Document ZnopvpervGLvdkOe1B6xnRkLZ

^ | " * J ^ hi I s-- ) -!H. . H 'is. f -'T^-fi. . . hi^- "-.- *,'. V *! !.* .b - '. !|J1- J .. + L ... 1 i*t 1 .- . .L'hr tv I - ' ,t >*> I- * i . . I i J1^' **+.* i'llL-t -- Sources : Wiley & roenterol s. Bioodndustrial 13 (1984) on P. !ncals via a i: suppi 1. jffs. V&r lood lead 50(1982) jpplevelns at the Environ- 1983: 5). aJth Acl- ens luftlollution O. e to lead Report .*nt Prothe Na rine and iska In- jdy proxicity in Copen- ScanJ J H'or* Occupational risk factors for brain tumors A case-referent death-certificate analysis by Terry L Thomas, MS,1 Elizabeth TH Fontham, DrPH,2 Sandra A Norman, PhD,3 Annette Stemhagen, DrPH,4 Robert N Hoover, MD THOMAS TL, FONTHAM ETH, NORMAN SA. STEMHAGEN A, HOOVER RN. Occupational risk factors for brain tumors: A case-referent death-certificate analysis. Scand J Work Environ Health 12 (1986) 121--127. Numerous studies have suggested that employment in the oil refining and chemical manufacturing industries may be associated with excess brain tumor risk. A case-referent study was undertaken to evaluate brain tumor risk by occupation and industry in three geographic areas (northern New Jersey, Philadelphia, and the Gulf Coast of Louisiana) with a heavy concentration of these industries. Seven hundred and eighteen white men dying from brain tumor at age 30 years or older were ascertained from death certificates for 1978--1981. The referents were men who died of other causes, excluding epilepsy and stroke. Usual occupation and industry were obtained from the death certificates, and the maximum likelihood estimates of the relative risk were calculated for specific industries and occupations. Small nonsignificant excess risks of brain tumors were seen among persons whose usual employment was in the petroleum refining, electrical equipment manufacturing, health services, and educational services industries. Compared with other white-collar professionals, health diagnosticians, teachers, and artists/designers had a significantly elevated brain tumor risk. Among blue-collar workers, the only group with a significantly elevated brain tumor risk was precision metal workers, who are exposed to metal dusts and fumes and substances used as coolants, lubricants, and degreasers. Key terms: artists/designers, health diagnosticians, precision metal workers, teachers. Recent investigations of mortality among workers employed in petroleum refineries and petrochemical plants have suggested an association between excess brain tumor risk and employment in the petroleum and chemical industries. A literature review indicated that numerous occupational groups have been suggested as having an elevated risk of brain tumors (27). The most prominent associations have been noted among whitecollar professional groups (suggesting diagnostic sen sitivity) and workers in a variety of occupations in which there might be contact with organic solvents, lubricating oils, polycyclic aromatic hydrocarbons, formaldehyde, polyvinyl chloride, acrylonitrile, and phenolic compounds. A case-referent study was initiated in three areas of the United States with a heavy concentration of the petroleum and chemical industries to investigate the role of petroleum and products of the refining process and the roles of a variety of other factors in the etiology 1 Environmental Epidemiology Branch, National Cancer In stitute, Bethesda, Maryland, United States. ' Pathology Department, Louisiana State University, School of Medicine, New Orleans, Louisiana, United States. 