Document M2Dk2ZOyRqE0pxom6J0BZkkj

Richardson L (1996): Self-asses= exposures.Am I Epidemiol 144321 iper H, B6gn D (1985): Obmnin niologc casecontrol studies J J. G6nn M (1986): Interwater agreeme p r e in a case-conuol study Br J Ind Facton for Cancer in the Workplace." ire assessment UI commumty-based stu tygene 3.41-58. ' CmperJ (1981) Discovenngcdrcino III A novel epidemiologic approdch !ixdson L (1989).Costs and statibti dt collecungarcupauon exposure in mol studies.Am J Epidemol I 3 0 I Richardson L, Dewar R. Gtnn M ie occupauonal environment. Method uqe case-referent momtonng system. 492 1988): Vanance and dissent A rea :pidemiol41 94'+958. A M E R I C A N JOURNAL OF I N D U S T R I A L M E D I C I N E 3 1 :287-295 (1 997) Review of Epidemiologic Evidence on Benzene and Lymphatic and Hematopoietic Cancers David A. Savitz, PhD,` and Kurtis W. Andrews. Pho Exposure to bencene is generally accepted as a cause of acute myeloid leukemia (AML),but the association with other cell types of leukemia and other lympharic and hematopoietic cancers is connoversiai. We compiled epidemiologic research on benzene and lymphaticand hematopoietic cancers in order to assess the panern of associations. Eighteen relevant community-based and 16 industry-based studies were located. Four of seven smdies of lymphatic and hematopoietic cancer in the aggregate identifed relative risks 41.8 or more, and eight of 14 total leukemia studies yielded rehtive risks in that range. Thefew available studies of specific histologic types of leukemia do not indicate larger or more consistent elevations in risk for AML compared to other leukemia cell types. Sporadic repons have linked benzene to non-Hodgkin's lymphoma and multiple myeloma. but mosr studies do not indicate a posime association. Limitations in s d y quality. particularly exposure assessment, pervade all of the studies reviewed, and the distinction between studies addressing benzene and those addressing jobs in industries that use benzene is somewhat arbitra?. Nonetheless, the epidemiologic evidence linking benzene ro leukemia in the aggresate, as well as for subtypes other than AML is no less persuasive than that for AML alone. Am-I. Ind. Med. 31~287-295.1997. 0 1997 Wdey-Liss. Inc. KEY WORDS:benzene; leukemia: lymphoma INTRODUCTION Benzene has been suspecrzd as an occupational carcino;zn since the 1920s [International Agency for Research on f-ncer, 19821. In 1981, the evidence that benzene is -:Tinogenic in experimental animals was characterized by .`:I International Agency for Research on Cancer [1982] as --..mired,"and the evidence that benzene is carcinogenic to .:'znans was found to be "sufficient." In part because the s i c e r s of primary concern. lymphatic and hematopoietic xmplasms, are rare, case series documenting "larger than :=petted" numbers of such cancers among shoe workers in - ? n m e n t of Epidemiology, Schooi of Public Health, University of North 3: - na. Chapel Hill, North Carolina. xmt Grant sponsor: the NationalInstituteof Environmental HealthSwences --- x g ; Contract Grant number: T32-ES07018. `-3-respondence to: David A. Savitz. PhD, Department of Epidemiology, CB =7G3. University of North Carolina. Chapel Hill. NC 27599. E-mail: :AID-SAVITZ@UNC.EDU k c a t e 0 24 June 1996 Italy [Vigliani & Saita. 19641 and Turkey [Aksoy et al.. 19741 provided important early information regarding increased risk. Although such observations lend general support to the notion that more cases than expected have occurred, in the absence of information on the population at risk and disease rates in some comparison population, the magnitude of association between benzene exposure and lymphatic and hematopoietic cancers cannot be assessed. Subsequent studies of benzene and cancer have attempted to measure exposure F d disease risk directly in order to derive dose-response relations. Although controversy remains about the exact shape of the function relating benzene exposure to acute myeloid leukemia (AML) [Austin et al.. 1980; Swaen and Meijers. 