Document evd08Y2w3jzy1qRYwgeXo4x64

1210 Occupations Associated with Multiple Myeloma Figgs et at -- Risk of Multiple Myeloma by Occupation and Industry Among Men and Women: A 24-State Death Certificate Study Larry W. Figgs, PhD Mustafa Dosemeci, PhD Aaron Blair, PhD Multiple myeloma is a malignant plasma cell proliferauon rare in people under age 35. Rates are higher among men than women, with the highest worldwide incidence rates oc- cumng among black African descend- This cancer surveillance investigation uses death cerrif?catesfiorn24 states ants. Although its etiology is largely ./or [he period I984- I989 lo identxv multiple m-veioma and occuparion unknown. epidemiologic efforts sug- . associations and to stimulare hypotheses. .4 case-control studj. of"multiple m.veloma was createdfrom 3,159.417 certokates in which 12,148 male and .iemaie cases were frequency matched by age, race, and gender with j v e conrrols per case. We screened 23I industries and 309 occupations. iT*omen demonstrated sign@cant excess risk among managers and administrators, gest several occupational associations. including agricultural. metal, rubbe:. benzene, wood, leather, textile. and pezroleum occupations. It is difficult. however. to assemble populations of men and women that are sufficiently posr-secondan: teachers. elementan, teachers, sociaI workers, other sales large to evaluate occupation and rare u,orkers,waitresses, and hospital maids. Men showed significant r i s k among disease associations. We used death computer system scientists, veterinarians, elementary teachers, authors, en- certificates for black and whte resi- gineering technicians.general oflce supervisors,insurance adjusters, barbers, dents from 24 states for the years riectronic repairers, supervisors of extracting industries, production supervi- 1984-1989 to generate clues to occu- sors. photoengravers. and graderldozer operator:. Men and women elemen- pational associations with multiple rar!. school teachers demonstrated the most consistent. statisticall>s*i,onrficant myeloma. increased risk qf multiple myeloma. Methods h m the Occupational Studies Section, Environmental Epidemiology Branch, National Cancer Institute, Berhesda. Maryland. .4ddress correspondence to: L. W. Figgs. PhD. National Cancer Institute, 6130 Executive B1Vd.. MSC 7364. EPN. Room 1IS. Bethesda. MD 20892-7364. -0W-h- l736i94i36 I 1 12 IOS03.00/0 Copyright 5 by , h x n c a n College of Occupational and Environmental Medicine The Xational Cancer Instirule (NCI). National Institute for Occupational Safety and Health. and the Sztional Center for Health Statistics a t veloped the death certificate data basused in this study. Participating S:at?S (Colorado. Georgia. Idaho. Indiana. Kansas. Kentucky. .Maine. ,Missour,. Nebraska. Yevada. New Hampshir-. New Jersey. New Mexico. Nonh Cx- olina. Ohio. Oklahoma. Rhode ISland, South Carolina. Tennessee. Utah. Washington. West Virginll. Wisconsin. Vermont) began coding occupation and industry titles or. death certificates in 1984. 301 31: states contributed in 1983. Occupation and indust? codes correspond 1.7 1980 US Census Bureau classificatioc , JOM Volume 36, Number 11, November 1994 121 of 23 I industries (IC) and 509 occupations (00.' From 3. i 59.1I 7 total deaths. .:I I45 multiple myeloma deaths (International Ciassitication of Diseases. h i n t h Revision (ICD))among white ( 3 7 . 9 5 )and black I 11.55) men and women were selecied. Five controls. dying of diseases other than cancer, were frequency matched to the cases by 3ge (5-year lige groups). gender. and race. '4subgroup ofcontrols. consisting entirely of individuals dying from cardiovascular disease. was also used for specific analyses. .Associations between IC/OC codeis) and multiple myeloma were derived from case-control design' using standard logistic regression methods.' Proportional mortality ratios (PMRs)were derived using me?hods deve!oped by hlonson' and sofrware developed at NCI" to identify eligible controls with an unrelated disease and no proportional mortality excess related to their occupation. Race-specific and gender-specific analyses of black female. black male. white female. and white male mortality odds ratios (ORs)were adjusted for age. autopsy status. and urbanirural residence. except where noted. For region-specific analyses. cases and controls were selected only from that region. with ORs adjusted for individual state contributions. Three-digit OC or IC c!assification codes were evaluated. Typicdly. decedents with specific OC or IC codes were compared to .'all others" codes. Occasionally. if socioeconomic status fSES) was thought to confound a result. the unexposed were a restricted group ot'occupations. Mlitarv, retired with no occupation reported. homemaker. student. volunteer. and never worked or disabled OC codes were