Document MGoYBgeB88V8B8exZMXm2Rr8V

it British Journal Of Industrial Medicine 1986;43:68569I -. ., .. ~ . . . .. . . ,_. .. An update of mortality among chemical workers' exposed to benzene G G BOND, E A M c L A R E N . C L B A L D W I N , R R COOK From Daw Chemicol IISA. Epidemiology. H & ES. Midland, Michigan 48674. U S A 4' . / _. " ABSTRACT Mortality was updated to the end of l9e2 for 594 employees cxposcd to benzene who had been studied previously and for an additional 362 exposed workers not studied previously. Cause specific mortality comparisons were made using United States white male, age, and calendar year adjusted rates. Total mortality was observed to have k e n significantly k l o w expectation, and this was particularly evident for deaths from accidental cau3cs. Mortality from skin cancer was significantly raised, although there were no unusual or common characteristics among the affccted individuals which would suggest a link with exposure to benzene. A non-significant excess of t d deaths from leukaemia was noted based on four observed cases; however, all four were myelogenous leukaemias and this represented a significant excess in that subcategory. These and other deaths of possible interest are reviewed in detail. Analyses by work area, duration of exposure, and cumulative dose index did not show patterns suggestive of a causal association between exposure to benzene and any particular cause of death. There is a large amount of evidence to support the hypo&& that acute myelocytic leukaemia may be a latent effect of overexposure to benzene' (B D Gold- stein, unpublished data). Data to support increased risks of other types of leukaemia or other cancers have 'been nponed, but arc in nced of further evaluation' (0Wong, unpublished data). The exact shape of the dose response curve for the knrcne kukaemia asociation is unknown and has k e n the source for spculation and controveny.'" One of the few studies which has been relied on for risk assessment was reported by Ott et al in 1978.' They followed 594 Dow, Michigan Division. cmployecr potentially exposed in any of three production a m using k n m e . Two deaths from leukaemia were observed compared with one expected to the end of 1973. Afler excluding cohort members with other hazardous exposure, they observed OM death from leukaunia against 0.9 expected. A third subject died of bilateral bronchopneumonia but had myeloblastic kukaemia listed on the death o c r t i b t c under "other significant conditions."This death was coded to pneumonia consistent with the rules follomd by the World Health Organisation.The present study updates the mortality experience of that cohort of 594 with an additional nine years of follow up to the end of 1982, and includes 362 exposed employes not studied originally. MctboQrod matedab A detailed description of the chlorobcnzol, alkyl k n =ne, and ethyl allulose operations in which thex employees had potential exposure to benzene was given by 011er d 6That reponalso should be consulted for summariesof industrial hygiene measurements used in exposurc classification, and the methods by which the cohort was identified. In brief, 594 employas were identified from annual census lists aa having worked for one month or longer ktwecn 1938 and 1970 at job, of interest in the three production areas. For the pmcnt study, the ansus lists w m consultcd again. This included a review of monthly census liru availabk beginning in 1965. These had not been used for the original study. An additional 228 employees who met the cohort definition of Ott cr aP were identified by this rc-vl.ew and are included in the present investigation. Monthly CCIISUS liru from 1970to 1978 wm consultcd to W r tain another 134 employees potentially exposed aner the closing dnrc of the original study. Thus the present study reports mortality among 956 employas. All job assignments icvolving potential exposure to benzene to the end of 1978 were coded and considered 685 * ... . . .. .. ~ . . . .. .. . . . 