1989: Paustenbach et al., 19941, with particular unceninty regarding risks in the low-dose range (around 10 ppm). it is generally accepted that benzene exposure in human populations is causally related to the development o f . W L . Most of the epidemiologic studies include information on several types of leukemias and often on other types of cancers as well. In fact. many of the early studies found the strongest evidence linking benzene to total leukemia rather number of such cases [Vigliani and Saita, 19 [Vianna & Polan, 19791 and multiple myeloma [Rinsky et al., 19871. The possibility of a broader array of lymphatic and hematopoietic cancers resulting from benzene has been suggested by Goldstein. based largely on toxicologic data [Goldstein, 1988; Goidstein, 1990;Hricko, 19941. Given a substantial number of epidemiologic studies 19891but we are less confidentthat all relevant non language publications were found. When multiple publications were produced from same study population, we tabulated results from the complete or up-to-date version of the study [Rinsky et 1987; Paxton et al., 1994; Wong, 19951. When re cer, most of which include cancers other than AML, we undertook a systematic effort to identify and compile the results in order to evaluate the epidemiologic evidence linking benzene toa range ofspecific types oflymphaticand hematopoietic cancers. MATERIALS AND METHODS benzene, we selected the most detailed report for in in the tables. In one instance, the more recent study`s are not yet complete, SO that we include parts of the e results [Yin et al., 19891, which are a subset of the results [Travis et al., 19941. The major division in the studies was between thos were conducted in a community population versus thos is known to be present (e.g., oil refineries, rubber mmufac- determine that benzene was present, and "high" meant that aggregate unless a subset of agents or workers was examined that directly addressed benzene. However, studies were not excluded for evaluating benzene as part of a relatively Some reports examined multiple groups delin level of exposure, certainty of exposure, duration of sure, or latency, which we included in the tabulations. results were available with and without adjustment criteria but d e r on the patterns of results, emphasizing the calculated using appropriate methods [Rothman and Boice magnitudes of relative risk estimates across studies. A second criterion for inclusion was an explicitcompcui- compared to those not thought to be exposed. usually and Human Services, 19801 to group cancers regardles generating a measure of relative risk (standardizedmortality the original classification scheme used by investigators. ____ -etc;). -Wedscr examined results for all lymphatic and hematopoietic Benzene and Lymphatic/Hematopoietic Canc TABLE 1.-Description of Community-bid Studies Reporting on Benzeneand Lymphatic and HematopoieticCancers Rstauthor bsarl satlhrg Exposum datasowce assigrmart Qualityof exposunt assessment controlfor COnfoUnders I (spard(1970) France In$nriew Self-report LOW Age, sex Ishimru(197l) Hiroshima8Nagaski interview Job title LOW Age, sex, city, distance from A-bomb llianna (1979) New York state Deathcertificate Job title LOW None Lines (1980) OlmstedCounty, Medicalrecords JobWe LOW Age, sex, residence, dura MinnesoOl tion of medical cowage Rodin (1986) Sweden Interview Self-report High None Linet [CUI (1987a) Baltimore Interview Self-report High Age, sex, race Linet [MM] (1987b) Baltimore lntecview self-report High Age, sex, year, hospital Schumacher (1988) North Camlina Death certificate Job title High Age, race, year Malone (1989) US (multicenter) Interview Self-report High Age, sex, race, residence, education Vai (1989) Crane (1989) ltaly Houston Hospil record Clinical evaluation High Age, sex, year lntenriew Self-report High None Siemiatycki(1991) Richardson (1992) - _~. Montreal Paris lntecview Interview Job description Job description High Age, income, smoking High Age, sex, ethnicity, resi- dence, hospital, prior . cancer therapy W e r r (1992) Boston Interview Self-report High Age, sex, residence Blair (1992) Iowa& Minnesota Interview Job We High Age, sex, state, year. edu- cation, pesticides, . family cancer history, hair dyes, directlswro- gate respondent Cran@ (1992) Houston Interview Job We LOW Age, sex, prior cancer therapy Heineman(1992) Denmark Tax records Job We High Age Cicwne (1993) Torino, ltaly Interview Job We High Age, sex. smoking m e overlap with subjects in Craneet al. (19139). I mias divided by chronicity and cell type (acute and chronic, but not all. studies adjusted relative risks for age, and where Iyrnphocydc and myeloid). appropriate for sex and race. Few studies included a broader array of potential confounders related to lifestyle and none attempted to incorporate information on other occupational RESULTS workplace exposures. Industry-based studies are concentrated in the chemical, Eighteen community-based (Table I) and 16 industry- oil. rubber, and shoe manufacturing sectors (Table 11),all of based (Table II) studies provided data on benzene exposure which classified exposure based on industrial-hygienist md lymphatic and hematopoietic cancers. Community- jud-ments. sometimes combined with measurements. Most based studies identified exposure from self-report of expo- approaches to exposure assessment were considered to be of w e or job title, with each having advantages and disadvan- high quality. Consideration of confounding was largely .ges (Table I). Only four studies integrated self-report,job limited to age and calendar time. - .