Document b5bBEyDNmZwvEOdr94kpVQEB0

X I Mortality Among Rubber Workers I 11. Cause-Specific Mortality, 1940-1978 I i i Elizabeth Delzell, S.D., a d Richard R. Mown, M.D. This report describes mortality occurring between 7940 dation, there were excess deaths from several cancers. andmid-7978 among29,087 menandwomen employedin The present study describes mortality through June 30, a rubber plant forat least two years. Mortalitypatterns for 1978, among these workers. In addition, men and women the period)uly 1, 7974, to )uly 7, 1978, were comparedto employed for two to four years are included in order to previouslypublishedfindings forjanuary 1, 1940, through determine to what extent their mortality experience lune 30,1974. Expectednumbers of deaths were basedon resembles that of long-term workers. U.S.general population mortality data. There were excess deaths from bladder cancer and leukemia among white Materials a d Methods male union members during both follow-up periods. Dur- Study subjects worked at a rubber manufacturingcom- ing recent follow-up of white male union memben em pany that has operated in Akron, Ohio, since the late ployedfor at least five years, there were excesses indeaths 1800s. The company's personnel records date from 1900 from three additional cancers: esophageal cancer (77 o b and contain information on name, demographic charac- send'4.8 expected), biliary and liver cancer (6 observed! teristics and employmentdates. Productionworkers have 3.3 expected) and /ymphoma and multiple myeloma 114 been members of the United Rubber, Cork, Linoleumand observecU5.8 exp&$l. Evidence from other studies of Plastics Workers of America (URW) Union since the late rubber workers suggests that observed excesses in deaths 1930s. URW records on members consist of name, social -from bladder cancer and leukemia are related to workdings suggest that occum re/evant to these disaas I recent years of that suff "cient time has not elapsed for such reductionsto result in ' decreased mortality. *Further inwtigation is required to clarify the contribution of occupational factors to ob served excesses in deaths from cancers of the esophagus Because numbers were small, andthe biliary passagesand liver andfrom lymphomaand non-white salaried males were also not included. The multiple myeloma. total number of eligible employees was 29,087, of whom 5,302 had worked for two to four years before July 1, 1971. and were not included in previous studies of these I workers. n1976, Monson and Nakano' reportedthe mortality Five study groups were defined on the basis of employ- experience of a large group of workers who had worked ment catego for five or more years in rubber manufacturing and who and were followed from January 1, 1940, through June 30, had 1974. Although there was no increase in overall mcitaky were 1 ~ apong rubber workers comparedto the U.S. general pop white male (NWM) union members, 4,200 female union From the Departmentof Epidemdogy,Harvard School of Public Health, 677 Huntinflon Avc, Boston, MA On15 members, 4,683 salaried males and 3,377 salaried fe males. supported bv a contract with the B.F. Goodrich Company and the United Rubber, Cork, Linoleum and PlasticsWorkers of America. Dr. Delrell was also supported b~ a trainmhip from the Natlonrl institute of acupatmnal k f e t y and Health(5 T15 O H 0 70961 Company censuses in 1974 and 1979 identified 4,470 active employ- among the 29,087 study subjects-Exact dates of termination were not obtained for theseworkers, Journal of OccupationalMedicinaNol. 23,No. 10/0ctober 1981 677 - ~ and duration of employment as of the close of the study on July1,1978, was estimated as follows. Workers active in1971 but not in 1974 were assumed to have terminated employment on January 1, 1973, while those active in 1974 but not in1979were assumedto haveterminatedon July1,1976. Workers who were active in1971 or 1974 but who died before the designated termination date were assumed to have left work on the date of death. Follow-up to ascertain mortality and to measure person-years of observation started on January1, 7940, or the second anniversary after startingwork, whichever was later, and continued to July1,1978. The company provided death recordsfor active and pensionedworkers, and a Social Security Administration (SSA) search identified deaths among workers not known to be deceased. The company and the SSA reported a total of 9,388 deaths during the period January1,1940, through July 1,1978 2,533 of these were not included in previously published reportsconcerningthese rubber workers.' Therewere no death, disability or retirement claims or reportedearnings for approximately 16% of the 26,232 subjects included in the SSA search. All subjects