1- I. .' P ........ ....... .... .'-. . . . . .: ......... _- .. . I -I ,I i i i i ! -i . . I~... .. , ij I ., . .... ... . . . /: 686 for exposure classiftation. Each job entry was assigned one of four exposure intensity levels (very low, low, moderate, and high) b a d on availabk industrial hygiene monitoring data.All jobsheld aner 1972 were classified as either low or very low exposure. A cumulative dose index was calculated for each cohon member by multiplying the representative TWA value for each intensity level by the number of days exposed at each kvel, and dividing by 30.41 dayslmonth to yield ppm months. As in the original paper of 011 cf u I ,t~he 'IWA value of I ppm was used for the very low level, 5 ppm for low level. 17 ppm for moderate kvel, and 30 ppm for high level. Follow UP of the cohort for vital statu determination was done ruing company records, the camings and benefits records of the Social Security Adminisvation to the end of 1979, and the National Death Index to the end of 1982. Death certificates were obtained and were d e d according to the Inter- national Classification of Di- (ICD) NICS in effect at the time of death by a nosologist experienced in multiple rcvision coding. Deaths were then tabulated by cause for analysis using the categories of the Sth ICD revision. Analysa were carried out by production a m , duration of exposure, a d by cumulative dose index. Employes may have contributed penon-years ar.d deaths to several categories. Allowance was made for a pouibk latency period by lagging uposurcs IS yean. Expected d a b were calculated by the indirect method ruing United S U t a white mak mortality rates via an updated venion of the hionson p m gramme (USDR 58).' S W r d i s a i mortality ratios (SMRs) w m alculated u the ntio of otmmed to expected deaths multiplied by 100, and Fisher's exact 95% confidence Limits a b u t the point cstimam' were computed wing thc programmer of Rothman and Boicc.p Examination of m d s across SMRs wu done using a zz trmd (at with one degree of f d o m as outlined by B d o w CI d.Io Previous studies at this plant had identified em- ployas with p u t exporun to mmic," ubestos," or high levels of vinyl chloride1' u possibly k i n g at Bond, M c h t e n . Baldwin. Cook incmstJ risk for sew XI malignancia. Analyses were made both including and excluding those employees from the present cohort. 6 RcrPlb Table I shows the vital and employment rtatus distribution of the cohort on 31 Dccember 1982. Follow up wes 98.4% complete; the I f employees lost to follow up were withdrawn from the life table on the date they wen last known to have been alive. A total of 225 deathr was identified, 25 of which were among the 68 employes who also had been exposed to askstos, . arsenic or vinyl chloride. Death certificates were obtained for all but one (0.4%) of the decedents; this death of unknown cause was included in the totals for analysis but was not allocated to any cause specific category. The distribution of employees by period 6f first exposure and duration of exposure is pmcnted in table 2. The median year of first exposure was 1948, and 784% of cohort rnanben were first exposed at least I5 years before the closing date of the study. The mean duration of exposure was 7.0 yean (median = 2.6). The 936 employees contributed a total of 24571 penon yean, for an average follow up of 25.7 years. Theaverage age at entry into follow up was 31.0yean. Table 3 presents the dislribution ofcohort members by exposure intensity level, and within each level by year of first exposure and duration of exposure. Improvements in exposure conditions over time and job mobility by employas resulted