added to the original 503 occupation codes of the 1950 census list. In all. 509 OC codes and 23 1 IC codes were screened. Where appropriate. separate three-digit codes such as secretaries (OC3 12) and stenographers (OC 3 14) were analyzed together. ORs are presented if there were five or more deaths among cases and the estimate was 22.0. or statistically significant. except where noted in table footnotes. Results Cases and controls displayed similar distributions in rural/urban residence. autopsy status. and geographic region (Table 1). Whites represented nearly five times as many total mul- tiple myeloma deaths as blacks. Ac- cordingly, more multiple myeloma/ occupation associations were examined for whites. Occupation Of the 301 occurrences of five or more exposed cases, 23 statistically significant (lower limit of 95% confidence interval (CI) 2 1.05) elevated ORs occurred. when only 7.5 would have been expected by chance alone. Women who were managers and administrators. post-secondary teachers, elementary teachers, social workers. other sales workers, waitresses. or hospital maids showed significantly in- creased risks. Men who were COT puter system scientists. veterinariar.. elementary teachers. authors, eng. neering technicians. general officer SL pervisors. insurance adjusters, bar bers. electronic repairers, supervisor of extracting industries. productior supervisors. photoengravers. o graderldozer opentors also showec significantly increased risks. Womer who were homemakers or had neve: been employed showed significantl;, reduced ORs, whereas men who were painters. construction laborers, or laborers or had never been employec showed significant decreased association. OCs with ORs 1 7.0 but nor statistically significant are also shown in Table 1. Across ncelgender groups, teachers demonstrated rhe strongest association with ORs, ranging from 26.8 for black female secondary school teachers to 1.3 for black male elementap TABLE 1 Population Parameter Distribution for Multiple Myeloma Cases and Controls Parameter Cases n (%I Controls n 196) Age grow CJ5 45-64 255 187 (01) 3,020 (25) 8,941 (74) 935 (01) 15,100 (25) 44,705 (74) Racelgender Black female White female 997 (8) 5,056 (42) 4.985 (8) 25.280 (42) Black male White male 946 (8) 5,149 (42) 4.730 (8) 25,745 (42) Rural/urban Metropolitan Nonrnetropolitan Other 7,720 (64) 4,426 (36) 2 (0) 37,ai7 (62) 22.385 (38) 38 (0) Autoosy Yes NO Unkown 465 (04) 9.363 (77) 2,320 (19) 5.580 (09) 45,900 (76) 9.260 (15) RegionNortheast Southeast Cantral West 1,242 (I0) 4,217 (35) 5.416 (45) 1,273 (10) 6,107 (10) 22.149 (36) 26,771 (44) 5.713 (09) * Northeast = Mame, New Hampshire, New Jersey, Rhode Island, Vermont: Southeast = Georgia, Kentucky. North Carolina. South Carolina, Tennessee, West Virginia; Centra = Indiana. Kansas. Missouri. Nebraska. Ohio. Oklahoma, Wisconsin: West = Colorado. Nevada. New Mexico, Utah. Washington, Idaho. 1212 Occupations Associated with Multiple Myeloma Figgs et al ~~ TABLE 2 Mortality Odds Ratios, 95% Confidence Interval, and Cases Exposed for Multiple Mveloma. According to Occupation Code Occupation Cases Exposed' Mortality Odds Ratio 9S0/o Confidence Interval Manager and professional specialty occupation [003-199]t Black females White females Black males White males 93 655 81 1,109 1.5 1.2-2.0 1.3 1.1-1.3 1.6 1.2-2.1 1.4 1.3-1.5 Management and administration (0 19)* White females Black males White males 120 20 416 1.3 1.1 -1.6 1.8 1 .l-3.1 1.4 1.2-1.5 Chemical engineer Whlte males 10 2.0 0.9-4.1 Computer system analyst and scientist (064) White males 5 7.9 1.8-33.8 Vetennarian (086) White males 6 3.2 1 .l-9.0 Post-secondary teacher (154) Black females 5 26.8 3.1-233.3 Elementary teacher (156) Black females White females White males 38 1.6 1 .l-2.4 224 1.5 1.3-1.8 64 1.8 1.4-2.5 Social worker (174) Black females 7 2.8 1 .l-7.4 Author (183) White males 6 3.4 1.2-9.9 Technical. sales, and administration support occupatron [203-3891 Black females White females Black males White males 58 838 58 758 1.3 1 .O-1.8 1.2 1.1-1.3 1.2 0.9-1.7 1 .o 1 .0-1.1 Engineer technician (216) White males 12 2.7 1.3-5.4 Other sales worker (274) Black females 8 2.6 1.1-6.3 school teachers. Elementan. school teachers' risks were consistently increased. industry Among all industries, of the 327 occurrenceswith five or more exposed cases (Table 3). 