not known to have died before July 1, 1978, were considered to be alive on that date. Death certificates for 444 (4.7%) of the reported deaths could not be located. The percentage of reported deaths with unavailable certificates varied by employ- ment group (4.0% for WM union members to 9.5% for salaried females), by duration of employment (1.4% for workers employed 15 or more years to 18.6% for those employed two to four years) and by age at death (2.9% for deaths at and above age 65 to 18.7% for those under 40 years of age at death). Deaths among workers employed two through four years and deaths under age 40 accounted for 11% and 3% of the total deaths, r e spectively. Decedentsfor whom a death certificate could not be located were assumed to have died of unknown cause on the date reported. Underlying cause of death was coded from available death certificates according to the International Classificationof Diseases, 7th Revision. For each employment group, observed numbers of deaths were compared.to numbers expected based on relevant U.S. mortality rates specific for cause, age, calendar time, sex and race. (Cause-specific mortality rates for Ohio are virtually identical to those for the United States'; the U.S. rates were used because they are available in greater detail.) To compute expected numbers, person-years of observation were allocated to five-year age and calendar time categories and multiplied by corresponding U.S. death rates.' The resultingquantitieswere summed over all ages and years to obtain total expected numbers. Ratios of ob served to expected numbers of deaths were expressed as standardized mortality ratios (SMRs), which are estimates in percentageform of rate ratios for rubber workers c o m paredto the US. general population, adjustedfor age and calendar time to the person-years distribution of the e m ployment group of interest. Confidence intervals were derived assumingthat observed numbers of death follow a Poisson distribution.' Mortality patterns for the interval covered in previous reports, )anuary 1,1940,through June30,1974, were corn pared to those for the period July1,1974, through July1, 1978. These intervals will be referred to as the 1940 through 1974 and the 1974 through 1978 follow-up perk ods, respectively. Because no new hirees after mid-1969 were includedinthestudy,the age distributionof personyears for 1974 through 1978 shifted toward older ages relative to the distribution for 1940 through 1974, and comparisons of SMRs betwen *& two follow-up periods could be misleading for causes of death associated with age. To address this issue, directly age standardized mortality rates were computed for all causes and all cancers for each employment group and follow-up perk od,using as a common standard the person-years distribution of all employees during the years 1940 through 1978. Also, for comparisons of site-specific cancer mor- tality in WM union members between the two follow-up periods, age standardized rate ratios were computed using the person-years distribution of men followed during the period 1974 through 1978. Workers employed for two through four years tended to be younger than those employed for five or more years. To facilitate comparisons of overall mortality b e tween the two duration of employment categories, all- cause death rates were standardized to the person-years age distribution of all employees followed during 1940 through 1978. Results There were approximately 441,985 person-years of o b servationfor WM union members, 28,729 for NWM union members, 134,397 for female union members, 110,987 for salaried males and 82,012 for salaried females. Average lengthof follow-upwas 28 years for WM unionmembers, 32 years for union females and 24 years for NWM union members, salaried males and salaried females. During 1940 through 1978 WM union members experk enced excesses in deaths from cancers of the biliary passagesand liver (12% excess) and of the bladder (17%) and from leukemia (21%) (Table 1). There were marked deficits in deaths from causes other than cancer and vascular lesions of the central nervous system. Bladder ' cancer and leukemia mortalitywere elevated during both follow-up periods. In addition, there were excess deaths from several cancers during 1974 through 1978 only, including cancers of the esophagus, large intestine, biliary system and liver, brainand lymphatic tissues and multiple myeloma. Inthe lymphatic cancer and multiple myeloma category (ICD codes 200-203), excess mortality was con- fined to ICD codes 202-203("other lymphoma"and multiple myeloma: 16 observd6.8 expected deaths). Examination separately of SMRs for men employed two through four years and those employed at least five years showed that