Document o96wLDMaNXEw7oBgrk0KEQzQw

REVISED (9/4/96) An Updated Mortality Study of Workers at a Petroleum Refinery in Beaumont, Texas Gerhard K. Raabe, Dr.P.H. Kim W. Collingwood, Ph.D. Otto Wong, Sc.D. 2 Mobil Business Resources Corporation Medical Department P.O. Box 189 New Hope, PA 18938 Applied Health Sciences, Inc. 81 Second Avenue, Suite 628 San Mateo, CA 94401 Address reprint request and correspondence to Dr. Raabe (Tel: 215-862-5718, Fax: 215-8623551). Short title: Mortality Study of Beaumont, Texas Refinery Workers Key words: petroleum industry, refinery workers, occupational mortality, lymphopoietic cancer, SMR, leukemia, mesothelioma, benzene Acknowledgments: We acknowledge the assistance of Barry Milcarek in this project. In addition, we thank Cheryl Williamson and Josette Hall for manuscript preparation and Deirdra Vachal for technical support. Abstract The mortality experience of 7,119 workers who were employed at a Beaumont, Texas refinery for at least one year between 1945 and 1987 was investigated. Mortality analyses based on standardized mortality ratios (SMRs) and 95% confidence intervals (95% CI) showed overall mortality was significantly lower than expected compared with the U.S. general population (SMR=82, 95% C1:79-86). Total cancer mortality was also lower than expected (SMR=92, 95% C1:84-100). Significant mortality deficits from several malignant and nonmalignant diseases .were reported. A significant mortality increase in the broad category of lymphatic and hematopoietic cancers was found (SMR=133, 95% C1:103-170). This increase was attributed to a nonsignificant elevation in leukemia of all cell-types combined (SMR=139, 95% C1:92-201) and a borderline significant increase in other lymphatic tissue cancer (SMR=158, 95% C1:101-235). The elevation in leukemia was confined to workers hired before 1950. Furthermore, the leukemia excess was shown to have peaked during the 1960s, with mortality no longer elevated post-1980. Analyses of cell-type specific leukemias showed a similar temporal pattern for acute myeloid leukemia (AML) which was not significantly elevated (SMR=136, 95% C1:59-268). Mortality from other leukemia cell types was similar to or lower than expected. Mortality from non-Hodgkin's lymphoma (NHL) (SMR=140, 95% C1:88-211) and multiple myeloma (MM) (SMR=121, 95% C1:55-230) were increased, but neither was statistically significant nor likely to be related to refinery employment. No death from asbestosis was reported, and mortality from mesothelioma and pulmonary fibrosis was lower than expected. Lung .cancer mortality for the overall cohort was similar to expected. For the overall cohort, analyses by duration of employment and time since first employment showed no evidence of any trends for increasing cause-specific mortality. Separate analyses of male workers employed in operator jobs showed mortality patterns which were more favorable than those of the total cohort. Maintenance craftworkers showed statistically significant elevations in mortality for prostate cancer (SMR=145, 95% CI: 107194), leukemia (SMR=179, 95% CI: 111-273), and other lymphatic tissue cancer (SMR=233, 95% CI: 138-368). Detailed analyses indicated that.~ among maintenance craftworkers mortality was elevated for AML, NHL and MM, but none was significant. Furthermore, no upward trend by duration of maintenance jobs was observed. A small increase of lung cancer was observed among maintenance craftworkers (SMR=120, 95% CI: 99-145), which was borderline significant. No relationship between lung cancer and duration of maintenance employment was found. In contrast, a deficit of pulmonary fibrosis was reported among maintenance craftworkers (SMR=62, 95% CI: 17-159). These findings were discussed in conjunction with results from other .refinery studies, and the limitations of the study were discussed. INTRODUCTION The Beaumont refinery has been operating since 1902 and is one of the oldest and largest full service refineries in the U.S. Currently the refinery processes 350,000 barrels of oil a day and produces a variety of petroleum products including gasoline, diesel fuels, fuel oils, jet fuel, kerosene, liquefied petroleum gas, asphalt, lubricating oil and greases, and feedstocks for the petrochemical industry. Refinery workers in general are potentially exposed to a wide range of petroleum-derived hydrocarbons and chemical substances used in the manufacturing of petroleum fuels and lubricants. Specific substances occurring in the working environment of petroleum refineries and previously evaluated by the International Agency for Research on Cancer (IARC) include: anisidine, arsenic compounds, asbestos, benzene, bitumens, butadiene, carbazole,. chlorinated hydrocarbons, chromium and chromium compounds, ethylene dibromide, ethylene dichloride, hydrazine, lead and lead compounds, mineral oils, nickel and nickel compounds, phenylenediamine, polycyclic aromatic compounds, and silica [IARC, 1989]. In 1984, Morgan and Wong completed a retrospective cohort mortality study of workers at the Mobil Beaumont, Texas refinery who worked for at least one year between .January 1, 1945 and January.I, 1979 [Morgan and Wong, 