19 demonstrated statistically significant increased associations viith multiple myeloma deaths. when only 8.2 would have been expected by chance alone. Women involved in blast furnace industries. hardware retailing, grocer) businesses. jewelry industries. and insurance showed significant associations. Other industries showing significant associations among women included persons working in physicians' offices. health services. elementary and secondary schools, and religious organizations. Men. in contrast. demonstrated significant associations in industries involved in the production of dairy produc's. plastic footwear, iron and steel. crdnance. electronic computer equipment. unspecified electrical equipment. water transporntion. wire-mho-telephone communications equipment. and urban transportation vehcles (buses). Men. but not women. Zmployed in collegesand univerjities also had elevated ORs. Both men and women showed significant associations with businesses invoived in real estate. finance and securities industries. church o r g a i tations. and educational insrimtions such as elementa?/seconW schools. Statistically significant decreased associations occurred among black males in agriculture. white females in eating and drinking retad trades. and white males in dwelling building repair and general government industries. Other ICs with ORs L 2.0 that were not statistically significant are also included in Table 3. General office supervisor (303) White males Payroll clerk (338) White females Postal clerk (354) White females 11 8 8 2.2 2.1-4.4 2.0 0.9-4.5 2.1 0.9-4.8 Elementary School Teachers Analysis Region-specific analyses indicatei that multiple myeloma was elevated among elementary school teachers 41,3 all geographic regons (Table - ' al JOM Volume 36, Number 11, November 1994 1213 Jol in- TABLE 24ontinued Mortality Odds Ratios, 95% Confidence Interval, and Cases Exposed for Multiple Myeloma, According to Occupation Code White elementary school teachers showed statistically significant increased associations when circulatoq diseases were used as the control 27 Occupation Cases Exposed' Mortality Odds Ratio 9596 Confidence Interval group or when the unexposed were limited to managers and professionals ed Traffic. shipping clerk (364) ;a- White females 3- 7 2.1 0.9-5.2 (OC = 001-199). ORs for multiple myeloma among elementary school teachers were reduced from the pre- IS. Insurance adjuster (375) vious analysis. but remained elevated Y- White males 10 3.3 1.5-7.3 above 1.0. When the unexposed were 2- restricted to secondary school teachers Service occupations [403-469] 'S, Black females 329 S. White females 377 1.l 1.o-1.3 0.9 0.8-1 .o (OC = I 13-154), ORs were elevated only among white females. .,cP Black males 124 1.o 0.8-1.2 .-:r White males 28s 0.9 0.8-1 .l Discussion :- Waiter (435) Death cerrificate use in cancer % Black females 5 4.9 1A-17.0 surveillance studies has a long his- tory.'-'" Although cause of death Miscellaneous food preparation and occupation/industrJ reponing (444) Black females 7 2.5 1.O-6.3 errors occur.' I-" death certificate studies provide sutliciently large Hospital. maid (449) Black females populations to evaluate rare occupa- 32 1.8 1.2-2.7 tions andlor cancers and inexpensively identify changes in associations over time. '' Multiple myeloma case I Barber (457) White males ascertainment is reliable. with 96% 26 1.7 1.1-2.7 detection and 98% confirmation Child care worker (468) Black females ntes. respectively." although agree7 2.6 1.O-6.6 ment appears to be poorer for women and nonwhites.'' Farming. forestry, & fishing occu- We caution that nonpanicipating pation [473-4991 Black females White females Black males White males states with Iarse. urban. and black 20 17 89 589 0.7 0.5-1.2 1.1 0.6-1 .a 0.7 0.5-0.9 1.o 0.9-1.1 population centers-such as Xew York, Pennsylvania. Illinois. Michigan. Texas. and California-led to less stable estimates of association among Sucemsor. f a n worker (477) White males African Americans and other minori5 2.5 0.8-7.3 ties.'' Also. death certificate occupa- tion and industry reponing varies Precision production, craft. and repair occupation [503-6991 Black females White females Black males White males 15 88 138 1,114 1.3 0.7-2.3 1.2 0.9-1.5 1.1 0.9-1.4 0.9 0.9-1 .o widely among population subgroups when compared to lifetime work his- tones. I* In addition. occupational histones were unavailable for this study. which may have led to nondifferential misclassification and biased risk esti- Farm equipment mechanic (517) White males 5 2.1 0.7-6.0 mates toward the null.!Y Because many comparisons were Eiectronic repair (523) White males made and chance associations may 12 2.1 1.1-4.1 have occurred, our finding should be viewed with a healthy dose of skepti- Carpet !nstaller (566) cism. The purpose of this analysis was White males 5 2.4 0.8-6.9 to generate clues to possible multiple Painter (579) White males myeloma-associated industries andi 23 0.5 0.4-0.a or occupations. which can be funher evaluated in analytic investigations. Plumber (585) Black males 5 2.2 0.8-6.3 By restricting OR reponing to OCCUpations and industries with five or 1214 Occupations Associated with Multiple Myeloma Figgs et a1 TABLE 24onfInued Mortality Odds Ratios, 95% Confidence Interval, and Cases Exposed for Multiple Myeloma, According to Occupation Code Occupation Cases Exposed' Odds Ratio 95% Confidence Interval Supervisor. extractive (613) White males a 2.7 1 .l-6.3 Productionsupervisor (633) White males 139 1.3 1 .l-1.5 Parternmaker (676) White males 5 2.3 0.8-6.7 Operator, fabncator. 