1 Clinical Epidemiology Unit, University of Pennsylvania, School of Medicine, Philadelphia, Pennsylvania, United States. * Cancer Epidemiology Services, New Jersey State Depart ment of Health, Trenton, New Jersey, United Slates. Reprint requests to: Ms TL Thomas, Occupational Studies Section, Environmental Epidemiology Branch, NCI, Landow Building, Room 4C16. Bethesda. MD 20892, USA. of brain and central nervous system (CNS) tumors. Cases and referents who died between 1978 and 1981 were selected from death certificates of usual residents of northern New Jersey, Philadelphia (Pennsylvania), and the Gulf Coast area of Louisiana. The present analyses were based on the occupational statements on the death certificates of the cases and referents. Materials and methods Three areas of the United States in which at least 5 of the workforce was employed in the petroleum and chemical industries during the past 20 years were chosen as study areas. These were northern New Jer sey, the Philadelphia standard metropolitan statistical area, and the Gulf Coast area of Louisiana. Cases were all white men, age 30 years or older, who died of brain or other CNS tumors (International Classification of Disease (31) ninth revision codes 191, 192, 225, and 239.7) and whose death certificates in dicated that they were residents of any of the selected study areas. The study period was 1 January 1979 to 31 December 1981 for New Jersey and Philadelphia, and I January 1978 to 30 June 1980 for Louisiana. Men who died of causes other than brain tumor, cerebrovascular disease, or epilepsy were the source of the reference group. The latter two conditions were excluded from the reference group because of the pos sibility that some of these deaths may have resulted from undiagnosed brain tumors. For each case, one m . pf f .001192431 U ...l`[ i L JLl.V O'T; S U ;i c in o rr J. t : it were rhe same aee at death and a ere re-aderus of 'he was mi 1 r:i c1 1. LI L\r *. ^ I ... j 1 V .U occupations Tor 12 study subjects cou^d no: be clas i:: D is "-cun merit* workers i occurred primarily among same study area as the case. Death certificates for 741 cases and referents were requested from appropriate state vital records offices. Four death certificates (one case and three referents) were not located, and 21 persons whose cause of death had been incorrectly classified as a brain tumor were eliminated from the case series. One reticulum cell sarcoma of the brain was also excluded. The final series for these analyses consisted of 718 cases and 738 referents. Occupational statements from death certificates were coded without knowledge of the case or referent sified into any of the census codes. These persons were included in the comparisons that used "all other occupations" as the reference, but were excluded from the comparisons with other white-collar professionals and other blue-collar workers. Usual industry was not recorded on 133 death certificates (31 of these also had the occupation missing), and these study subjects were included as unknowns in the analyses by industry. Maximum likelihood estimates of the relative risk (odds ratios) for brain tumors and 95 Io confidence intervals (6) were calculated for specific occupations and industries. Analyses by occupation were adjusted forces. Table 3 shows the odds i among the white-collar occu; ratio greater than 1.0 was workers with blue-collar wor professionals had a significa tumors compared to all ot finding waS.due to statistica for brain tumor among healt and writers/artists/entertai excess of brain tumors am surveyors. Specific occupati status of the study subject and according to the index for the potential confounding influences of age and of occupations from the 1980 Census of the Popula marital status in these data, and risk estimates by in tion (2). In several of the analyses, occupational codes dustry were adjusted for white-collar/blue-collar status were grouped to include persons with presumably of the usual occupation. The odds ratios were con Table 2. Odds ratios (or brain similar life-style characteristics and socioeconomic sidered to be statistically significant when the 95 % status. The term "white-collar professionals" included confidence interval did not include 1.0. Industry*1 occupations classified as executive, administrative, managerial, and professional specialties (census codes 003 through 199). "Other white-collar" jobs included technical, sales, and administrative support occupa tions (census codes 203 through 389). Included in the "blue-collar" category were service, fanning, forestry, fishing, precision production, craft and repair occupa tions, operators, fabricators, and laborers (census codes 