~...I .- i t t l t , and expert evaluation [Siemiatycki,1991;Richardson Four of seven studies of lymphatic and hematopoietic et a E 3 9 9 Z Heinemarr errrf39!92+ G h z h is optima1 in such studies. iieneeOetleaertfaieLs-LtsWtn3b.3r,e-s-cdCa nHc5e5reshiat kEae&g gweihgbaetnezesnheo-wediwo-foldGrW-atemei'%thtteh-&e --- - ~ _ 2 S a i et al. I19831 study largely uninformativeduem small -.-- -.- i 7 ! i i i II I 1 x- "+;.. i L. -TABLE W. Descriptionof Industry-basedStudies Reporting on Benzeneand Lymphatic and HematopoieticCancers " - (lurlitrd axpoars mif a confoundaa , *... ,.. Aksoy (1974) Istanbul, Turkey Shoe manufacturing Lovr None Thorpe (1974) Ewope OiVpetrochemical Hioh Age Rushbon(1981) Tsai (19B3) UnitedKingdom TeXaS OiVpetmchemical OiVpetmchemical \ Hi$ High Age, duration of work year hired Age, sex, year, race OecWfle(l983) us `checkoWay(1W us Chemical RUbber High Age, year High Age Bond Mkhgan Chemical Hsh Age, year Austin (1986) Wong (1987% b) Ulinois us OiVpetmchemical Hgh Age Chemical Hsh Age, Year -Yin (19E9) ott (1989) Hutley (1991) AJW(1992) (1994) wens (1995) Greenland (1994) Tram (1994) China us United Kingdom Australia us Massachusetts China All industries Chemical Coke and coal OiVpemchemical Pliofilm (rubber) Transformer assembly Multiple industnes High Moderate Moderate Hgh High High High Age None Age, year Age Age, year Age, year of hire, year of death Age, sex 4 ,1 7 1 L Albb- AIPHS,AustralianInsMuteof PetroleumHealth SutvallanceProgram TABLE 111. Results of Studies Reportingon Benzene and Total Lymphaticand Hematopoietic Cancers Firstauthor (year) study Mng Exposed cases Relative risk 95% ci Vai (1989) Tsai (1983) Decouflb (1983) Bond (1986) Wong (1987a) AlPHS (1992) Travis (1994) Communlty Industry Indusby Industry IndusEry industry Indusliy 15 0 4 7 308 8b 1OC 76 a2 13.3 8.0-22.2 0.0 0.0-3.3 3.8 1.1-10.2 1.3 OS-2.8 1.3 0.9-1.8 1.o 0.5-2.1 1.9 0.7-4.6 2.3 0.8-6.5 3.4 1.9-6.1 1Contlnuousbenzene exposue: ~ntermiltenbt anzene exposmz -.(Mediumversuslowbenzenearposve. dHighversuslow benzene ~ ~-- _ _ _ _ _ _ - -~ numbers. Although the comparison between benzene- exposed md the genenl-population in Wong's [1987al study provided evidence against an association with TABLE IV. Resultsof Studies Reporting on Benzeneand Total Leukemia Firstauthor 0 studv satting m- CBSBS risk Girard (1970p Communtiy 30 2.9 lshimam (1971P Community 24 1.8 Urn (1980) Communtiy 4 3.3 shorpe(1974) hdlw 8 1.2 Rushton(1981) hdusby 18 2.3 Tw (1983) hdusby 0 0.0 clecCm(1!383) hdusby 3 6.8 Bond(1986) hdusby 4 1.9 -Ausbn (1986) Wong (1987a) Yin (1989) Hwley (1991) hdusby hdustry hdusby hdusby 15 NA F 1.4 ld 0.7 25 5.7 5 0.4 PaXm(1994) 14 3.6 rnland(1994) NA` 0.9 NAP 1.4 W%ceuUtee.rsCU. myQdcombned. -exdu(ed antinumemosue. A3 to exclude presumedx-ray rather than benzene ~xposUm++ -g TABLE V. ~esultsof ~tudikReporting on Benzenemi specific HistologicTypes af Leukemia Benzene and Lymphatic/Hematopoietic Cancer ---.r-y 29?' TABLEVL Results of Studies Reportingim Benzeneand HOn-Hodgkin's Lymphoma I Girard (19m) Community LSrnS 1 0.8 0.1-7.4 mustry 4 2 5 NA Vianna (1979) Community Ls 499 2 0 1.8-2.1, wustry 2 1.5 NA Schumacher (1 988) Community LYRS, Other LC 56b 0.8 0.5-1.1 Community 9 3.7 1.2-10.7 lff 0.9 0.4-2.1 Community 61p 1.4 0.4-2.2 Siemiatydd 43b 1.5 0.9-2.5 (1991) Community All NHL 266 0.7 O.tl.1 3lC 0.9 os-1.5 ff 0.8 0.3-1.9 Malone (1989) Community NA 1.l 0.6-2.0 Scherr (1992) Community All NHL N4 1.2 0.5-2.6 Mye'Gld Blair (1992) Community All NHL 153 1.1 0.9-1.4 gccone (1993) Community 9 1.7 0.65.5 Checkoway Bona WtW Industry 4 4.4 1.2-11.4 (1 984 Industry Ls 6 3.0 NA Ausnn (1986) Industry 15 NA NA Bond (1986) Industry WRS ott (1989) Industry 5 1.o NA Wong ( 1 W q lndustty WRS 2 1.8 0.2-6.6 3' 1.1 0.2-3.3 Acite 19 1.2 0.0-6.4 Gimd(1970) Community 17 .RichanSon (1992) Community 224 3.0 1.1-8.1 ott (1989) Industry All NHL 1.3 0.8-2.3 Greenland 5 1.0 NA ? ' Akso:i :974) Industry 15e 2 8 1.3-5.9 (1 994) Industry All NHL 26 2 2 1.4-3.2 NA 1.0 0.2-4.5 Acute Myemd Flodin 11986) Crane t1989) Community Community 0 6 .LS, lymptmsarmna: RS. reticulsarcoma; LC, tymphabc c a n m HHL mn-Hodglan's -0.0 lymphoma bwhltes. 