8 repair occupation (703-8891 Black females White females Black males White males 78 405 371 1.018 1 .l 0.8-1.4 1 .o 0.9-1.2 0.9 0.8-1.1 0.9 0.8-0.9 Photoengraver(735) White males 6 4.0 1.4-11.7 Miscellaneousmachine operator (777) Black females 7 2.5 1 .O-6.4 Product inspector (769) Black females 5 3.7 1 .o-13.9 Bus driver (808) Black males 8 2.4 1 .O-5.6 Suoemsor, moving equipment (843) White males 5 2.5 0.9-7.6 Grader, dozer operator (855) White males 12 2.7 1.2-6.1 Constructionlaborer (869) White males 61 0.7 0.5-0.9 Stevedore (876) Black males Stock handler (877) White females 9 2.3 1 .O-5.1 14 1.5 0.8-2.7 Laborer (889) Black males 96 0.7 0.6-0.9 Homemaker 1914) Black females White females 364 2.562 0.8 0.7-0.9 0.9 Never employed (917) White females White males 20 0.4 0.3-0.7 23 0.4 0.2-0.6 ~~ +* Cases exposed = number of cases with that usual occupation. [ ] = 1980 census 3digit industry code. t ( ) = 1980 census 3digit occupation code. All estimates of grouped job categones are reported regardless of stabstrcal signmcance or number of deaths. more deaths and ORs 2 2.0. or which were statistically significant. we adopted a conservative approach that should help focus on the associations that are more likely to represent neu leads or confirm previous reports. It is also impomnt to remember that because these analyses were limited to decedent cases and controls. comparabiiity may present special limirations.''-`2 While we lacked important lifestyle information such as alcohol and tobacco use, these factors may nor be strongly associated with multiple myeloma and should not confound risk estimates."-" Teachers were the most consistentl>- elevated group. particularly eiementary teachers. This association. whch persisted across gender, race. and gecgraphic region, has been reported elsewhere, 14.:6.17 .Dut not for women separately. It does not appear to be srrictl!. a socioeconomic phenomenon. be- cause escesses still occurred when manager and professional codes (OC = 00 1-199) were specifically selected as unexposed occupations. Compari- sons with post-secondary teachers. however. eliminated the excess for all but whte females. Additional caution regarding es- cesses among teachers is necessaq. however. Since the population at risk is unknown. appropriate control seleaion becomes very ' important. Good health is a selection factor related to initial and continued emplo! ment.".'9 Stewart and Hunting3' demonstrated that both the PMR and OR may overestimate cause-specific mortality in workmg populations Nifj: strong healthy worker effects: teachers are generally healthier than other workers." When teachers were compared with managers and other proffisionals (OC = 001-199) in an effoc to control for SES. ORs were reduced. but excesses remained. A companwr. of elementary teachers (OC = 156 with post-secondary teachers !oc 113-153) found deficits for all gender groups. except white womer. (OR = 1.4). This suggested that th? association is more likely to be reiar:: to teaching than to simply having 2 white-collar job. Excesses were observed among s.s -. era1 occupations that used electn:z.- 1216 TABLE 24ontmued Industry Category [Group1 (industry specific) Ordnance (292) White males Electronic computer equipment (322) White males Other electrical machinery. equip rnent and supplies (342) White males Aircraft and pans (352) Black males Ship/boat building and repair (360) Black males Miscellaneous manufactunng industry (391) White females Not specified industry (392) Black females Transportation. communications, and other public utilities [400-472] Black females White females Black males White males Bus/urban transit (401) Black males Water transport (420) Black males Communications radio/TV broadcast (440) White females Wire/radio/phone(441) White males Water supply (470) BlacK males Wholesale trade Durable goods 1500-5321 White females Black males White males Electronic goods (512) White females Nondurable goods [540-5711 White females Cases Exposed' 14 6 51 5 5 20 10 7 95 122 488 9 12 5 38 6 19 3 78 5 21 Occupations Associated with Multiple Myeloma Figgs et a1 Mortality Odds Ratio 2.3 95% Confidence Intewal 1.2-4.3 3.7 1.3-10.8 1.6 1.2-2.2 3.1 1.O-9.6 2.1 0.7-5.9 1.8 1.1-2.9 2.2 1.0-4.6 0.9 1.4-2.1 1.2 0.9-1.5 1.2 1.o-1.5 0.9 0.8-1 .O 2.5 1.l-5.6 2.4 1.2-4.8 2.8 0.9-8.3 1.6 1.1-2.3 2.3 0.9-6.1 1.9 1.l-3.2 0.5 0.2-1.6 1.2 1.O-1.6 2.8 0.9-8.3 1.4 0.9-2.3 tion of the mortality difference between US blacks and whites found in other studies examining factors such as and SES.63Among 22 OCcupations in which black and white men have five or more cases (includes occupations not statistically signifi- cant and/or with ORs c 7.0). 