403 through 889). Industry was coded according to the 1972 four-digit Standardized Industrial Clas sification (SIC) (17). Because there was often in sufficient information to code to the four-digit level of the SIC, analyses were limited to major two-digit categories. Information on age at death, type of brain tumor diagnosis, marital status, and study area was also abstracted from the death certificates, and it formed the basis for these analyses. Sixteen persons (6 cases, 10 referents) were reported to be unemployed or "never employed." Occupation Results About half of the cases ascertained were from New Jersey (N = 361), and a third were from Philadel phia (N = 243). The smallest proportion was from Louisiana (N = 114). The age distribution resembled the pattern seen for age-specific brain cancer death rates among white males in the United States (13). Seventy-six percent of the study subjects were between the ages of 50 and 79 years at death. Table 1 shows the distribution of cases by deathcertificate diagnoses. Most cases for whom the death certificate specified the cell type of the brain tumor had glial tumors. The majority of these was glio blastoma multiforme, while a much smaller propor tion was astrocytoma or unspecified gliomas. Menin gioma and other cell types accounted for a very small Agriculture, forestry, fish (01Mining (10--14) Construction (15--17) Manufacturing (20--39) Chemicals (28) Petroleum refining (29) Electrical equipment (36) Transportation, public utilities Wholesale and retail trade (5i Finance, Insurance, real esta Services (70--89) Health services (80) Educational services (82) Public administration (91--9- Nationa! security (97) Unemployed and unknown (? * The reference group used 6 Numbers In parentheses a c Adjusted for age at death. Table 1. Distribution of brain cancer cases by death-certificate diagnosis. proportion of the cell types specified. A diagnosis of only "brain tumor" appeared on 30 % of the death certificates. Less than 15 % of the certificates had a diagnosis of a "malignant brain tumor" without Table 3. Odds ratios for bra study areas, 1978--1981.* Occupation6 Diagnosis Number Percentage specification of cell type. Because a very small percent Glial tumors Glioma (not otherwise specified) 366 77 51.0 10.7 age of the case series was nonbrain CNS tumors, the term "brain tumor" has been used throughout to refer 1 Ail white collar (003--389) * I White collar professionals I Executive, administrative, r Glioblastoma multiforme* Astrocytoma (grades 1 and II) Oligodendroglioma 201 28.0 to all CNS tumors. 82 11.4 2 0.3 Maximum likelihood estimates of the relative risk Engineers, architects, survi Math and computer scienti Natural scientists (069--08 Ependymoma 4 0.6 for brain tumor by industry are shown in table 2. The Health diagnosing occupat Meningioma Other specified diagnoses6 Malignant brain tumor 28 3.9 odds ratio for manufacturing industries was 1.0. There 12 1.7 was no excess risk of brain tumors among persons Health treating occupation Teachers and counselors ( Social scientists and urba< (not specified) 95 13.2 whose usual employment was reported in the chemical l Social recreation and relig Unspecified brain tumor Spinal cord tumor (not specified) 212 5 29.5 0.7 industry, and the odds ratio for petroleum refining J Lawyers and judges (178-- 1 Writers, artists, entertaine was only slightly above 1.0. The only manufacturing . Technical support occupa: Total 718 100.0 * Includes astrocytoma specified as grade III or IV. industry with an odds ratio above 1.5 was SIC 36 (manufacture of electrical equipment). The elevated Sales occupations (243--2 Administrative support oc ** Includes four schwannoma, two hemangioblastoma, one risk of brain tumors seen among persons whose usual ' The reference group for craniopharyngioma, two medulloblastoma, one meningiosarcoma, one retinoblastoma, and one sarcoma of the central employment was in service industries was confined to unemployed. 6 Numbers in parentheses nervous system. health and educational services. The excess of brain c Adjusted for age at dee V # .