95% Uestimated based on 90% CI 1.1 0.3-3.8 CBlackS, 9% a m a t e d based on 90% CI. ! Crane 11992) Wong :1!3!35) Community Industry 4 6 3.3 0.5-22.3 d A n y exposue, 95% CI esbmated based on 90% U 5.0 1.8-1 1.o QsubsDurbd erqow-e,95% CI estrmated basedon 90% CI. Conbnuasaxpawre. slntemrtterdmpJsure. ''BCaalsecdulactnedrdfrarmr eTt aaLbk(1498b0as)jeodb-oenxpsoeisfu-rreepomrtaetdnrexposw.e. Basea cn ?MOrmpabonal HazardsSurvey (Sieberetal., 199l)pb-exposuremab'1x 'Any exposre. I 'Medium or eqam?. internal cohort comparisons not calculable due to an absence of cases in the unexposed cohort. but dose-responseanalyses showing a concentration of risk in the most highly exposed suppon a positive association with relative risks in the range group Wong. 1987bl. For example, in the ,pup with over of 3-5 in the most highly exposed groups. Overall. given the 720 ppm-months of occupational benzene exposure, the diversity of methods and likely bias toward the null from S M R was 2.8 (95% confidence interval 0.6-8.1). based on nondifferential exposure misclassification. the replicated three observed cases. r!idence of increased risk is notable across a series of .Analysis of specific cell types of leukemia provides the relatively highquality industry-based studies. most direct suggestion of a general rather than specific Leukemia in the aggregate provides stronger evidence association of benzene with leukemia (Table V). Studies for an increased risk, with relative risks above 2 in two grouped leukemias in different ways, e.p.. ag-mgating all community-based studies [Girard and Revol, 1970; Linos et lymphatic or all acute leukemias. so that the number of al.. 19801 and four industry-based studies [Rushton and repons for individual cell types is limited.Also. imprecision Xlderson. 1981: Decouflt et al., 1983: Yin et al., 1989; is more problematic for specific histologic tvpes of leukemia Puton et al.. 1W4](Table IV). Two studies yielded relative than for leukemia in the aggregate. Both studies examining 1.1 k estimates of 1.5 to 2.0 [Ishimaru et al.. 1971: Bond et al.. lymphocytic leukemia and two of three studies addressing I '861. but several studies generated relative risk estimates chronic lymphocytic leukemia generated relative risks of 1.5 or more. Two of four studies of myeloid leukemia and acute 4I, & &-: - TABLE VII. Resdlsof Studies Reporting on Benzene and MultipleMyeloma RefWellC9 (Year) shrdy setting Exposed -58s Relatire rirk 95% CI Linet (1987b) Heineman (1992) Ott (1989) Wong (1995) Community Community Mushy mushy NA 137a 52b 5 4 12 0.4-3.6 1.3 1.O-1.6 0.8 0.6-1.1 1.4 NA 2.9 0.8-7.5 *Possibleexposue bPmbable exposue leukemia. The quality of studies is quite mixed, limited particularly for the community studies which contribute over half of the relative risk estimates, perhaps because it is difficult to a c m e a sufficient number of cases in industrybased studies to examine specific histologic types of leukemia. The overall pattern does not indicate m association limited to AMI- with the evidence for lym +iocytic leukemia similar to that for myeloid leukemia. Non-HodZAn's lymphoma was general1 not associated with an o v e d l elevated relative risk (Table ;I). with only three studies providing support for such J association [Vianna and Polan. 1979: Checkoway et al.. 984: Bond et al., 19861 and a number of well-designed. large studies failing to report increased risks [Schumachr and Delzell, 1988; Siemiatycki. 1991; Blair et al.. 19921. Multiple myeloma. first linked to benzer . in the report of two cases by DeCouflC et al. [I9831 .vas strongly associated with benzene exposure in the st^ -y of Pliofilm workers [Rinsky et al.. 1987; Wong, 19951 .n which four cases were observed versus 1.37expected ( T d e VU).Other studies [Linet et al.. 1987; Ott et al., 1989; tieineman et al., 19921 found only very weak associations (relative risks under 1.5). DISCUSSION The pattern of epidemiologic study resuls across broad and narrow ,goups of lymphatic and hematopoietic cancers does not lead to a conclusion that benzene is ;Icause of AML alone. Considering the strength and precisim of associations, the data are most supportive of a rather broad effect on lymphatic and hematopoietic cancers or leukemia in general. Since these diseases are subsets of one another, extremely strong associations for AML would have to influence the total leukemia results. which would. in turn. influence the poietic cancers [Young et al., 19811. Thus, although th findings reflect overlapping groups, they do not sim indicate an association with a single subset of cases. Won& [19951 recently highlighted the strong association betweex benzene exposure and AML in the Pliofilm cohort ( s ' ~ observed 1.19 expected, relative risk of 5.03). yet b 5 subtraction from the observed and expected numbers of totap leukemias [Paxton et al., 19941, the residual of leukemi& other than AML are also associated with benzene expos& (eight observed, 2.70 expected, relative risk of 2.96) [Savih and Andrews, 19961. The perception that the association between