15 OC- cupations showed black men with higher ORs than white men. six showed white men with higher ORs than blacks, and one occupation showed an equivalent association. Similarly. among 39 industries in which black and white men appeared in sufficient numbers for comparison (includes industries not statistically significant or with ORs c L O ) , 27 showed black men with hgher ORs than whte men, 1 1 show white men with higher ORs than black men, and one industry produced equal estimates for both groups. It is unlikely (x' goodness of fit: P c 0.005) that chance alone would account for this racial distribution among occupations or industries. Differences in exposure levels withm specific occupations may contribute to hgher rates for multiple myeloma deaths among blacks than whtes in the United States.'.@Health care access and quality might also contribute to US racial mortality differences. .4mOng industries with white male and female cases available in the same indusT. 77% showed male ORs hgher than female. whereas 45% showed female associations higher than male. Larger associations for women than men for occupational exposures have been reported elsewhere.'5 It may be that blacks and women must accept "riskier" tasks. within the same job. than whte men. Blacks and women could also be more susceptibleto occupational exposures. Understanding the occupational.' industnal associations and gender differences observed here involves a broader understanding of bioIooJC3! and sociological factors. .4larger proportion of the unexposed among women consisted of persons with limited industrial exposures (eg. house- wives). If women classified as u m posed truly lack exposure. it ma!' t.2 easier to detect associations among women than men in this data base. JOM Volume 36, Number 11, November 1994 TABLE 240ntinued Industry Category [Group] (Industry SpediC) Black males White males Electncal goads (512) White females Retail trade [580-691] Black females White females Black males White males Hardware stares (581) White females Department stares (591) Black females Grocenes (601) White females Bakeries (610) White females Miscellaneous vehicle dealers (622) White males Eating and dnnking (641) White females Jewelry (660) White females Finance, insurance. and real estate [700-7121 Black females White females Black males White males Security. commodity. brokerage and investment (710) White males Insurance (711 ) White females Real estate (712) White females Business and repair setvices [721-760] Black females White females Black males White males Dwelling/building service (722) White males Cams Exposed' 9 99 Odds Ratio 1.3 1 .l 5 2.8 47 1.2 429 1.1 51 1 .o 428 0.9 9 2.7 9 2.0 64 1.7 9 2.3 5 2.8 85 0.7 12 2.7 5 0.6 132 1.5 1 1 0.7 176 1.2 18 2.1 50 1.7 32 1.6 1 1 1.1 45 0.9 33 1 .o 159 0.9 6 0.4 950/. Confidence Interval 0.6-2.7 0.9-1.4 0.9-8.3 1 .o-1.2 1 .o-1.2 0.7-1.4 0.9-1.1 1.2-5.9 0.94.3 1.3-2.2 1.04.0 0.9-8.3 0.6-0.9 1.2-5.9 0.3-1.7 1.2-1.8 0.7-1.4 1 .O-1.4 1.2-3.7 1.2-2.3 1.1-2.3 0.6-2.2 0.6-1.2 0.7-1.5 0.7-1.O 0.2-0.9 1217 1218 Occupations Associated with Multiple Myeloma Figgs e t al - TABLE 3-Cr-ntinued Industry Category [Group1 (Industry specific) Computer/data processing (740) White males Cases Exposed' 7 Personal services [761-7911, Black females White females Black males White males 11 45 33 159 Barber shop (780) White males 25 Entertainmentand recreation [800-802] White females White males 14 43 Theater/film (800) White males 21 Professional and related services [812-8921 Black females White females Black males White males 181 760 139 668 Physicians' office (812) White females 22 Dentists' office (820) White females 8 Health services (840) Black females 10 E!emenrary and secondary school (842) Black females White females Black males White males 71 342 36 140 College and university (850) White males 57 Other social services (871) Black females 5 Religious organization (880) White females White males 25 56 Public administration [goo-9321 Black females White females Black males White males 26 109 46 255 Other general government (901) White males 69 O#dOdrtaswRatio 3.5 1.1 0.9 1.o 0.9 1.7 0.9 1.2 1.8 1.6 1.2 1.4 1.1 1.8 2.1 5.0 1.6 1.5 1.6 1.5 1.5 2.2 1.6 1.4 1.6 1.o 1.2 1.o 0.7 95% Confidence Interval 1.3-9.2 0.6-2.2 0.6-1.2 0.7-1.5 0.7-1 .O 1.l-2.6 0.5-1.7 0.8-1.6 1.1-2.9 1.3-1.9 1.o-1.1 1.1-1.7 1.o-1.2 1.l-2.8 0.9-4.8 2.1 -12.2 1.2-2.1 1.3-1.7 1.1-2.3 1.2-1.8 1.1-2.0 0.8-6.5 1.1-2.6 1.l-1.9 1.O-2.6 0.8-1.2 1.l-1.8 0.8-1.1 0.6-0.9 .. T-JOM Volume 36, Number 11, November 1994 TABLE 34ontinued Industry Category /Group1 flndustrv soecific) Cases Exposed' Mortality Odds Ratio 95% Confidence Interval Administrative human resources programs (922) Black females 8 2.3 1.O-5.2 Industry not reported (990) Black females White females Black males White males 16 62 40 117 * Cases exposed = number of cases in that industry. t [ ] = 1980 census 3-digit industry codes. $ ( ) = 1980 census 3-digit occupation code. All estimates of grouped job categories are reported regardless of statistical significance or number of deaths. 