in most having been exposed at more than one intensity level. Nearly a quarter (229 subjects) of the cohort members were exposed in jobs categorid u 30 ppm l W A benzene exposure, and nearly a third (311) were exposed in jobs categorixd as 17 ppm TWA. Table 4 shows the overall mortality comparisons for rlated causa of death. The all causes SMR for this cohort wu 84 (or 83 if the arsenic, asbestos, and vinyl chloride exposed subset is excluded). This rcprcsented a significant deficit m m M d with expectation but was consistent with !';e general pattern of lower mortality among WC~IO~CU from this plant." A ' i 9% 888 IW I92 n3 146 15s 13s 3m31 31s 6J 246 233 IS 1: ... .. .I ~- . .. . .. . . - .. . . .. . ... .- . ...'. -. . .... ... . ..-.. ... . . . An updote of mortality am- chemical workers eXpaKd lo k e n e - 687 held lrved opcntor orrigmenu with potentid high level ~ Z U Wexposure until he ntired in 1962. His atimrrsd career dosap was 4211 ppm m o n t h He died aged 80 of myclomonocytic kukrmu'r; no nee row wu prfonned. The i n t m r l bttwsm the ontet 1w Ip10-19 1930-SP 196069 6 39 IM It1 35 45 72 53 ~ 37 55 41 76 3x U do 4 137 3Sl 141 I89 of d h U 8nd d C d l W B I -nod to hrvc kta two months. Cuc number S w u not included in the originalOtt cohort. He began work with thecompany in 1928and 19mn 46 74 I8 0 I38 worked for one month in an 8 r u of potmthl ancnic Told 265 333 ZW IS 936 exporurr before moving to chlombrnzol wherc he worked flnt u an oprotor then u 8 thin forunm in jobs clurifkd 01 low or very low benotm exporurr until hi8 rrtirrmcnt in 1952. His estimatedcareer dor- age was 343 ppm months. He did in I967 a p d 80 of significant cxccu w u observed for skin caner bued acute myelomout kukacmia. on four o k r v d deaths. A substantial and significant As w u tnw for the three deaths from kukocmia & k i t was observd for rhc cotegory of rcciQnul obsewal by 011 et a/, both the new k u k u m i u had deaths. No other significant m u l b wrc found for thc the potential for unquantifisd brief. but potentidly other 54 causa of death evrlurtaj. relatively high, exporuru to bcnzrne. In the alkyl btn- Four deaths were OMfrom kukumio. m c a r a , kntmekwlswm marunduhigh as283 resulting in a non-significant QCCU for this ategoy. ppm in some u m p k in ethyl dlukme, u hi& u 937 The QLC summaries for h e x four dab lad for P pprn. Therefore. aution must bt cxcrcirad whm Anh case clouifkd to pneumonia u1given in tabk 5. interpreting the cumulative do= indica. Cuc n u m b 1-3 o r i g i d l y rcpomi by 011et There wu 8 non-rignifkant CXCCII of deaths a/.* Cuc number 4 died in 1980 rRn tbe original OMin theategory of o h and unrpecitidd a n - foilow ended. This mur began work at the oom- om.A micW of the deathccrtiAcltedirgntma for the p n y in 1943 u an o p n t o r in S a m fabriation. seven'Rr individuals did not show 8 mmmon shortly thtrrrRcrhcmovad to alkyl knzlcntw&rc he tumour type nor any that were of lymphtic tirrw. . nuVaY roul YI w 53 U 49 46s 1940 1910-9 4 3 6 16 J I I3 IO I I IS I 3 4 -1930-9 > 1-9 I9m 4s I2 33 5 76 aD 6 27 I2 II IS 5 10 8 84 83 1I3 n - -T d < I940 41m6 SJ 2 IW-9 IU I6 1 9 S 9 38 6 > 1-9 1910 II7 21 io1 IO 9a 8 26 I6 21 50 n7 12 9 w IS 27 -9 3 733 60 2% 1in01 I39 . . .. .. ....... ......-,,... ,. , ,. ...,.. .. . . . ... ... .. . . . . ..._. ... , .. . . .. ... . .. . . . . .. ... .. ,. .. ...... .... .... .... . . . . . ... . . . . . . . . .. .. . . . . . . . . . . ... . .. -.. 6aa Bond. M c h e n . BCJdwin. Cook l2S 268.6 J9 31-7 2 1-11 13 15.1 0 1-3 3 2.8 7 44 0 1.7 0 1.1 2 3.0 20 19.1 0 0.8 ia 18.1 0 03 4 1.0 3 3.2 0 0-3 I 1.5 I 1.4 0 0.1 I ' 1.7 0 0.1 2 1.1 0 0-7 4 2.1 I 1-3 7 34 4 3.8 3 e7 I15 1347 13 15.5 2 1.9 5 7.3 4 I:. 