with the exceptions of lungcancer and lymphatic cancer and multiple myeloma, cancer excesses occurred only among men who worked for at least five years. Men employed for two through four years experiencedslight excesses in lungcancer (16 observedll3.7 expected)and in lymphatic cancer and multiple myeloma (3 observed/l.9 expected) during 1974 through 1978. Table 2 gives age-standardized rates and rate ratios (RR) for selected cancers among menemployedfor five or more years. Several patternsfor causes of death included in this table are noteworthy: 678 Mortality Among Rubber Workers/Delzell and Monson * i & i ! t f c t ! i t -Table 1. Mortality Amonam-kale Union Members According to Causa a d Fdlow-Up Mod. All causes Malignant neoplasms (140-205) Digestive organs and peritoneum (150-159) Esophagus(150) Stomach (151) Large intestine(153) Biliary passages and liver (155, 156) Pancreas (157) Lung (162. 163) Prostate (177) Bladder (181) Brain (193) Lymphatic and multiple myeloma (200-203)l Leukemia (204) Vascular lesions of the central nervous system (330-334) Diseases of the circulatory system (400-468) Diseases of the respiratory system (470-527) Diseases of the digestive system (530-587) Accidents, poisoning and violence (800-998) Other known causes Deaths without certificates(W total deaths) 1940-1974 SMRt O b d 1974-1918 SMR obr SUR 86 94 101 88 103 103 104 89 91 103 117 82 "P"w 7 e 3 2 2 .t 97 86 70 82 64 61 n=213 5694 90 1103 87 1093 102 259 96 396 113 78 103 25 191 11 105 98 107 12 103 115 115 29 105 33 179 7 112 58 64 9 84 275 81 67 88 95 101 26 102 49 115 11 117 2224 147 Ir 54 M* 183 7 S l P 91 m.Ol109 57 %*117 f 11 121 531 89 100 96 2647 78 463 85 246 88 82 74 251 96 43 83 328 62 32 64 385 77 63 63 3.7% n=61 5.5% n=274 1940-1978 Obr MI C16 -6797 8 5 - 89 1352 91 101 -474 94- 113 36 73 145 110 8 5 - 124 -144 8 9 - 124 40 80 152 67 65- 107 342 79- 98 121 8 5 - 122 60 8 9 - 150 31 62- 130 -76 86- 137 68 94 153 631 88- 104 3110 8 2 - 88 328 66- 82 -294 74 93 -360 57 71 448 4.0% . 'International Classification of Diseases, 7th Revision tStandardized mortality ratio: 100 x (observed no. of deaths + expected no. of deaths). Expected numbers are based on age-, time-, and cause-specific death rates for U.S. white males -Sobs observed number of deaths -gCl confidence interval of SMR + Q L D.ut Y , & L T & ~ & N W * = 2 vse,79 1Codes 200, 202 and 203 are not included before 1950 1. Esophageal cancer rates in W M union members tri- us.Pled between the two follow-up periods, while rates for white males remained essentially unchanged. 2. Large intestine cancer rates rose only slightly for both W M union members and US. white males. 3. Bladder cancer and leukemia rates among W M union members did not change appreciably, nor were there marked trends in these rates for US.white males. 4. Other lymphoma a d *multiple myeloma mortality n* rates nearly doubled among US.white males adquad- rupled among W M union members,for whom there was a greater than twofold excess in deaths from these diseases during the period 1974 through 1978. Further examination of mortality among W M union members who worked for at least five years and died in I 1974 through 1978 showed that the observed excesses in -Table 2. Morbllty from Selected Cancers Among White Male Union Members Employed for at Least Five Years, ' According to Follow-Up hrlod. burr of lkrth FOHOW-UPPlflod 1940-1974 Obr' mt RRS 1974-1978 Ob$ m RR Esophageal cancer 23 1.3 83 11 3.8 228 1.65 1.75 Large intestinecancer 108 8.6 128 27 9.2 124 6.75 7.55 Biliary and liver cancer 27 1.7 79 6 2.0 180 2.25 1.15 Bladder cancer 46 3.8 136 11 3.8 131 Leukemia n--.--.T.(32..489 ~ 152 11 2.95 3.8 2.29 2.85 Other lymphoma and multiple myeloma1 18 1.2 to9 14 4.8 242 I 1.15 2.05 ~~ *S5s@rvcclnumber of deaths t h t e (x 10,000 person-years). Rates for 1940-1974 are adjusted for age according to the distribution of men employed five or more years and followed 1974-1978 +Ratio (x 100)of standardized rate in rubber workers and standardized rate in U.S. white males These numbers are rates (x 10,000person-years) tor U.S.white males, age standardizedto the distribution of men employedfive or moreyears and followed 1974-1978 1Only deaths after 1949 are included Journal of Occupational Medicine/Vol. 23,No. 10/0ctober 1981 679 -Tabla 3. MorblHy Amon- peUs ing to clusc and Follow-up Rriod. CIUSm Of Ikrth(IC0 NO.) All causes Malignant neoplasms (140-205.) Mgesthre organs and peritoneum (150-159) Esophagus (150) Stomach (151) Large intestine (153) Biliary passages and liver (155. 