benzene and AML dominates the overall pattern 0 epidemiologic study results is not supported by our review. Restricting our analyses to epidemiologic studies excludes a number of lines of evidence. Extensive laboratory research [Goldstein, 19881 was not considered, even though such studies provide an important component of the overall assessment of the health effects of benzene. Case series on benzene and leukemia provide evidence both for specific associations with AML as well as other cell types of leukemia Fternational Agency for Research on Cancer, 19821, but were not included in the tabulations unless explicit comparison was made of risk among exposed and unexposed individuals. Even epidemiologicstudies of industries in which benzene is known to have been used were excluded unless benzene was examined explicitly. The extensive data from the petroleum industry [Delzell et al., 1988; Savie and LMoure, 19841and rubber industry [McMichael et al.. 19761 are potentially relevant to assessing the health r i s k associated with benzene, but without some restriction to the subsets of workers with benzene exposure the data are difficult to apply. We did not exclude any studies for failure to meet minimum quality criteria. Exposure assessment. particularly in community-based studies, is subject to substantial error and may well reach a point at which the authors' assertions that benzene is under investigation can be challenged. Inferences based on aggregations of job titles. without careful review by industrial hygienists [Vianna & Polan, 19791 or other corroboration of inferences regarding benzene exposure. are subject to a great deal of uncertainty. Self-reported exposure of cases and controls [Girard and Revol, 1970: Malone et al., 19891 may be superior to assessment by job title if the study subjects have sufficient understanding of their work environment. The optimal approach in community-based studies is the synthesis of self-report with expert evaluation [Siemiatycki. 199I ] or specialized questionnaires designed to elicit essential information for assessment of exposure [Stewart and Stewart. 19941. Few community-based studies attained that level of quality. %%ere cases and controls contribute subiective _the- potential for-biased r e p o r t i z or inferences based on job titles o r Benzene and Lymphatic/Hematopoietic Cancer - 293 more objective data, the errors are likely to be similar carcinogenic. Even in studies that were published, the for cases and controls and the resulting biastoward failing to decision to provide results specifically for benzene among. identify associations that are truly present an array ofagents of interest may have been influenced by Industry-based studies have used more objective expo- whether an association was found. One might expect such sure assessment methods, relying on blinded coding by bias to be strongest for A m , in which failure to find sxprts or inferring exposure based on job characteristics. associations is seen as aberrant. Some involvement with exposure assessment experts was In light of the research findingsreviewed here, sufficient dmost always involved, ensuring a higher minimum quality. uncertainty remains to encourage additional study with ?;onetheless, even with certainty that benzene is present in refined exposure assessment. Clarification of the pattern of ht: \vork environment and that some indibiduals are ex- benzene associations with specific leukemia and lymphoma posed. there is often substantial uncertainty regarding which cell types requires extremely large studies, so that cohort individual jobs involved exposure to benzene and the studies with sufficient statistical precision are required. If magnitude of that exposure. Quantitativeinferences relating based on mortality as the endpoint, such studies may be !eveis of benzene exposure to risk of cancer is impossible in lacking in sufficient diagnostic precision ro address this ~1 but a few of the epidemiologic studies reviewed [Pausten- issue. Community-based case-control studies. which have bach et al., 19941. Given the diversity of approaches to the potential for attaining adequate study size and accurate csposure assessment and their varying quality, the potential diamoses, are severely limited in their ability to assess for the intensity and accuracy of exposure messment.to exposure across diverse industries. Although a number of produce notably different risk estimates s M d be noted. agents have been effectively reconsrmcted by experts [Gerin Failure to isolate benzene from clojzly associated et al.. 19841, benzene poses particular challenges: It was agents that may themselves cause lympharicmd hematopoi- rarely used in pure form, but