* 4 +_4 Industry* Agriculture. forestry, ffah (01--08) MMntng (10--14) Construction (10--17) Msnulaeturtng (20--38) Chemicals 08) Pstrolsum refining (2H Bsctrical equipment (38) Transportation, pubUe mow (40--48) Wholsssls and ratail trado (80--88) Finance, insurance!. real aatata (80 87) Saivteaa (TO--88) HanWh services (88) Educational services (92) Public admMstration (81--87) National security (87) Unaetploysd and unknown 88 Number exposed fWiwilv 13 19 88 SO 70 224 219 37 34 19 15 28 18 74 70 180 120 28 24 108 86 18 12 20 11 44 39 10 4 88 87 Crude odds ratio 0.7 IjO 06 t.1 1.1 16 16 1.1 0.8 1.2 1.3 1.6 1.9 1.3 26 0.7 AdHisted odds ratio* 02 13 0.9 1.0 1.0 1.2 1.7 1.1 0.7 1.0 1.4 1.7 13 1.4 46 0.9 88 % CvmTvQVVIOV Interval 0.4-- 16 0.4-- 46 06-- 16 06-- 16 06-- 1.7 06-- 26 0.7-- 46 06-- 16 06-- 1.0 06-- 1.9 16-- 1.9 0.7-- 3.4 0.8-- 4.4 06-- 26 16--20.9 06-- 16 VAB.0001192433 J si-1- 'i "V ^ ;fLcsy 'H(|'I|%',; w'hiel:**-.- :, pfb. V-*'-5" i -i i * f t p j 4 I ( f V * i t * l1 J 1 i i * '( k I P I 1 I 1 * 4 ;j* *i ? h * i -i \m t Table A. u d y a re a s a' * :, f :: r - : '0 7 S - - 1981 a 4 mv e: t < > " whdemoii :r Louro + essiorials in the New Jersey, and Phi)ade p oa Occupation N umber exposed Cases Referents Crude odds ratio Adjusted odds ratio13 Industrial engineers Health diagnosing professionals Teachers Artists and designers 5p ~i ' j 13 2 4.4 11 3 2.4 13 -- r 4 5.2 3.9 a The reference group used for each odds ratio calculation was all other white-collar professionals. b Adjusted for age at death and marital status. %95 confidence interval 1.0 -- 41.5 0.7--28.2 9 Table 7. Distribution of c t i Occupation * All white-collar professio Other white-collar worket All blue-collar workers Unknown Tablo 5. Odds ratios for brain tumor among blue-collar occupational groups in the Louisiana, New Jersey, and Philadelphia study areas, 1978--1981.* Occupation6 Service occupations (403--469) Farming occupations (473--489) Forestry, fishing, logging (494--499) Mechanics and repairers (503--549) Construction trades (553--599) Extractive occupations (613--617) Precision production occupations (633--699) Machine operators, assemblers, inspectors (703--799) Transportation and material moving (803--859) Helpers and other laborers (663--889) Unemployed and unknown (999) Number exposed Cases Referents 48 64 13 18 --3 34 29 53 68 33 91 74 49 69 47 52 45 53 26 48 Crude odds ratio 0.8 0.7 * 1.2 0.8 1.0 1.3 0.7 0.9 0.9 0.5 Adjusted oddsratioc . 0.8 0.8 * 1.3 0.8 1.1 1.3 0.7 0.9 0.9 0.6 * The reference group for each category was all other occupations including unknown and unemployed. b Numbers in parentheses are the 1980 Census of the Population codes. c Adjusted for age at death and marital status. 95 % confidence interval 0.5-- 1.2 0.3--1.7 -* 0.7--2.2 0.5--1.2 0.2--6.9. 0.9--1.8 0.5--1.0 0.6--1.4 0.6-1.4 0.4--1.1 Table 6. Odds ratios for brain tumor among precision metal workers in the Louisiana, New Jersey, and Philadelphia study areas, 1978--1981.* Occupation All precision metal workers Machinists Tool and die makers Sheet metal workers Number exposed Cases Referents 47 26 26 16 74 84 Crude odds ratio 2.2 1.9 2.0 2.3 * The reference group used for each odds ratio calculation was all other blue-collar workers. b Adjusted for age at death and marital status. Adjusted odds ratio* 2.1 1.8 1.8 2.2 95 % confidence 1.2--3.8 0.9--3.7 0.5--7.4 0.6--9.0 were 13 artists/designers among the cases and none among the referents. Two of the six cases who were designers had occupations as industrial designers and might have actually had drafting occupations; how* ever, coding rules caused them to be included in the designer classification. Two of the remaining four cases were interior designers, one designed furs, and one was unspecified. Among the six professional artists, there were three commercial or graphics artists, a wildlife artist, a freelance cartoonist, and one artist/art teacher. The only blue-collar groups for whom brain tumor odds ratios were above 1.0 were precision production workers and mechanics/repairers (table 5). Among precision production workers, only precision metal workers had