156) Pancreas (157) Lung (162, 163) Bladder (181) Lymphatic and multiple myeloma (200-203)t Vascular lesions of the central nervous system (330-3341 ~seasesof the circulatory system (400-468) . Diseases of the respiratory system (470-527) Diseases of the digestive system (530-587) Accidents, poisoning and violence (800-998) Other known causes Deaths without certificates (% total deaths) 1940-1974 SMR* ob8 65 77 82 81 80 55 154 123 * 83 251 45 17 3 5 2 3 4 13 79 106 38 8 28 5 55 28 46 29 n=8 3.2% wkrr-up pr# 1974-1978 lero-lsn SMR Ob8 SUR ob, 95% CI 82 72 -68 323 61 76 91 18 80 63 62-103 121 ? 90 24 58-135 81 1 81 4 22-20? 0 0 66 5 21-151 -155 2 81 4 22-209 249 1 170 4 46 436 272 3 160 7 6 4- 3 3 1 46 3 72 16 41 - 117 56 1 1 - 3 1 ~ -9 171 6 62 372 -65 31 44 92 82 26 80 132 6 7 - 94 85 4 46 12 24- 81 -116 5 45 10 21 82 -25 2 51 30 35 73 19 2 42 31 n=6 8.3% 11114 4.0% 'Expected numbers for SMRs are based on age-, calendar time-, and cause-specific death rates for U.S. non-white males tCodes 200, 202. and 203 are not included before 1950 esophageal cancer and other lymphoma and myeloma the following divisions (observed deathdexpected deaths concentrated among men who started working between based on rates in U.S. white males): aerospace (512.6).ie I w a i m (W.nand hose and belt (7/3.0). Esophageal cancer was elevated in men who worked for at least five years in * rubber reclaim (510.9). cei? 11 observedl3.8 expected; p Mortalityfrom all causes amongNWMunion (Table 3) was 32% lower than expetteddur ttiple myeloma:14 observedl4.7 expected).Preliminary study period. Although mortality from analysis of mortality by work area for all WM union mem causes wils also lower than expected, bers revealed excesses in other lymphoma and multiple during both fo myeloma among men employed for five or more years in the biliary-p I -Tabla 4. Morblity Amonp Famla Union Ismbars Accordinp to Causa and Follow-Up Period. 1 I Caurm of Oath (IC0 No.) All causes Malignant neoplasms (140-205) Digestive organs and peritoneum (150-159) Stomach (151) Large intestine (153) Biliary passages and liver (155, 156) Pancreas (157) Lung (162. 163) -Breast (170) Uterus (171 174)t Cervix (171) Corpus (172) Bladder (181) Erain (193) Lymphaticand multiple myeloma (200-203)t Leukemia (204) Vascular lesions of the central nervous system (330-334) Diseases of the circulatory system (440-468) Diseases of the respiratory system (470-527) Diseases of the digestive system (530-587) Accidents, poisoning and violence (800-998) Other known causes Deaths without certificates (W total deaths) 1940-1974 SMR' 8 Ob8 79 730 82 205 76 52 74 8 92 26 106 8 38 4 97 16 84 47 99 30 152 21 20 2 157 5 102 6 88 10 87 7 70 69 73 250 72 23 74 32 81 38 56 66 n-47 6.4% FoWow-Up MDd 1974-1978 SMR Ob8 SMR 85 57 43 90 36 70 0 26 75 181 123 223 0 223 30 49 81 76 125 103 113 65 n=27 233 36 8 2 3 1 0 2 9 7 2 5 0 3 1 1 28 92 16 10 9 15 11.6% 80 77 69 77 79 100 28 75 83 108 149 58 119 124 74 79 73 74 87 79 86 58 n=?4 1940-1978 Ob8 963 241 60 10 29 9 4 18 56 37 23 7 5 9 11 8 97 342 39 42 47 81 7.6% 95% CI -75 85 --68 87 52 88 3 4 - 142 5 3- 114 46-190 8 - 73 -4 4 - 1 1 8 62 107 -78- 149 95 224 2 3 - 120 -3 8 - 278 57 236 3 7 - 133 -34- 156 59 89 66- 82 -62- 119 57 107 62-112 1 *Expected numbers for SMRs are based on age-, calendar time-. and cause-specific death rates for U.S. white females tCodes 171. 200, 202 and 203 are not included betore 1950 580 MortalityAmong Rubber Workers/Delzell and Monson -~ Tabk 5. hlorblhy Among&brbd Mdos According to huro and Fo#anr-Up Period. Causa of Dwth (IC0 No.) All causes Malignant neoplasms (140-205) Digestive organs and psritoneum(150.159) Esophagus (150) Stomach (151) Large intestine (153) Biliary passages and liver (155, 156) Pancreas (157) Lung (162. 163) Bladder (181) Brain (193) Lymphatic and multiple myeloma (200-203)t Leukemia (204) Vascular lesions of the central nervous system (330-334) Diseases of the circulatory system (440-468) Diseases of the respiratory system (470-527) Diseases of the digestive system (530-587) Accidents, poisoning and violence (800-998) Other known causes Deaths without certificates(% total deaths) 1940-1874 SMR' Ob8 75 . 77 72 76 77 68 18 73 81 120 97 110 99 75 71 51 51 36 58 n=47 866 167 51 4 13 14 1 9 48 9 6 13 9 73 406 33 31 41 68 5.3% m - u p )wlrd 1874-1978 SIR Ob, SYR 69 59 101 146 39 105 337 31 50 49 74 67 45 64 66 128 25 39 61 n=12 198 35 16 2 1 6 3 1 10 1 1 2 1 8l58 26 3 7 12 6.1% 70 72 77 90 72 76 61 64 73 195 93 101 88 73 70 69 47 36 48 n=59 1940-1978 Ob8 M % C I 1064 202 67 6 14 20 4 10 58 10 7 15 10 88 494 59 34 48 80 5.5% -65- 74 63 83 60- 98 33-196 39- 121 47- 118 315616--- 115 119 95 5 0 - 193 3 7 - 191 57- 167 -42 58 64 53 ---- 162 90 76 89 33 27 - 66 48 'Expected numbers for SMRs are based on age-, calendar time-. and cause-specificdeath rates for U.S. white males tCodes 200, 202 and 203 are not included before 1950 "* tissues and multiple