as one among several solvents; etic cancers could introduce confounding. In the rubber it is no longer used as a commercial solvent, so that industry. for example, associations betwen benzene and lymphocytic leukemia were reported [Ckckoway et al., IYS41.yet other solvents were as or more smngly associated riith lymphocytic leukemia, making isolation of the etio- present-day measurements provide no information relevant to past exposures; and the benzene levels of concern are relatively low. In spite of limitations, morcaliry follow-up of historically exposed cohorts should continue. with attention logic agent difficult. Environments assw<ated with the to identifying diagnostic subtypes of these cancers, as well petrochemical industry, shoe manufacturing. and painting as refined exposure assessment in commue--basedstudies. contain a wide array of organic solvents aod other poten- Even where exposure has terminated, continuedfollow-up is II tially carcinogenic chemicals which wax not fully ad- w a r r a n d to fully understand the long-termconsequencesof dressed in the published reports. In fact ao matter how benzene exposure. arefully job titles and work settings are examined, the X recently completed study in China may provide the I d5liry to isolate benzene from concomimc txposures will clearest evidence to date regarding benzene and lymphatic b? limited. and hematopoietic cancers [Yin et al.. 1991: Dosemeci et al., Dia-gnostic quality of lymphatic and hematopoietic 19943. Methods, but not detailed results. from this large cancers can be problematic, especially in part time periods cohort and nested case-control study of multiple factories and for deaths rather than incident cases. Diagnoses would have been described, and it promises to provide refined generally be superior in community-based studies in which exposure and cancer data on a cohort of nearly 80,000 incident cases are identified through hospital records or exposed men and women, as well as 36.000 unexposed tumor registries. Although death cemficam typically used workers w i n et al., 19941. Differences in leukemia cell type in industry-based studies, are accurate for broad groupings distributions between Caucasian and Chinese populations, l i k leukemia, they are subject to substantial error in making howeker. may affect the generalizability of their results, at inferencesabout specific histologic types or'leukemia [Percy least in regard to the types of leukemia affected by benzene rft ai.. 19811. Such misclassification across histologic types exposUR?. a.ould tend to dilute associations that occur only within a specific subtype of leukemia, and reduce pcision if many CONCLUSION cases are relegated to the "unspecified"category. The methods for identifying all relemnc research to be The epidemiologic evidence tinking benzene to leuke- i:icluded in this review are vulnerable to bias as well. The mia in the aggregate, as well as acute and chronic lympho- <'-cision of the authors to publish mic!es and include cytic and myeloid leukemia. is no less persuasive.than that .elected results may be affected by the pattern of those for X\lL alone. Specificityof association \vasonce cited as a . -. ... ... -. .. . . ~ ~ s . u ~ E ~ a ~ u a t l o n _ s _ o f . ~ b e nanzde ncaencer that. djd-not- 294 Savitz and Andrews - \ ahxeSyanalogy+totkr known tarcinogens,it -wouM 4XXa+X-F.A4eng-J-F-Zlw J-S.-Wachekler SKneller R, Bla%l-.F- . bPJurpnsingifbenzene-s only cafc_inogeniceffect was on Gokon study among workers exposed lobenzene in China: 11. E.*%assessment.Am J Ind Med 26:401--111. MIL. Limitations in the literature. particularly in regard to exposure assessment and the ability to isolate benzene from .-concomitant workplace exposures. preclude more definitive Flodin U. Frednkspon M. Axelson 0. Persson B, Hardell L ` Backgroundradiation, electrical w o k and some otherexposuresa,.. aith acute myeloid leul.emia in a --referent study Arch En\ iror F z. conclusions. 31:77-84 ACKNOWLEDGMENTS Gerin M, Siemiatycki J. Kemper H. Begin D (1985): Obtainin,0 (k-;L'tional exposure histones in epidemiologic case-control studies. J O::,: Med 27:420426. We would like to acknowledge the support provided by Dr. Herschel Hobson to initiate this project and the helpful review of the manuscript b! Dr. Andrew Olshan. Kurtis Andrews was supported in part by National Institute of Environmental Health Sciences Training Grant T32Es07018. REFERENCES Akmy M. Erdem S. DinCol G 1974):Leukemia in shoe-workers exposed chronically to benzene. Blood 44:837-84 1. Austin H. Cole P. McCrau-DS I 1986):A case-control study of leukemia at an oil refinery. J Occup Med 2S:I 169-1173. AusM H. Delzell E. Cole P I 19801:Benzene and leukemia:A