7-5 82- I39 13401 46-147 0-284 22-3 I 3 59-3m 0-217 0-335 8-24 I 64-162 Mi 59-I 57 0-1230 loel024 I P-274 0-1130 2-37 I 2-39a 0-3699 1-328 0-3689 22-657 0-527 J24 2429 7e3w Ip-zrn U-1151 70-101 4 sI43 13-UO 22-Id0 6-n *a6 200 242.3 49 499.2 2 I4 12 13.6 0 Ia 2 3 2.5 6 4.4 0 1.J 0 1.0 2 2.7 14 17.2 0 0-8 I 3 16-3 0 0-2 4 0.9 2 24 0 0.3 0 I .4 I 1-3 00 I I4 0 0.1 2 I .o 0 0.6 3 1.9 I 1.2 1 3.4 4 3.4 I 06 I06 121.4 I I 13.9 2 1.7 5 64 4 17.2 7 5.1 a3 100 I24 -a9 I18 --I 38 75 -81 -bo 441 --70 -80 -64 -199 162 a4 205 I18 164 a7 79 I18 76 23 I29 12-95 76132 1-51 &in 0-307 2s-ISl 50-297 0-246 0-369 9-268 45- I37 0461 42-136 0-1845 i21-113a 9-258 0- I230 0-164 -2 4 2 9 2-Ma 0-3689 26122 0-615 3341 24 8WU 32-301 4-929 71-106 40-142 14435 2sI77 64 49-149 , In addition to the death, from rplutk and pmi- dassikd 8s potentially uposcd to several different aous anaemia (omuch) obsemd by 011er d,'OM levelsof benzeneincluding hi& until his n t i m c n t in o t h e r d a t h o o c u n s d f r o m ~ o f M o o d o r b l o o d 1965. His atimated aeoct dose was 6027 ppm forming hut. This mm was h i d in 1923 and month. He dicd in 1980 a@ 80 of myelofibrosis. w o r k d in many diffmt pbr indudin#an area of 'Ibe analysis of morulity in the total cohort was potential ancnic cxpowrr until 1939wkn hc s t t k d carrid out dowing for a IS year minimum latency into various opcntot p b r in alkyl benzene. He was &ad. This effeaively m o v e d tttc-. 5nt cxposcd .. .. , .- -I d. ,. . .. . ... . .. :. . . . . alter 1967 fmm the rarlydr.Tbc mortality pttrm, obwmd for tho chloroknroi urd ethyl a ~ o e dauibcd above wem not rltucd, rlthougb aome Of UIUI. the caw cpscific SMRa were W rli#htly. Table7ptacntr ObwrVed lad expaxed dacha fmm Morulity results by work ua uc prrwatcd in rllanccricombid by yara ofexporuemd~nmvd table 6. Some mrploycu worked in rnultfpk mu, s i n e 6nt exposure. Morulity w u romewhrtm i din therefore they m y be countai more thrn once in t& thc period from IS to 19 yan after &tt upoj~n; table.Totalmortalitywuoonristcntlyblow expactsd however, it wu h @ y indcpadcnt of duntion of .. - lewis t a w the three work uaa. Although tbac mpo8urc. were no statistially dgnif&antrim by work area, Tabk 8 show the resultsof dose nrponae analyses marginal ex- of skin cancer were o k m d only for r c W aw of dah ategoria U y x a wrc for the alkyl kntcnc md ethyl Odlulors a m & lad mrdcbyflntIr~ngexpoturracroyeur,mdtbcn15 ex- lymphopoieticanccr wuobumd only forthe y a n to allow for tbe porribk e x p d o n of a latency alkyl benome a m . A slight mcua of lung oocctWU period. Duntion of expoaure is m important corn- obscmed among t h a who worked in ethyl c d l b , poncnt of m y cumulrtive dorc index md is ruongiy w b e m desciu of this cancer wae okcrred for tk conclrtai with survival. Failure to lag expoaura to other two production arcas. Smrll cxassa in tbc a t - compcnvtc for thi8 m y rudt in miuing 8 d o e .. . e p r y of cancer of other and uupiikd ata wtlt =pone relation. There were no signilkant mortality .. .. ._ 1 .... ...... .: - . . .. . . . . .. .... .. .. .. . . . . . . . . . . . . . . . . . .. . . . -. . . . . . . ....... ~ .. ... . .. . .. . ... __*__ .- 690 Tablo 8 Morrdfty -y expMvrtt by esthrotrd &atlw Bond. M c h r e n . Baldwin, Cook dose of benzm expwnre. exrludhg employees with competing All aua (001-9W) (w99 ppm moorha ~999ppmmonlh lomppnmoothr Toul All a- (140-m)) 0499 115.' 43 42 30 27 IS14 35.2 55.6 212.3 30.0 7.3 11.9 499.2 9-7 2-5 4. I 163 0.6 0.1 0.2 0.9 39 07 1.1 68 1.2 03 0.4 1.9 746 17.9 21.9 1214 76 122 76 83. 90 I.S-0. 