myeloma.omphosarcoma, reticule iarcoma and Hodgkin's disease were responsiblefor the ortaiity in the lymphatic candeKa category. There were no leukemia deaths, while two were expected during the overall study period. The 21% excess in digestivecancer deaths for the period1974 through 1978 was accounted for by cancers of the large intestine, biliary passages and liver and pancreas. Among female union members(Table41, there were ex- cesses in deaths from cancer of the cervix, bladder and brain during 1940 through 1978. For most other causes, there were deficits in observed deaths, with SMRs ranging Trom 28 for pancreas cancer to 100 for cancer of the biliary passages and liver. Excesses in bladder and brain "cancers were based on five and nine observed deaths, respectively.The increase in bladder cancer mortalitywas seen only in the years 7940 through 1974, while the brain cancer excess occurred only in the years 1974 through 1978. There was no excess mortality from respiratory cancer, even among women who worked for at least 20 years (6 observd8.8 expected). Excess cervical cancer deaths occurred during both follow-up periods. Further analysis of the cervical cancer deaths showed that elevated mortality was confined to women who were -fable 6. Mortslity Among Salaried Females According to Cause and Follow-Up Period. C8uo of Death (IC0 No.) All causes Malignant neoplasms (140-205) Digestive organs and peritoneum (150-159) Lung (162. 163) Breast (170) Uterus (171-174)t Cervix (171) Corpus (172) Bladder (181) Brain (193) Lymphatic and multiple myeloma (200-203)t Leukemia (204) Vascular lesions of the central nenrws system (330-334) Diseases of the circulatory system (440-468) Diseases of the respiratorysystem (470-527) Diseases of the digestive system (530-587) Accidents. poisoning and violence (800-998) Other known causes Deaths without certificates (H total deaths) 1940-1974 SMR' Obs 75 192 89 65 106 17 74 4 90 16 45 4 20 1 39 1 282 2 135 3 161 6 3a 1 68 15 61 47 55 5 66 9 62 14 49 18 n-19 9.9% FOMW-Up mod 1974- 1978 SMR Ob8 SMR 56 49 70 51 13 79 83 5 100 0 0 45 100 6 93 0 0 47 0 0 17 0 0 30 0 0 200 0 0 103 78 1 139 0 0 29 58 5 65 51 16 58 26 1 46 23 1 56 115 6 72 36 3 46 n=4 0.2% n=23 1940- 1978 Ob8 95% C I -241 62 a0 -78 23 62 63 - 98 150 -4 12- 115 22 58 140 5 1 5 - 110 1 0 - 94 1 2 0 23 --176383 -3 7 1 21 56 1 -- 300 287 162 -20 39 100 63 45- 74 6 17-100 10 27- 103 20 44- 112 21 9.5% 'Expected numbors tor SMRs are based on age-. calendar HIM-. and cause-specificdeath rates for U.S.whlte females tCodeS 171, 200. 202 and 203 are not included before 1950 Journal of Occupational MedicineNol. 23,No. 10/0ctober 1981 681 i - -T I M 7. MOf'bIW froRl All CWWS and from All 0- AccordlRg tO ER@wBrolrp a d Fdlorr-Up W. EJwh=m em White male union members All causes ' All cancers Non-whlte male union members All causes All cancers Female union members All causes All cancers salaried males Al causes All cancers Salaried females All causes All cancers wkrr-up mhd 1W-1974 1914-1971) sb- Ob5 ClpdrMI* mt m Crrdrrn. sbmfdlmd nata 5694 13.8 (0.2)$ 1103 29.1 (0.6) 1093 2.6 (0.1) 259 6.8 e.31 251 10.0 14.7 (1.0) 44 1.7 2.5 (0.4) 72 20.1 15.7 (2.0) 19 5.3 3.8 (0.8) 730 6.0 6.7 (0.3) 233 17.9 6.1 (0.5) 205 1.I 1.8 (0.1) 36 2.7 1.1 (0.2) 066 9.o 11.3 (0.4) 198 13.3 10.0 (0.7) 167 1.7 2.2 (0.2) 35 2.4 1.8 (0.3) 192 2.8 6.5 (0.6) 65 0.9 1.9 (0.3) 49 13.9 13 1.o 4.3 (0.6) 1.1 (0.3) 'Crude annual rate (x 1.OOO person-years) tAnnual mortality rate (x 1,000 person-years), standardizedfor age to the person-years dlstribution of all employeesfollowed 1940-1978 $Numbers in parenthesesare standard errors (x 1,000 person-years) of the standardized rates employed for at least15 years (16 observedl6.7 expected), who died 25 or more years after starting work (15 o b servedP.6 expected) and who started work before 1940 (16 observedl8.1 expected). The uterine cancer SMR for union females increased from 99 in the period 1940 through 1974 to 181 from 1974 to 1978. This change resulted from a marked decline in age-adjusted uterine cancer death rates for U.S. white females from 5.7 per l0,OOO person-years in 1940 through 1974 to 2.9 per 10,OOOperson-years in 1975 through 1978,' while similarly age-adjusted rates for union females changed only slightly from 5.9 to 5.2 per l0,OOOperson-yearsduring the same intervals. Compared to US.white males, salaried males (Table 5) experienced no increased mortality during the overall study period. For theperiod1974 through1978 there were excesses in deaths from esophageal and biliary and liver, cancer, based on two and three deaths, respectively, and a 28% increase in deaths from diseases of the respiratory system. For salaried females (Table 61, .there were excesses, based on small