1219 TABLE 4 Multiple Myeloma Cases Exposed, Mortality Odds Ratios, and 9596 Confidence Interval for E!ementary Teachers (Occupational Code = 156)by Race/Gender Group and Geographic Region Northeast' Southeast Central West Gmup Casest Exposed 95% MOodmds'@ Conti- Ratio dence Interval Cases Exposed ORadtdios 9596 95x Confi- Cases Mom'iv confi- dence Exposed Odds dence &posed Interval Ratio Interval M0-W ORadtdios 95Yo Confr- dence Interval Black females White females Black males White males 1 27 0 7 1.8 0.1-23.0 28 1.9 1.2-2.0 +$ 67 7 1.9 0.8-4.8 19 1.6 1.0-2.4 9 1.6 1.2-2.1 102 1.2 0.5-2.8 3 1.7 1.0-2.9 27 1.9 0.9-4.1 0 1.5 1.2-1.9 28 2.0 0.5-8.2 0 1.7 1.1-2.7 11 $$ 1.2 0.80-1.3 $$ 2.4 1.10-5.: * Northeast = Maine. New Hampshire. New Jersey, Rhode Island, Vermont: Southeast = Georgia. Kentucky, North Carolina, South Carolina. Tennessee, West Virginia; Central = Indiana. Kansas, Missoun, Nebraska. Ohio. Oklahoma, Wisconsin; West = Colorado. Nevada, New Mexico. Utah. Washington. Idaho. 7 Cases exposed = number of cases with that usual occupation. $ Mortality odds ratlo not calculated. Miettinen and Wang' introduced deaths. and I b J all circulatory disease Conclusions 4 monaiity odds ratio analysis as an alternative to proponional mortality ratio analysis: the advantages of ORs are discussed there and elsewhere." Briefly. ORs are less influenced by over- or underrepresentation of other causes of death. and they contrast deaths among exposed and unexposed individuals. In addition. Mettinen deaths. The cardiovascular disease PMR among clementary teachers in our 24-state data base indicated no association with elemenrap teaching (PMR = 1.0). A comparison of controls with circulatory disease (ICD = 390-459) produced little difference from "all jobs" ORs but produced higher OR estimates than managerial These data provide a simple. cconomicsl surveillance tool for occupational cancer risks among women and men. For multiple myeloma. excesses occurred among elementary teachers and for several occupations with potential exposure to solvents. metals. and EMFf. and Wang emphasized selecting con- and professional controls. Therefore. References 4 4 trols from an auxiliaq disease not associated with the exposure of interest. To dddress this concern. we used some hedthy worker effect is present among cardiovascular disease controls, a major component of "ail jobs" I. Riedel DA. Pottern L,M.The epidemiology of multiple myeloma. Hrmafol Oncol Clin LV .4m. I!W2:6:X?S-247. two control groups: (0)ail noncancer effects. 2. US Department of Commerce. 1980 1220 Occupations Associated with Multiple Myeloma Figgs et al TABLE 5 Multiple Myeloma C a s e s Exposed, Mortality O d d s Ratios. and 95S0Confidence Interval for Elementary School Teachers (Occupational C o c e = 156)Using Different Control Groups for Comparison Control Groups Unrestricted Exposure Category Restricted Exposure (Job) Categories Restricted Disease Category All Jobs',? (Occupational Code = 001-999) Managers 8 Professionals',$ (OccupationalCode = 001- 199) Post-secondary Teachers', (OccupationalCode = 113-754) All Circulatory Diseases11 (ICD = 390-459) Race/Gender ~ Black females White females alack males White males 95% 95?& 95 70 95% Cases Mortality Confidence Mortality Confidence Mortality Confidence Mortality Confidence Exoosed Odds Ratio Interval Odds Ratio Interval Odds Ratio Interval Odds Ratio lnteval 38 1.6 1.1-2.1 1.3 0.7-2.2 0.2 0.4-1.O 1.3 0.8-2.2 224 1.4 1.2-1.7 1.3 1.o-1.5 1.4 0.6-3.5 1.5 1.2-1.8 10 1.8 0.8-3.9 1.6 0.6-3.9 0.0 0.0-inf 1.8 0.6-5.5 64 1.9 1.4-2.6 1.4 1.o-1.9 0.6 0.3-1.2 1.8 1.2-2.6 * 1980 census of population: Alphabetical Index of Industries and Occupations. US Department of Commerce. T All jobs = all job codes 007-999.occupational code = 156 $ Managers and professional = all job codes 001-199 'Managend and Professional Specialty OEupatiOnS", excluding OCCUDatiOnal codes = 156 5 Post-seconoary teachers = all job codes 113-154 "Teachers. postsecondary.' 1 IC3 = International Classification of Dneases. 9th rev 'Circulatory Diseases', IC0 = 390-459 Census qf rhe Popuiarion: .diphabetical Index of Industries and Ocmparions. Publication no. PHC 80-Rj. Waslung- ton, DC:CS Government Printing Of- fice: 1982. 3. Mieninen OS. Wang JD. An alternative to the proportionate morraliry ratio. .dm J Epidemioi. 1981:114144-118. 4. Preston DL. Lubin JH. Pierce DA. and McConney ME. EPICURE: Risk aggression axd data analysis software. Seattle. WX: HiroSoft International Corporation: 1993. 