statistically significant odds ratios for brain tumors when all other blue-collar occupations were used as the reference (table 6). The most com 124 mon occupations within this category were machinists, tool and die makers, and sheet metal workers. The odds ratios were elevated to a similar level among per sons in each of these occupational categories. Because brain tumor risk was elevated among cer tain white-collar groups and some skilled blue-collar workers, an attempt was made to address the diagnostic sensitivity issue. Table 7 shows the distribution of cases in occupational categories by death-certificate brain tumor diagnosis. Almost one-half of the blue-collar workers and unknown occupations had unspecified brain tumor diagnoses on their death certificates, while 32 Vo of white-collar professionals had unspecified diagnoses. Glial tumors accounted for over 60 Vo of the death-certificate diagnoses among white-collar professionals and less than 50 Vo among other workers. With blue-collar workers as the reference, white-collar Table 8. Odds ratios for t 1978--1981. Occupation White-collar workers vs t White-collar professional Health diagnosticians vs Teachers vs other profes Artists/designers vs othe Precision metal workers Machinists vs other blue Tool & die makers vs otl Sheet metal workers vs < workers had a signif tumors with known ce types (table 8). A sirr sionals when all other as the reference. Su table 8, were therefon collar professionals a professional groups a as the reference for t general, odds ratios v unknown cell types c in each of the group: Discussion Caution must be exei of the present study, are supposed to have usual occupation dur that the information i tion; thus it may not \ cause death certificate industry, one cannot c study subjects might 1 occupational statemt be a good descriptior groups, however, be Professionals or othe often. VAB.0001192434 Tabls 7. Sisvitv.J'cn o4 casts cy occuaaticnai category and deaf* -:erti'icate ciagnosis ;CNS = central nervous system] Occupation Glial tumor Num Per ber centage Death-certificate diagnosis Other ceil type Malignant brain tumor Num Per ber centage Num Per ber centage Unspecified brain or CNS tumor Num Per ber centage Total number All white-collar professional Other white-collar workers All blue-collar workars Unknown 109 61.6 95 49.2 181 47.3 11 42.3 11 6.2 8 6.1 19 S.0 2 7.7 21 11.9 19 14.4 50 13.1 5 19.2 36 203 40 30.3 133 34.7 8 30.8 177 132 363 26 TaMefL Odd* ratio* for brain tumor by d*ath-cartiticat* diagnosis in tha Louisiana, New Jersey, and Philadelphia study areas, 1978--1981. Occupation tMil|Mal|a piMlMfE are MuMdlllf HMHllMA Whitre nnllrer omfrererefeMElre vi otlirer mtiltre nnilif ipoiIac HPOtth dhpRQllWPR DP PRP pfpIPPplOMlP ArtittaMpoipoprp dp othpr pipiMPipMip PfPflftfllM Ml MARuk are rettlOr hjim rnflM~ mmlfreri MpOMflMp Dp flppf MhMHNNMP DMMhtIP Tool 1 iPre idOMi vre oHhrer MreMfiOttrer wortorett SHott mptpl mottmm dp ottior MvipoIp poApi Cell type known Exporered Ad)u*tPd cases odds ratio 183 13 180 24 10 5.1 6 3.1 9* 24 24 18 23 4 14 3 14 OpU typp unknoim cease odm^aMo 118 1.1 57 13 3 2.7 5 84 4 23 2.1 10 13 3 1.7 8 33 worken had a sigaificantfy elevated risk for brain tumors with known oed types but not for uaknown ced types (table t). A stnrihr result was seen for profes sionals whea an other dUikoIm workers mere ased as the reference. Swhscouent results, nresaatad in table t, wm therefore rsloiMretl with aB other whitecollar pwifs--lonP as die rdsranee Ior each of the professional groapa and all other bias cedar workers me mennoe w dr pmiiw smui wormra* in eneral, odds ratios were elevated for known and for unknown oed types of brain ttunors among workers in each of the groups shown. Diacuaaion Caution must be cncrdacd ia interpreting the results of the present trady. Even though deadi certificates ere supposed to have infbnuatioa cm the deceased's usual occupation during Ms lifetime, it is very Ukciy that the information recorded is often the last occupa tion; thus it may not be the deceased's lonprt job. Because death centneates record onfy one oocapation and