review of the literamreand risk assessment. Am J Epidemiol 127:419439. Australian lnstitute of Pemleum Health Surveillance Program (19921: Health,Watch. Ninth Repon. Departmentof Public Health and Community Medicine, Universityof Melbourne. Melbourne.Australia. Girard R. Revol L (1970): La frkquence d'une exposition benkniqnt L cours des hhopathies -maves. Nouvelle Revue Francaise d'Hemaro;Lx;A= 10:477-484. Goldstein BD (1988): Benzene toxicity. Occup Med: State of ~7.t i Reviews 3541-554. Goldstein BD (1990): Is exposure IO benzene a cause of human mulc-h. myeloma? Ann NY Acad Sci 609:23-230. Greenland S, Salvan A. Wegman DH. Hallock MF, Smith TJ (1% : case-control study of cancer mortaliv at a transformer assemblyfacilin ir Arch Occup Envrron Health 66:49-%. Heineman EF, Olsen JH.Pottem Lhl. Gomez M. Raffn E, Blair A 1 !?-I Occupational risk factors for multiple myeloma among Danish men. C; -? Causes Control 3555-568. Hill AB (1965): The environment and disease: Association or c a m : : Proc Roy SOCMed 58:295-300. Hoar SK, Morrison AS. Cole P. Silverman DT (1980): An occupation ZL exposure linkage system for the srudy of occupational carcinogenm- ~ Occup Med 22:722-726. Blair A. Linos A. Steran PA Burmeister LF. Gibson R, Everett G. Schuman L, Cantor KP t19921:Evaluation of risks for non-Hodgkin's lymphoma by occuparionand industry exposuresfrom a case-control sNdy Am J h d Med 231301-312. Bond CG, McLaren EA. Baldwin CL, Cook RFt (1986): An update of monality among chemical Uorkers exposed to benzene. Br J Ind Med 43:685-691. Checkoway H, Wilcosky1.Rolf P. Tyroler H (1984):An evaluation of the assoCiarionsof leukemia and rubber indusn). solvent exposures. Am J Ind Med 5239-249. Chinese Epidemiologic Study Group of Leukemia and Aplastic Anemia (1992): Risk factors analysisof leukemia and aplastic anemia in China (in Chimes). Acta AcademiaeMedicinaeSinicae 14185-1 89. Ciccone G, Mirabelli D. Lais A. CavarottiP. Rege-Cambrin G, Davico L, V i s P (1993): Myeloid leukemias and myeludysplastic syndromes: Chemical exposure, histologk subtype and cytogenetics in a case-conml study.CancerGenetCytoeeoa68:135-139. Craoe MM,Keating MI.TiujiUoJM,Labanhe DR,Fran'kowskiRF (I 989): Environmental exposures in cytogenetically defined subsets of acute nonljmphocytic leukemiaJAMA 262:634439. Crane MM, Godwin JE Annegers JF, Keating MJ (1992): Is histological subtype a marker for environmental exposures in acute myelogenous . leukemia? Cancer EpidemiotBiomarkersR e v 1:183-388. W f l t P, Blattne~WA. Blair A (1983): Mortality among chemical workas exposedio b c n aa~d~other agents. Environ Res 30:16-25. . Delzell E,Austin H,Cdc P(1988): Epidemiologic studiesof the penoleum industry.OccupMed:Stueof theArt Reviews3:455474. ,DosemcciM,Hayes,RB, Yii S-N, Linet M,Chow W-H,Wang Y-2, Jiang ZL,Dai T-R,plang W-U, chao 2-J, Ye P-Z'Kou Q-R, Fan Y-H,plang Hricko A (1994): Rings of controversy around benzene. Environ f i z x Perspect 102:27&281. Hurley JF, Cheme JW. MacLareo W (1991): Exposure to benzene &mortality from leukaemia: Results from coke oven and other coal prcc-r workers. (Letter).Br J Ind Med48:502-504. lshimaru T,Okada H, Tomiyasu T. Tsuchmom T,Hoshino T,Ichimanr M (1971): Occupational factorsin the epidemiolw of leukemia in Hiroshh and Nagasaki.Am J Epidemiol93:157-165. International Agency for Research Cancer (1982): "IARC Mono& K. the Evaluation of the Carcinogenic Risk of Chemicals to Humans. SOEX Industrial Chemicals and Dyestuffs." World Health Organization Inter-tional Agency for Research on Cancer. Lyon. France, 94-148. Linet MS, Stewart WF, Van Natta ML. McCaffrey LD, Szklo M (198;: Comparison of methods for determining occupational exposure in : case-control interview study of chrwic lymphocytic leukemia. J Occur Med 29: 136-141. L k t MS, Harlow SD, McLaughlio JK (1987b): A case-control md\ d multiple myeloma in whites: Chnmic antigenic stimulation, -+or: d drug use. Cancer Res 47:297%2981. h o sA. Kyle RA, OFallon WM, Kurland LT(1980): A case-conuul sr~? ofoccupational exposum and leulacmia. Int J Epidemiol9131-135. Malone KE, Koepsell lD,D a h g JR. Weiss NS,Moms PD,Taylor Pv. Swanson GM, Lyon JL. (1989):Chronic lymphocytic leukemiain relarimIC chemical exposures. Am J Epidemiol1301152-1158. McMichael AJ, Spinas R GambleJF.Tousey PM (1976):Mortality amW rubber workers: Relationship to specificjobs. J &cup Med 18:178-185. On MG, Teta MJ,Greenberg HL (1989): Lymphatic and--eh tissue cancer in a chemical manufrturing mvironment. h.J Ind h i d 16631643. *. .ri(4 Benzene and Lymphatic/Hematopoietic Cancer 295 ~ I GT. ownsend JC, Fshbeck WA, Langoer RA (1978): Mortality ,,,,ong individuals ~ c c ~ p a t i ~ n aelxl yposed to kmme. Environ ~ ~ 33&:3-10h. pJustenbach DJ, Bass RD. Rice P (1994): &mme toxidty and risk ,S,sment, 1972-1992: hplications for future rcgalatioaEnvinm Health p.