5 . Monson RR. .Analysisof relative survival and proponional mortality. Compur Biomed Res. 1974:7:325-332. 6. Dosemeci M. Helsel W. May C. O/E ARaivSiS Command Procedure for Use with rhe 1984-1 989 A4onalir.v Files: Users Guide C'ersion4.0. June 1997,. 7 . Petersen GR. Milham S Jr. iVIOSH Research Repon: Occuparional.Monaliry in (ne Siare of Cul[fornia. 1959-1961. DHEW (NIOSH) Publication no. 80104. Cincinnati. OH: National Institute for Occupational Safety and Health: 1980. 8. .Milham S Jr. OccuparionaiMorraliry in Washington Siare. 1950-1979. DHEW (NIOSH)Publication no. 83-1 16. Cincinnati. OH: National Institute for Occupational Safety and Health: 1983. 9. Dubrow R. Sestito JP. Lalich NR. Bur- nett CA. Saleg JA. Death certificatebared occupational mortality surveillance in the United States. .dm J Ind 'Wed. 198:: 1 1:329-347. IO. Loomis DP. Saviu DA. Occupation and leukemia mortality among men in 16 staIes: 1985-1987..4m J/nd.Med. 1991; 19:509-S I. 11. Selikotf U. Use of death cenificates in epidemiological studies. including occupational hazards discordance with ciini d and autopsy findmgs. .4m J IndMed. 1997,:?1:469-480. I? Selkoff U. Death cenificates in epidemiological studies. including occupational hazards: inaccuracies in occupational categories. .4m J Ind Med. 1992:2?:493-504. 13. Selikoff U. Seidman H. Use of death cenifiuzes in epidemiological studies. including occupational hazards: variations in discordance of different asbesto.s-associated diswes on best evidence ascertainment. Am J Ind &fed. 1997,:7,2:481497. 14 Dubrow R, Wegmen DH. Cancer and occupation in Massachusetts: a death certificate study. .4m J Ind Med. 1984: 6:207-330. 15 Percy C. Stanek E. Glwckler L. Accuncy of cancer death cenificates and irs effect on cancer mortality statistics. .dm J Pidiic Heairh. 1981: I :242-250. 16. Steenland K.Seaumont J . The accuracy ofoccupation and indusry data on death xrtificates. J OCCI~.IlFed. I Y84:26:258296. 1.: Burnett C.A. Measuring work-related nealth disparities for minority popula:ions. In: Procctdinrs qt-rhe 1993 Pithlic Flculrii C ~ w c ~ r c vown Records and Sia1'isricx Towud., ihc Y u r JOOO. DHHS Publicarion no. 94- 12 14. Hyattsville. MD: Zational Center for Health Statis- tics: 1994. 18. Schade WJ. Swanson GM. Comparison of death ctrtificate occuparion and in- d u n n data with lifetime occupational histones obtained by interview: vana- tions in the accuracy of death certificate entries. . ~ J JV Ind .Wed. 1988:14:121- 136. 19. Checkowa! H. Pearce NE. Crawford- Broun DJ. Research .Wehods in OCCII- paiionai E?iu'emioiog~..Yew York: Ox- ford Cniversiry Press: 1989. 70. Gordis L. Should dead cases be matched to dead controls? .-fm J Epia'etnroi. 1982:115:1-5. 31. Mchughiin JK. Blot WJ. Mhel ES. Mandel JS. Problems in the use o i dead controls in case-control studies: I. Gen- en1 results. .4m J Epidemiol. 1985: 12;: 131-139. ?--?. Mchughlin JK. Blot WJ. Mhel E S . .Mandel JS. Problems in the use of de3d controls in cz-control studies: II. EfTecec: of excluaine certain causes of death. J Epidernioi. 1985: i 21:485-494. 23. Brown LM.Gibson R. Burmeister LF. Schuman L.M. Everett GD. Blair .A. Ai- coho1 consumption and risk of leukeml;. non-Hod-gkin's lymphoma. and multlpl: myeloma. L ~ ~ i iRkcs. 1991:16:979-9~; 24. Brown L.M. Everett GD. Gibson R. %J- meister LF. Schuman L.M. Blair .L Smoking and risks of non-Hodekin 1) m- phoma and multiple myeloma. Cdnc'c'' Causes Cvnrrril. 1992:3:49-55. 75 -d. Heineman EF. Z t h m SH. McLaugh]lr J K Vaughr JB. Hrubec Z. A prospeC!l\L' JOM Volume 36, Number 11, November 1994 1221 study of tobacco use and multiple myeloma: evidence against an association. Cuncw CJrrscs Conrrd. I993:j:j 1-36, 26. Heineman EF. Olsen JH. Pottern LM. Gomez bl. Ratfn E. Blair A. Occupational risk A C ~ OfKor multiple myeloma among Danish men. Cancer Carires COn!TrJ/. 1992:3:555-568. 37. Boifetta P. Stellman SD. Garfinkel L. A cax-control study of multiple myeloma naied in the .Amencan Cancer Society Prospective Study. Inr J Cmcer. 1989: 13:55&559. 25. .Monson RR. Observations on the healthy worker effect. J Occiip .Wed. 1986:3135-433. 29. Mc.L.tichae1.AJ. Standardized mortality ratios and the *healthy worker effect": scratching beneaththe surface. J Occrip .Wed. 1976:18:165-I 68. 30. Stewan W. Hunting K. b l o d i t y odds ratio. proponionate mortality ratio. and hedthy worker effect. .-lmJ Ind .Wed. 1985: I4:35-3 5 5. 5 1. Olsen JH. Jensen OM.Occupation 3nd risk ofcancer in Denmark. Scand J I t ' d Environ Health. 1987:13:7-89. 32. Demers P. Vaughn T. Koepsell T. et d. .A cse-control scud? of multiple myeloma and occupation. .-im J Ind .bled. 1994 (in press). 33. London SJ. Thomas DC. Bowman JD. Sobel E. Cheng T-C. Peters JM. Exposure to residential electric and magnetic fields 3nd risk of childhood leukemia. .AmJ Epiuemrol. 