industry, one cannot evaluate risks far ad jobs in which study subjects might have been "ever employed." The occupational statement on the death certificate may be a good description of longest occupation for some group, however, because it is unlikely that trained professionals or other skided workers dump careers often. A second limitation is the use to identify brain tumor da fof metastaam from otbar some of the J,rg--pps listed as "braia "braia tumor" may not erfwdy b tumors. This Is proMririy not a major ever, became a review of death National Cancer Survey areas bidketadjihat about 10 fk ofthe dtath certificates that dstadalnPananer of the brain as underlying caure were not confirmed by hospital reoords (19). Thus ouly about 72 of our measeswmuw adsdlmmoses. Fbmdy, became we have made muhiple some stailitirady significant muds would due to dwnoe alone. Despite them of death-certificate occupational provide useful dues for plaaafa* detailed complete occupatiooal history data terviews, particularly for skided worfcetsfior profes sionals. Creenwald ct at (7) suggested that d tivky may account for the elevated brain observed in numerous occupational words, braia tumors might be more nosed among groups that have good access care than among those who might not medical surveillance, la our study, excess risk was observed among white-collar member* of the military; however, < tivuv docs not cn:irc!y e> plain these 'inJings. Teach ers. health diagnostician;, and trusts/ designers had elevated risk of brain tumors even when all other whiie- teachers in a O.'il jr i; \Lirvey (20). Artt>ts are ex posed to numerous chemicals. including inks, solvents, paints, and pigments (14). In our study, most of the , ". Green w. aid P. Frio Earle K. Diagnosti explanation for an Med 23 (1981) 69- coltar professionals were used as the reference, and artists/designers were commercial artists or draftsmen 8. GuraJnick L. Mon. there is no reason to believe (hat they should have better access to medical care than other white-collar professional groups. The distribution of diagnoses in table 7 shows that there are some variations in the specificity of the death certificate diagnoses by oc cupational category. Although it is not known whether professionals are more likely to have a brain tumor diagnosed than other workers, the presence of a higher whose primary occupational exposures presumably are to inks and solvents. Three surveys of mortality by occupation have shown elevated frequencies of brain cancer among teachers (S, 8, 20). In these surveys and in our study it was not possible to determine whether this ex cess might be confined to one teaching specialty (eg, laboratory science teachers) or whether the excess risk i i I t i among men 20 to US Government P (Vital statistics sp 9. Hadjimichael OC. RE. Mortality at lemployees in a nut Med 25 (1983) 48 10. Harrington JM, O Otologists, 1974--: proportion of celt type diagnoses on the death cer occurs among all teachers. 1 11. Katz RM. Causes tificates might result from greater access to medical diagnostic techniques. Thus, if diagnostic sensitivity were responsible for excess brain tumor risk in a par ticular group, one might expect elevated odds ratios Our study showed a significantly elevated risk of brain tumors among precision metal workers, including machinists, tool and die makers, and sheet metal workers. These findings are consistent with occupa Occup Med 2S (I* 12. Key MM. Henscht IR. EdeL.ed.Oc recognition. US ( ingtoa. DC 1978. for known tumor cell types but not for unknown tumor cell types. In table 8, risks for brain tumors of known and of unknown odl types were elevated among all of the groups listed. Even though some groups had a cell tional mortality surveys that suggest excess mortality from brain tumors among machinists (5, 8). More deaths from brain tumors than expected were re ported among workers in three nuclear fuels and 77--lit). 