numbers, in cancers of the bladder, brain and lymphatic tissue and multiple myeloma. These occurred only during the earlier follow-up period. The deaths from cancer of the lymphatic tissues and multiple myelomawere from Hodgkin's disease(2 observedn.lex- pected) and from other lymphomaand multiplemyeloma (4 observedll.2 expected). In contrast to union females, salaried females experienceda marked deficit in cervical cancer deaths during both follow-up periods. Table 7 gives all-cause and all-cancer age-adjusted mortality rates for each employment group. NWM union members experienced the highest overall mortality rates during both follow-up periods and the highest cancer mortality rate for the years 1974 through 1978. All-cause and all-cancer mortality rates either increased slightly or -Tabk 8. Morblltv from All Clurn Combimd Accordlna to E m ~ k m a nQt IWPand Duntlon af Em~lovnwnt. .- . . WhM male union members 2-4 Ls Non-white male union members 2-4 25 Female union members 2-4 25 Salaried males 2-4 25 Salaried females 2-4 25 Obr 730 6067 40 284 120 843 101 963 50 191 cnda 7.8 17.4 6.2 12.7 3.3 ' 8.6 3.7 11.6 1.3 4.5 8bdudind Mlt 13.0 ( O S ) $ 13.8 (0.2) 20.9 (6.0) 14.8 (0.9) 4.5 (0.4) 7.1 (0.2) 8.8 (1.0) 11.2 (0.4) 3.9 (0.8) 6.3 (0.5) 'Crude annual rate (x 1,OOO person-years) tAnnual mortality rata (x 1.OOO person-years), standardizedfor age to the penon-yearsdistributlon of all employees followed 1940-1978 $Numbers in parenthesesam standard e m (x 1,OOO person-years)d the standardized rates 682 Mortality Among Rubber WorkeWDetzelland Monson remained unchanged during the two follow-up periods for WM and NWM union members and decreased for fe male union members and salaried employees. Except for the sharp drop in cancer rates in females, these trends are consistent with patterns observed in the general populh tion. The approximate 40% decline in cancer mortality rates amongfemale rubber workers was not paralleled by a similar decrease in cancer mortality rates for US. white females, which fell less steeply from about 2.4 per 1,OOO person-years in the period1940through1974to about 2.2 per 1,ooO person-years in 1975 through 1978.' For union females the percentageof reporteddeaths inwhich cause of death could not be determined because of missing death certificateswas nearly two times higher in the years 1974 through 1978 than from 1940through 1974.This dif- ference may have contributed to the decline in their cancer SMR (Table 4)and rate betweenthe two follow-up periods. However, cancer SMRs corrected by distributing deaths with missingcertificates by cause according to the distribution of certified deaths were 88 for the period 1940 through 1974 and 64 for 1974 through 1978.It is therefore unlikely that failure to locate death certificates completely explains this finding. Except for WM and NWM union members, workers employed for two to four years experienced lower overall mortality than those employed for at least five years (Table 8).Among WM union members, rates were similar in the two duration of employment categories. NWM union members who worked for two to four years experienced higher overall mortality than men who worked for five or more years, although the magnitude of the standard errors of the rates suggests that this difference is consistent with sampling variability. Numbers of deaths from specific causes among short-term workers were generally insufficient for meaningful analysis. Apparent excesses in deaths from certaincancers amongWM union members employed for two to four years were mentioned earlier in this paper. Discussion The present study evaluated mortality occurring b e tween January1,1940,and July 1, 1978,among 29,087 men and women who worked for at least two years at a large rubber manufacturing facility. These workers have now been observed for an average of 27 years per e m ployee. Before discussing specific findings based on their mortality experience, it is important to note several limitations imposed by the methodology used in the study. First SMRs and RRs reported in this study are probablyconservative estimates of causal associations, if any, b e ween rubber industryemploymentandspecific diseases. Failure to identify all deaths during the study period, unavailability of death certificates for nearly 5% of the reporteddeaths and use of the US. general populationas a comparison group are likely sources of underestimation.' Lack of information on disease determinants other than employment and demographic characteristics constitutes another limitationof the study. Without this information, it is difficult to judge whether observed associations reflect causal links between disease and work- related exposures. Finally, many cause-specific SMRs were