92 100 62 204 49 Bo --6763 441 -133 ta3 12s -I67 uo I58 sa 123 83 87 IS yra?s Oh Ios 33 17 165 17 9 7 43 8 4 I 13 4 0 0 4 3 0 2 5 2 0 3 53 21 IJ 89 ~~ GP SMR I 174 25.7 374 180.5 09 129 72 91 25.0 I08 -5 7. I s9 0 4 84 39-I I10 8-6 1.0 3.0 13.S 04 0-I 0. I 0.6 -2.3 132 0.5 b,0.7 214 3.5 142 -0.9 222 0.2 0.3 333 I 4 214 62.6 13.7 Ian5 Dl 75 %a3 92 trends for m y awe of deatb miurted. Oddly enough, mortrlity from awed awes w u somewhat mired among those c x p o d from u)o to 9!l9 pprn months. AU four &in ~ l l c mwm o k n d to have km in the lowst dose awry. Dhrrk. cxpoctcd. Taken together with the two deaths from kukrrmir previously reported for this cohort, the total still doer not represent a statistically significant -CXCCII of toul k u k d a . It docs represent, however, a riguiffant excess of mortality from myelogenous kuhmia(4obrcnsdHnur0.9expected;p 0911) -whcn mortality data compikd from the Surveillance, Epidemiology and End Ruulu programme arc used as a compujlon.t* t 'lk datb from myelofibrosis is intcruting since thisconditionis bciieved to represent a primary over- growth of a ~ V fibCroblastic a l l line, analogous to mydomo~ I' There have been stvera~ nporuof myelotlbrorir among anployetr with alleg- edly lmvy bcllEcM exporun (B D Goldstein. unpublisheddata). Among thedeathsfrom other lym- phoreciculu ancer wu one multipk myeloma. This did not rtprrrcnt an ex- over expaxation and is nocrbk only because Dscou& er d suggested a link be- thh condition m n d knomc.' An unexpected finding w u a signrkanr CAW of . .. . . . . . . . .... .. 7. . ... .. . . . . .. .. .. . . . . . . . . . . . . . - .. .. -. . - ... ... .... ,* . c An update of mortalily m o n g chemical workers exposed to benzene 691 mortality from skin cancer, bared on four o&urvod a small number of events, ampcting exposures to cases. A review of the death artiflater : h o d that other @tentially hazardour materiais, and the unar- ' thm of the four were malignant melanomas, the tain contribution of unquantified brief exposum. fourth was described BI a squamous a l l arcinoma. There were no unusual or common charactcrirtia We gratefully acknowledge the technical support among the affected individual's work historia that received from Dr M Gerald Ott, Karen Wymcr, Jan- would suggest a link with benunc or any other work- ice Cartmill, and Bernard h i c h . Angie Grzcgomyk place exposum. All four were classified into the low- typed the manuscript. a t cumulative dose category. %our repom of mortality among benzene exposed workers have not noted exass skin canctr, or more specifically mela- noma, and thus its finding here may represent ~JI isolated or chance occurrence'-* (B D Goldstein and 0 Wong. unpublished data). Mortality from accidental causes of death was obscerved to have becn substantially below expectation. even by indirect comparison with other employees from this There is no availobk explanation for this deficit but a@n it may rcprrrcnt a chance arurrena. Analysis by work area duration of exposure, and cumulative dose index did not show p t t r n u sug- gcstivc of a causal d a t i o n ktwscn expourn! to b e m e and any p o r t i ~ ual ~t g o r y of dath. Three of the four patimuwith kukaemir, OM with multipk myclomqmd one with mydoflbrosisworkod in alkyl benzcne. This may indicate that lorn additional CQ- factor may be imponant in the actioloa of that deaths. Other materids prwcnt in this work UQ included ethylknzmc, divinylknom, tolume, xylene, isopropyl knamc, styrene, dbopropyl bcn- m e , a d alpha-methyl styme. I Askoy reccntly suggested a link between exposure to benzene and lungc a w . * We found no evidcucz to support such an auocirtion in out study. LungM c c r was not found to be excessive in tbe toul cobonor in I the cohort krs exclusion^ (13 OM@ntt 16.3 expected), nor was thm a tmd of irrraring lung a n a r risk with incrusing exposure. Tbt s t d y hd 95% p o w to d e w I twofold ririr d l u g with a one uilcd t a t 8d dph8 = 005, rad 57% power to detect a om and a half-fold risk. Although this study provida supper\ lor tbe UIO. ciation k t w a n exposure to md acute myelogenous kukaanir report& by otkra' (B D Goldstein), wen1 factors obmplhte w of tkrc data for risk aucsrmmt. Thcrtincludetheprobkmof ,