13. Mason TJ, McfCaj JF. Atlas of cum 1969. US Governn ... I97J. (DHEW pu type recorded more often than others, thdr risk of "cell weapons fabrication fadUtfes (3,9,30) which had large 14. McCann M. Artist type known" tumors was similar to their risk of "cell machining operations. Blue-collar (hourly) workers type unknown." Elevated risk was not due to an ex in an aviation electronics plant had a significantly New York, NY II IS. Miller BA. Blair / sionai artiste: A p cess risk for known tumor cell types alone. devated frequency of brain cancer (24), and all five of Toxicol Oncol 6 ( No excess risk of brain turnon was seen among per sons whose usual employment was in the chemical manufacturing industry. Studies of individual plants the hourly workers with brain cancer were machinists. A slight excess brain cancer risk was also reported among automobile machinists hired after 1948 (4). | 16. Mchoboo WJ, Sri Brain tumors ansoend petrochemical NY Acad Sd 381 and chemical operators in Texas indicated that there Machinists are exposed to metal dusts and fumes, 17. Office of Manager may be elevated risk of brain rumors associated wkh employment in the chemical manufacturing industry (16, 21,29), and a study of pharmaceutical workers lubricating oils, and solvents (12). Epidemiologic studies of numerous occupational groups with ex posure to lubricating oils and solvents have also re classification man Office. Washingt* 18. Olin OR. Aldbot Swedish shemiro showed a significantly elevated frequency of brain ported excess mortality from brain tumors (27). viron Rat 22 (198 tumors among production workers (26). Previous Findings of this preliminary analysis indicate that I 19. Percy C. Stanefc J epidemiologic analyses indicated a possible excess risk devated brain tumor risk may be associated with cer I death certificates statistics. Am J P of brain tumors among workers ht petroleum refineries tain occupational exposures. Interviews arc being con 20. Peterson CR, Milt (22, 25, 28), but the present analysis indicated only ducted to obtain lifetime occupational histories from stateof CsBfornia a small nonsignificant excess risk of brain tumors the next-of-kin of cases and referents. A hospital associated with usual employment in petroleum re record and pathology review will allow case-referent Office, Washiagtc 'Tkation no 80-- It fining. analyses by the cell type of the brain tumor. i Elevated odds ratios for brain tumors seen among persons in the health and educational services are * t accounted for by the excess risk seen among health diagnosticians and tcnchers. Epidemiologic studies have suggested elevated brain tumor risk among registered female mines (11) and veterinarians (1), but not among other health professionals. Exposures in the medical retting fachnle numerous biological agents, and chemicals like disinfectants, antiseptics, and pharmaceutical preparations (12). Similar occupational exposures occur among chemists (18), pathologists (10), and anatomists (21) reported to have an elevated brain tumor risk. A proportionate mortality ratio analysis of deaths among professional artists showed a significantly elevated frequency of brain tumors (15); however, this devated frequency was not associated with any par ticular art specialty. A slight excess frequency of brain cancer deaths was also seen among artists aad art t. Blair A. Hayes HM. Monalky patterns among US wrerinariam. 1947--1977: Aa mpidded study. fat I RpMsailol II (ttfi) >91--397. 2. ""Tint itffm-- It-- nsmm of fnpglatlon naisilkd index of ladanries aad oampaiftaejUS Ooeanuprnt Printing OfBos. Weifclnpee.DC 19*2. (PHCriMU). 3. Craglc PL, BorisOft, Qaahsn JlLTaofccwty WO, FrySA. A morarifey etuiftrof asu exposed to dementa! ameury. I Oocup Mid 28 (HB4) 811--821. 4. Decoufle P. Further taajyrir of caaasr mortality pat- teras among worhon reposed to ratting oH mins. J Hail Cancer lost 81 (1978) 1025--10)0. 5. Dutwow It. Wopaaa Ml. Cancer and occupation In Massachusetts: AdaMh conMteare nuay. Am J lad Med 8 (1984) 297--239. 6. Part JJ. Point and imtrvl estimation of dm common odds ratio in the comhittarion of 2x 2 tables with fixed marginals. Btonssuika 57 (1970) 4?l|-475. `ire:: * - i I : ! . I :* : ilctrm: I ^; *C jK. ll.