based on small numbers of events and were statistically unstable. Therefore, sampling variability cannot be discounted as an explanation for some of the observed excesses and deficits, especially amongemploymentgroups other than the WM union members. For this reason. the remaining discussionwill focus on findings in the latter study group. Reports" concerning mortality during 1940 through 1974 among the rubber workers included in this study noted excess deaths from the following cancers: (1) blad der cancer and leukemia in WM union members; (2)lym phatic cancer and multiple myeloma in NWM union members; (3) cervical cancer, bladder cancer and l y m phatic cancer and multiple myeloma in females; and (4) lymphatic cancer and multiple myeloma in salaried males. Among females there was excess cervical cancer only in union members, and excess lymphatic cancer and myeloma occurred only in salaried employees. The results of recent follow-up through mid-1978 suggest several additional findings. Among WM union members, neither SMRs standardized rate ratios for bladder changed appreciably indicating-ihaithe ex bet workers ha& not diminished in recent yea explanations are possible: (1) there have been no reductions in occupational exposures associated with these cancers; (2)sufficienttime has not elapsed for such reductions to result in decreased mortality; and (3) persistent differences in nonoccupational risk factors for these malignanciesexist betweenrubber workers and U.S. white males. Several studies have identified smoking as a deter- 'minant of bladder cancer.' lo Emphysema and lung cancer are also strongly positively associated with smoking, and mortality from these diseases would be elevated among the rubber workers in this study if they smoked more than the general population. However, WM union members experienced deficits of 14% and 16% in lung cancer and emphysema deaths, respectively,and it is thus unlikely that they smoked more.than U.S. white males in general. WM union members employed fbr at least five years experienced twofold excesses in deaths from esophageal cancer and from other lymphoma and multiple myeloma and an 80% increase in biliary and liver cancer mortality during the period July1, 1974,through JulyI, 1978.Al- though previous studies of rubber workers noted im creased mortality from stomach and large intestine cancer111l * 1 3 and linked these excesses to occupational expo~ures,3~~t~he~r1e5have been no prior reports of elevated esophageal cancer among rubber industry employees. Alcohol and cigarette consumption are the major known risk factors for this malignancy. However, evidence cited above suggests that rubber workers in this study did not smoke more than the general population, and a similar argument can be advanced against the possibility of higher alcohol consumption among rubber workers: WM union members experienced a 17% deficit in deaths from diseases of the digestive system, and the overall SMR for cirrhosis, which is strongly associated with alcohol intake, was 79. Mancuso and Brennan" suggested that gallbladder Journal of Occupational MedicineAol. 23,No. 1O/October 1981 683 * and bile duct cancers are increased among rubber work- ers. To support this hypothesis, they pointed out that the rubber industry employs numerous hepatotoxic agents and that several nitrosamines chemically related to sub- 3. Momon RR and Fine LJ: Cancer mortality and morbidity among ~ b kwrorken. / Natl Cancer /nst61:1047-1053,1978. 4. Monwn RR: Analysis of relative survival and proportional m a tality. Comput Biomcd Re 735-332.1974. 5. Rothman KJ and Buin JD IC E p k h i d o g i c Anahis with a P r e stances used in rubber manufacturing have been shown grammable Calculator. NIH Publication No. 79-1649. Washington, experimentally to be liver and biliary carcinogens. Biliary and liver cancer mortaliw rates among WM union members in the present study rose only slightly between the two follow-up periods, and the marked increase in SMR D.C.: US.Govt Printing Office, 1979. pp 26-29. 6. National Center for Health Statistics (NCHS). Vital statistics of the United States.Vol It, hbrtality, Part A Washington, D.C.: U.S. Govt Printing Office, 1940-1976. 7. Goldsmith J RWhat do we expect from an occupational cohort?/ seen for the period1974 through1978 resulted almost entirely from a sharp decline in rates for U.S. white males after 7969 that was brought about partly by changes in rules for classifying this cause of death." It is possible Occup Med 17:126-127,1975. 