h i m . S . . b'atanabe. T. & Ikeda. \I.(1980) Cyrogenetics and c ) tokinetics of cultured I! mphoc! tctj irom benzene-cyposrd it orkrrs. / t i ! . 4 r c . h . Oc,cvrp. Etirirmi. H~~uifi1i6.. 31 -41 N'hite. M ..Infante. P. & Chu. K . 19%').A q u a n ! ~ i ~ l i i veesiinute of leukemia rnortahr! ~ ~ s o c i a t eudi i h oiiupational r i p o w r e to heii/znr. /?/.SA . i t / d . , 2, 195-203 Wilcosli!. T . C . . ChrckoNay. H.. 3larshall. E.G. Jnd T!rolsr. H..\.(19841CJncer mortalit> Jnd solwnt exposures in the rubber industr). . . I t t / . ltrd. H . I . , ~ ..~.S.VOC J , 45. 809-8 I I Wolf. P. H . . .\nd)elko\ic*h. 0..Smith. .A & Tbroler. H . (1981I .A case-control ,iud! of leukemia in the U . S . rubber indu>:ry. J . O ~ c ~ r.p\ /.t ~ / . . 23, 103-105 N'onz. 0. ( I 387 I .-lu /iitiirsrr.r-it,itk* /jiorr'i//i.\. .rrirrii, ot c ~ h t v ~ ic/ /ci li.ork(>rs~ ~ ( . ( , / ~ p ~ ~ i / ~ / t / ~ ~ / / ~ c vpo.scd I( hvi:Lwe. Washin p o n . D C.. Chemica 1 M Ln~u ilict urers .I\sjc7cia [ion. December. 198-7 b ' H O 1 % 1 I E t i i ~ r r o r i t ~ ~ c ~H~ci foalr/li Crlrc,r/(i I<< .4rwrirc . Geneva. n ' o r l d Hc;llth Organimrlon. pp. 143-145 18 McMlCHAEL APPESDIX (added in proor) The extended follow-up of the NlOSH cohort t Rinsky et d..19851 has now &en published LRinskg et 01.. 1987)'. These data appear to provide a better bas: for quantitatirc nsk assessment than d o earlier studies. The authors estimated the cumul2ttit.e exposure 07 individual workers to benzene by use of a job- exposure mainx to link detailed job histories with indusinal hbgiene measurement (historical air sampling of ten7enc concentrations). The overall SMR for leukaemia ( 9 declthsi in benzene-exposed workers was 337 (95Ob conhdence intenal. lS&641). and that for multiple m>eloma (1deaths) uas 409 I 110-1047) For leukaemia. a docc-response relationship u a s evident. uiih SMRs increasing from 109 1 2uses1 [ o .3X ( 2 cases). I186 ( 2 cases) and 6637 i 3 cases). uith increases in cumulative ben7ene exposure from less than 40 ppm-gears. to 40-199. 200-399 and 400 or more. respectitel!, (400 ppm-)eurs IS equivalrnr to a mean annual exposure of 10 ppm over 3 I0-?ear working lifetimet The authors concluded that an exponential funciion best descnbed the dosesresponse relationship. and that a substantial decrease in the risk for death from leukaemia could therefore be achiebed by louerin: xcupaiional exposure to benzene. including (importantly) exposure within the ranse 10 ppm-0.1 pprn R i n A k R 4 . Smith. 4 B Homung. R . Filloon T G . 1 o u n g R J .Okun. 4 H s( L m d r i g m P J ( 1%-1 Bcnzrnr m d \CUkdCmid dn epidcmiolo!!ic nsk dsscssmcnt \til t n e l f I f d 310 Io44-ioio WORLD HEALTH ORGANIZATION INTERNATIONAL AGENCY FOR RESEARCH ON CANCER- RC ENVIRONMENTAL in !ffJ CARCINOGENS in METHODS OF ANALYSIS ms AND EXPOSURE MEASUREMENT I:! VOLUME 10 - Benzene and Alkylated Benzenes EDITORS L. FISHBEIN & I.K. O'NEILL 0. 7s C C 1 I A R C !ken(& Publimuons No. 85 INTERNATIONAL A G E N C Y FOR RESL4RCH 01 C.\\CER LYON I988