1991:134:913-937. 34. Cdeman M. Berd V. .A review of the epidcmiolopcai studies of the health effects of living near or worlung with electncrty generation and transmission equipment. In! J Eprdrrnroi. 1985: 17: 1-13. 35. Ahlbom .A. A review of the epidemiologic iitenture on magneric fields and cancer. Scand J W'urk Environ Hrultlr. 1985: !1:337-343. 36. Saviu DA. Calle EE. Leukemia 3nd occupationai exposure to electromagnetic tie!ds: review of epidemiologcal surveys. J O C C.W~e~d. I98729:17-5 I . 37. Tomenius L. 50-Hz electromagnetic environmcnr and the incidence of childhood tumors in Stockholm County. Biurirc!r~i,nuqnrricr.1986:T: 19 1-107. 5s. Wenheimer N. Leper E. Ele~tricalwiring configurations and childhood cancer. . i m J Ep~detnrol.1979: 109:273-3YJ. 39. Milham S Jr. Mortality in workers ex- -posed to elearornag~eticfields. Environ Hrallh Perspecr. 1985 :62 2 87 300. u). Pearce NE. Reif JS. Fraser J. CYe-con- trol studies of u n c e r in Xew Zedand electrical workers. Inr J Epiderniol. 1989:18:55-59. 41. Eriksson M. KYlsson M. Occupational and other environmental facron and multiple myeloma: a population-based use-control study. Br J Ind .Wed. 199249155- 103. 42. Semendw RM..Momson HI, fiedel 0. Wiiluns K .Multiple myeloma mortality and agnculture practices in prairie provinces in Canada J Ocnrp Med. 1993;35:557-56 1. 43. Linet MS.Harlow SD. McLaughlin JK. A cue-control study of multiple myeloma in whim: chronic antigenic stim- ulation. occupation. and drug use. Cancer Res. 1987;47:2978-298 1. 44. DeCouile P. Blattner WX. Blair A. Mortaliry among chemical workers exposed to benzene and other agents. Environ Res. 1983:jO: 16-25. 45. Devine BJ. Barron V, Kaplan SD. Tcxaco monaliry study: I. Morrality among refinery, petrochemical, and research workers. J Ocrzp .Wed 1985:17:445347. 46. Kaplan SD. Update ofa mortality study of workers in perroleurn retinenes. J OC- cup .Wed. 198625:514-5l6. 47. Wong 0, Raabe GK. Critical review of cancer epidemiology in petroleum industry e m p l o y e s with 3 quantitative mcta- analysis by u n c e r site. .-lm J /nd .Wed. 1989:!5:253-310. 18. Pottern LM. Heineman EF. Olson JH. RaEn E, Blair A . Multiple myeloma among Danish women: employment history and workplace exposures. Cancer Causes Conrroi. 1997-:3:427-432. 49. Mason TJ. Cancer mortality in US counties with plasrics and related indusmes. Environ Health Perspecr. 1975:l 1:79-84. 50. Berthwaite PB. Ptxce N. Fnser J. Cancer risk in gainten: study based on the New Zedand Cancer Regstry. Br J Ind .Wed. 1990:17:742-746. 5 I . Burmeisrer LF. Cancer in Iowa farmers: recent results. .4m J Ind .bled. 1990:15:295-301. 52. Pearce NE. Smith AH. Howard JK. Sheppard RA. Gila HJ. Teague C.A. Casexonuol srudy of multiple myeloma and farming. Br J Cancer. 198654493- 500. 53. Pearce NE. Smith AH. Fisher DO. Malignant lymphoma and muitiple myeloma linked with agricultunl occupations in a New Zealand cancer registrybased nudy. .dm J Epiderniol. IY85: 121:225-237. 54. Gilrs GG. L i c k s JN, Balkie ,MJ, Low- enthd kM, Panton J. Myeioprolifentive and iymphoproliferative disorders in Tasmania 1972-1980: occupational and familial aspects. J .Vad Cancer Insr. I98&72: 123;- 1240. 55. Cantor K P , Blair A. Farming and mortality from multiple myeloma: a casecontrol study with the use of death cerukicates. J AVarl Cancer Insr. 1984: 7'1:251 - 5 5 . 56. Gallagher RP. Threlfdl WJ. Cancer mortality in metal workers. Can .Wed .-lssoc J. 1983:129:1191-1194. 57. Flodin U. Fredrikwon M. Penson B. Multiple myeloma and engine exhausts. fresh wood. 3nd creosote: 3 case-referent srudy. .lm J Ind Med. 1987:12:5 19-29. 58. Hansen ES. A mortality study ofDanish stokers. Br J Ind .\fed. 1992:49:48-52. 59. Schneider D. Greenberg .MR.Wartenberg D, Kipen H. The study ofcancer in minorities. .V J J f e d . 1991;58:?13-719. 60. Cooper R. Steinhauer M. Miller W. David R. Schaakin .A. Radsm. society. and &sese: an exploration of the social and biolopcal mechanisms of differential morrality. Inr J Health Sew. 198 1: 11:589114 6 1. Mancuso TF. Lung cancer among black migrants. Intenction of host and OCCUpationd environment fac:on. J O C ~ Z . ~ .Wed. i 977: !9:j 3 1-53?. 63. 4lancuso TF. Sterling TD. Lung czncer among Slack and white migrants in the U S . J .Varl.Wed.issuc. 1975:67:106-1 I t . 65. ,McWhoner WP. Scharzkm .AG. Horm JW, Brown CC. Cmtnbuuon of socioeconomic sutus to black/white diiferences in cancer incidence. Cancrr. 1989:53:983-987. 64. b e s LAG. Hankev BF, .Miller BA. Hanman .AM. Edwards BK. Cmcer Srarrsricr Rrvrew 1973-59. NIH Publication no. 9 1-2789. SeIhesda. MD: National Cancer Institute: 199I. 65. S p i m R. Stewart P.A. Lce JS. et 31. Retrospective cohon mortality study of workers ;11 an a i r c n ~maintenance facility. I. Epidemiologrcd re~ults.Br J I d .Wed 199l:18:515-530.