8. Cde P, Monson RR, Haning H, and Friedell GH: Smdting and cancer of the lower urinary tract. N E t @ / Med 264:129-134,197l. 9.Wynder EL and Goldsmith R The epidemiology of bladder cancer. A second loo&.Cancer 40:1246-1268,1977. that proceduresfor diagnosing, reportingor codingbiliary 10. MorrisonAS and Cole P: Epidemiology of bladder cancer. Urol and liver cancer as a cause of death among study dece dents differed in some way from those used for general population decedents. Such differences, if present, could account for the excess seen during recent follow-up. Clin N Amer 3 13-29,1976. 11. McMichael AI, Andjelkovich D, and Tyroler H A Cancer mor- tality among rubberworken: An epidemiologic study. Ann N Y A u d Sci 27l:125-137,1976. 12.McMichael AI, Spirtas R. and Kupper LL: An epidemiologic Among the cancers of the hematopoietic and lymphat- study of mortality within a cohortof rubber workers, 1964-72./ Occup ic tissues, leukemia has been the su6ject of muc $'%gation in W'rubber industry, w h @ m W i Med 16:45&464,1974. * 13. Waterhouse IAH: Current status of cancer risk in the rubber i w dustrv. in Advances in Medical Oncolonv. Research and Education. ProcAings of the 12th InternationalCancer Congress. Buenos Aires, 1978. Vol. 111. Epidemiology, 1. M. Birch (Ed.). New York: Pergamon y' 4) also be elevated in rubber workers. Little is known about the etiology of the various lymphatic cancers and multi- ple myeloma. Associations have been reported for oc- cupational groups exposed to chemicals," 22 23 and to radiation." In conclusion, results of recent follow-up were consistent with mortality patterns previously described for these workers' and for other groups of rubber work- Press. 1978. DD 97-705. 14. McMichael AI. Spirtas R. Gamble IF, and T o ~ PqM Mortality among rubber workers: Relationship to specific jobs. / Occup Med 18:178-185, 1976. 15. Andielkovich D. Taulbee I , Symons M. and Williams T Modity of rubber workers with reference to work experience. I Occup Med 19397405.1977. 16. Mancuso TF and Brennan MI: Epidemiologic considerations of cancer of the gallbladder, bile ducts and salivary glands in the rubber industry. / Occup Med 12:333-340.1970. 17. Andielkovich D, Taulbee 1. and Symons M Mortalityexperience ers."13171826 Compared to their general population courv terparts, men and women employed in the rubber industry experience deficits in overall deaths and in deaths from most specific causes but sustain increasedmortality of a cohort of rubber workers, 1964-1973. / Occup Med 18387-394. 1976. 18. Fox AI, Lindars DC, and Owen R: A survey of occupational cancer in the rubber and cablemaking industries: Results of a fiveyear analysis, 1967-7l. Br / Ind Med 31:140-151, 1974. .from several cancers. Although cancer excesses observed during recent follow-up were based on small numbers 19. US. DHEW: Eighth Revision International Classification of Diseases Adapted for Use in the United States. PHS Publ. No. 1693. and may haveoccurred by chance or becauseof systema tic differences in nonoccupational risk factors between Washington. D.C.: US. Govt. Printing Office, 1968. 20. McMichaelAI, Spirtas R, Kupper LL, and Gamble I F Solvent ex- ' posure and leukemia among rubber workers: An epidemiologic study. rubber workers and the general population, the possibility." / O c c ~ pMed 17234-239.1975. that they are related to industrial exposures should be considered.Currently beingexamined in detail are associations between specific work areas within the rubber industry and the occurrence of esophageal, biliary and 21. Li FP, Fraumeni IF, Mantel N. and Miller R W Cancer mortality among chemists. / Natl Cancer lnst 431159-1164.1969. 22. Tabershaw IR and Caffey W R Mortality study of workers in the manufacture of vinyl chloride and its polymers. / Occup Med 16509-518. 1974. liver, lymphatic and other cancers in order to identify sub 23. De Kraay WH: Pesticides and lymphomas in Iowa. / Iowa Med groups of workers having high cancer risks. This informa tion will aid in focusing efforts to control workplace exposures to hazardous substances. Soc L8:50-53.1978. 24. Wagoner JK, Archer VE, Carroll BE, et al: Cancer mortality pat- terns among US. uranium miners and millers, 1950 through 1962. 1 Natl Cancer Inst 32:787-801.1964. References 1. Monson RR and Nakano KK: Mortality among rubber workers. 1. White male union employees in Akron, Ohio. Am / Epidemiol 103284-296.1976. 2. MonsonRR and Nakano KK: Mortality among rubber workers. 11. Other employees. Am / Epidemkd 103297-303.1976. 25. Matanoski CM, Seltser R, Sartwell PE, et al: The current mortali. ty rates of radiologistsand other physicianspecialists: Specific causes of death. A m s / pidcmiollOl:199-2l0.1975. 26. Fox A] and Collier PF: A survcy of occupational cancer in the rubber and cablemaking industries: Analysis of deaths occurring in 1972-74. Brit I Ind Med 33249-264.1976. z f 684 Mortality Among RubberWorkers/Delzell and Monson 1 t