-l t K \n *A . v 'oi. kk, A jAwnier I IrarSc K ;1 .ir u.t.. *i:<. I-; An c^Ktetwologit: RJ. K: ihheck `A a . \ i .rive:,ligation oi twain umvirs i-\plu:ia:io f. r an appurer.! h'un uaior excess J Occup among chemical par.: employees using a sample-based Med 23 (1981) fV0--694. cohort method. J Occup Med 25 (1983) 387--393. Ouralniok L. Mortality hy occurs :on and cause of death Schoiccnfeld D. Warshauer ME, Zauber AG, Meikle JG, among men 20 to 64 years of age: United States. 1950. Hart BS. A prospective study of morbidity and mortality US Government Printing Office, Washington. DC 1963. in petroleum industry employees in the United States -- (Vital statistics special reports, vol 53, no 31. A preliminary report. In: Peto R. Schneiderman M, ed. Hadjimichael OC, Osifcld AM, D`Atri DA. Brubaker Quantification of occupational cancer. Cold Spring RE. Mortality and cancer incidence experience of employees in a nuclear fuels fabrication plant. J Occup Harbor Laboratory, Cold Spring Harbor, NY 1981, pp 247--260, (Banbury report no 9). Med 25 (1983) 48--61. Stroup NE, Blair A, Erikson GE. Brain cancer and Harrington JM. Oakes D. Mortality study of British pa thologists, 1974--80, Br J Ind Med 41 (1984) 188--191. other causes of death in anatomists. Am J Epidemiol 120 (1984) 500. Katz RM. Causes of death among registered nurses. I Occup Med 25 (1983) 760--762. Key MM. Henschel AF. Butler J, Ligo RN, Tabcrshaw IR. Ede L. cd. Occupational diseases: A guide to their Sweeney MH, Ahrenbolz SH. NIOSH health hazard evaluation report HETA 80-250-1529. National Institute for Occupational Safety and Health, Cincinnati. OH 1984. recognition. US Government Printing Office, Wash ington. DC 1978. (DHEW (NIOSH) publication no 77--181). Mason TJ. McKay FW. Hoover R, Bios Wi, Fraumcni JF. Atlas of caacer aiortality for US counties: 1990-- 1969. US Government hinting Office, Washington, DC 1975. (PHEW publication no (NIK) 75-780). McCann M. Artist beware. Watson-GupdU Publications. New York, NY 1979. Miter BA, Blair A. Mortally patterns among profes sional artists: A preliminary report. J Environ Pathol Toxicol Oncol 6 (1985) 303--313. Nicholson Wi, Sddmaa H, Settoff U, Tan D, Cteric E. Brain tumors among openiing mglwren in the dtemiad and petrochemical industry in Texas and Louisiana. Ana NY Acad Sri 381 (1982) 172--180. Theriault G, Goulet L. A mortality study ofoil refinery workers. J Occup Med 21 (1979) 367--370. Thomas TL, Decoutie P. Mortadhy among workers employed in the pharmaceutical industry: A preliminary investigation. J Occup Med 21 (1979!) 619--623. Thomas TL, Waxweiler RJ. Brain tumors and occupa tional risk factors: A review. Scand J Work Environ Health 12 (1986) 1--15. Thomas TL. Waxweiler RJ, Crandall MS, While DW, Mourc-Eraso R, Itaya S. Fraiimeid JF. Brain cancer among OCAW members in three Texas oil refineries. Ann NY Acad Sri 381 (1912) 120--129. Waxweiler RJ, Alexander V. Lcffingweg SS, Haring M, Lloyd JW. MortaBty from brain tumor and other causes in a cohort ofpetrochemical workers. J Natl Cancer Inst 70 (1913) 73-81. daxsificatioa manual. 1972. Uf Govenuncnt Printing Office, Washington, DC 1972. Olh GR, AftOhom A. The cancer mortality among Swedish dhendas graduated during three decades. En viron Res 22 (1980) 154--161. Percy C, Staaek E, Gloeelcicr L. Accuracy of cancer death certificates and its effect on cancer mortality statistics. Am I Public Hmhh 71 (1981) 242--250. Peterson OR. Milham S. Oecupntioaal mortaBty hiihe state of California. 1999--01, US Government Prhuing Wilkinson GS, Vodc CL, Aeqnavdte JF, TteUcn OL, Reyes M, Brackbfll R. Wiggs L. Mortality among plutoniura and ocher workers at a nuclear facility. In: Epidemiology applied to health physics: Proceerings of the sixteenth midyear topical meeting of die Health Physics Society. National Technical teTonmaion Service, Springfield, VA 1983, pp 328--337. (CONF430I0I). WorM HcaMi Organteation. Manual ofthe international and cawcs death, 1975 revision. World Health Organization, Geneva 1977. Office, Washington, DC 1980. (DHEW (NIOSH) pub lication no 80--104). Received for publication: 12 November 1985 1 I r ri.V *-!i \ II Trt -v - VAB.0001192437 -i i rdsftj ill