=nvtSuppl 101(S~pp16):177-200. pLxton MB. Chinchill VM, Brett SM.Rodricks JV (1994): lmkemia risk ,wiJtcd with benzene exposure in the Pliofilm ooborcI. .Monabty update n,~ e~posuredistribution.Risk Anal 14147-154. petry C. Stanek E, Gloeckler L (1981): Aawacy of cancer death &ticat6 and its effect on cancer mortality statisrics Am J Wlic Health -1:2-12-250. Richardson S, Zittorn R Bastuji-Garin S, Lasserrc V, Guihemeuc C, cjdiou M.Gguie F, Laffont-Faust I (1992): Oxupuional risk factors for ,cute leukaemia: Acase-control study. Int J Epidemiol11:1063-1073. Rln,k: U. Smith AB, Homing R, Filloon TG. Young RJ, Okun AH, Ludndnpn PJ (1987): Benzene and leukemia: .b~epidaniologicrisk ,.,uxsmtnt N Engl J Med 3 16:1044-1050. Thorpe JJ (1974): Epidemiologic s w e y o f l e u k h in prisons poantidly exposed to bemamzJ Occup Med 16375-382 Travis LB, Li C-Y,zhang Z-N, Li D-G.Xn SN. Chow W-H, Li G-L. Doserneci M. B l a W. Fraumeni JF Jr, Hay= m, Lioct MS (1994): Hematopoieticmniignanciesand related disorders among bmzcneexposed workers in China Leukemia and Lymphoma 1491-1@2. Tsai SP,Wen CP. Weiss NS,Wong 0.McClellaoWA.Gibson RL (1983): Retrospective mortalify and medical s w e i l l a n a studia of workers in benzene areas of mincrics.J Occup Med 25:685-692 US Department of Health and Human Services { 1980): T h e International Classification ofDiseases," Ninth Revision, Clinical Modification. Volume 1. DHHS Publ. NO.(PHS)80-1 260. Vai T,Radice L. Catenacci G, Biscaldi GP,Guerciiena S. Pesatori AC. Bertazzi PA (1989): Studio a distanza di 304 casi di sospetta patologia da benzene osservafi negli anni 1950-1971 (in Itatian). Med Lav 80397404. Vianna NJ, Polan A (1979): Lymphomas and occupational benzene exposure. Lancet 1:1394-1395. Kothman XU. Boice JD Jr (1982): "Epidemiobgic Analysis with a programmable Calculator." Epidemiology Resouroer Boston, Massachu- dtS. Rurhton L. Alderson h4R (1981): A caseconrrd irudy to investigate the .i,,ociaion between exposure to benzene and deaths fmm leukaemia in oil xnner). workers. Br J Cancer 43:77-84. j.ni[i DA. .And?ews KW (1996) Risk of myelogenous leukaemia and inuhprc nyeloma in workers exposed to benzene tknnt. &cup Envim \led 1996:53:357. b I t z Dx. Moure R (1984): Cancer risk among oii rerimry workers. J Occup Med 26562470. Vigiiani EC, Saira G (1964): Benzene and leukemia. N Engl J Med 2712372-876. Wong 0 (1987a): An industry wide mortality study of chemical workers occupationally e x p o d to benzene. L General Results. Br J Ind Med U 3 6 5 - 3 8 1. Wong 0 (198%): An industry wide monality m d y of chemical workers occupationally exposed to benzene. II. Dose-ressnse analyses. Br J Ind Med 44:382-395. Wong 0 (199Y):Risk of acute myeloid leukemia and multiple myeloma in workers exposed to benzene. Occup En\ironMed52:380-384. Schsrr P.A. Hutchison GB, Neiman RS (1992): Sioa-Hodgkin's lymphoma Wu W ( 1988): Occupational cancer epidemiology in the People`s Republic mndoccupational exposure. Cancer Res (Suppl)52EO3s-5-Ws. of China. J &cup .Med 30:968-974. SchumacherMC.Delzell E (1988):Adeath-ceniiimtc casz~aonosltudy of non-Hodgkin's lymphoma and occupation in men in Xorth Carolina. Am J Ind `.Id 13:317-330. S e r e r 1i.K Jr. Sundin DS. Frazier TM, Robinson CF (1991i:Development, use. md availability of a job exposure matrix based on Xational Occuparional HazardSurvey data.& J Ind Med 20:16>174. Siemiatycki J (1991): "Risk Factors for Cancer in the Wkkplace." CRC Press. Boca Raton. Florida. Stewart WF. Stewart PA (1994): Occupational case-conml studies: I. Collecting information on work histories and work-related e?rposures.Am J Ind \fed 26297-312. s\\.::n GMH. Meijers M V (1989): Risk assessment of leukaemia and t`c.`cilpanonalexposure to benzene. Br J Ind Med 16.826-8Z4). Yin S-N. Li G-L Tain F-D, Fu Z-I, Jin C. Chen Y-J. Luo S-J.Ye P-Z Zhang I-2, Wang G C . Zhang Z-C,Wu H-N. Zhong gC (1989): A retrospective cohort study of leukemia and other cancers m benzene workers. Envkon Health Perspect 82207-213. Yin S-N.Lmet .M. Hayes RB, Li G-L.Dosemeci M.WangY-Z,Chow W-H, Jiang 2-L, Wachlder S,Zhang W-U. Dai T-R Chao X-J. Zhang X-C. Ye P-Z, Kou Q-R Meng J-F, Zho J-S,Lin X-E Din%C-Y. Kneller R, Blot WJ (1994): Cohort study among workers exposed to benzene in China: I. General methods and resources. Am J Ind Med 26:383-100. Young JL Jr. Percy CL,Asire AJ (eds)(1981): -Surveillance, Epidemiology, and End 'Results: Incidence and Mornlity Data. 1973-77." US Department of Health and Human Services. MI4 Publication No. 81-2330, National Cancer Institute Monograph 57. National Cancer Institute, Bethesda. Maryland