1984]. The vital status of these workers was ascertained through December 31, 1978. Based on a comparison to national mortality rates, workers at the Beaumont refinery experienced a 19% deficit in overall mortality and a 4% deficit in total cancer mortality. A statistically significant excess of lymphatic and hematopoietic cancers was observed among white males. The excess was attributed to a statistically significant increase of deaths from leukemia, although nonsignificant elevations were also noted for lymphosarcoma and reticutosarcoma, and for the category of other lymphatic tissue cancer. To continue surveillance of the mortality experience of the Beaumont refinery workers,. enrollment in the study was extended over an additional nine years of employment eligibility (1979-1987). Vital status follow-up was extended to December 31, 1987. The objective of the update was to expand the previous study, and to continue monitoring the r~ortality pattern.s of the Beaumont refinery workers. Because of the expanded cohort and the extended observation period, the precision of cause-specific mortality risk estimates would likely be improved, and additional detailed analyses for some causes could be performed. MATERIALS AND METHODS The expanded cohort consisted of all Mobil employees who worked a year or more at the Beaumont refinery between January 1, 1945 and December 31, 1987. The cohort members were identified through a combination of personnel records located on-site and computerized payroll files. Information abstracted from these records included social security number, name, sex, race, date of birth, date of employment, employment status at the closing date of the study, vital status at the closing date of the study, and date of retirement, separation, or death when applicable. Employment histories of the cohort members were updated from 1978 to 1987. For each employee, the employment history data consisted of the beginning and ending dates for each job, job titles and departments. Several sources were used for vital status ascertainment, including the company's personnel database, Pension Benefits Information, Inc., the National Death Index, and the Social Security Administration. Death certificates were obtained from company personnel files or from state vital records offices. A qualified nosologist coded the underlying cause of death according to the Eighth Revision of the International Classification of Diseases (ICD) [World Health Organization, 1967, 1969]. Overall and cause-specific mortality was analyzed with the OCMAP program, a modified life-table computer program that calculates cause-specific standardized mortality ratios (SMRs), the ratio of observed to expected deaths expressed in percentage [Marsh and Preini~nger, 1980]. Expected deaths were derived from the National Center for Health Statistics"rflortality rates for the U.S. population, which are specific for cause, sex, race, and 5-year intervals of age and calendar period [Monson, 1987]. OCMAP also computed cause-specific mortality relative to cumulative duration of employment and time since first employment. Entry into vital status follow-up was January 1, 1946 for workers with at least one year fulltime employment prior to that date or, for all others, one day after completion of.one year employment. SMRs for specific death caiegories were examined relative to time since first employment and by duration of employment. The method discussed by Breslow et al. was used to test for trends in SMRs [Breslow et al., 1983]. To investigate mortality by job category, male employees at the Beaumont refinery were classified into two major job categories: (I) operators (having a minimum of 6 months of employment in operator jobs), and (ii) maintenance craftworkers (having a minimum of 6 months of employment in maintenance craft jobs). This latter subcohort of craftworkers consisted principally of workers involved in a 2 broad range of labor and maintenance activities in the mechanical and service divisions of the refinery. In addition, because of the a priori interest in lymphatic and hematopoietic cancers, SMRs for major cell-type specific leukemias (acute myeloid, chronic myeloid, acute lymphatic and chronic lymphatic leukemia, or AML, CML, ALL and CLL), non-Hodgkin's lymphoma (NHL) and multiple myeloma (MM), which were not routinely calculated in the OCMAP program, were also performed. U.S. age-specific mortality rates for cell-type specific leukemias based on data from the National Center for Health Statistics (NCHS) [Selvin et al., 1980] and similar rates for NHL and MM compiled by the National Cancer Institute (NCI) [Pickle et al., 1987] were used in computing the expected deaths from these causes. In addition, because of the use of asbestos at the refinery in the past, separate analyses for asbestos-related diseases were also performed. Expected deaths were based on U.S. rates for pulmonary fibrosis obtained from. NCHS (unpublished data by special request) and for mesothelioma complied by NCI [Connelly et al., 1987]. RESULTS Descriptive Statistics A total of 7,11.9 workers were eligible for inclusion in the updated study (Table I). This repr.esented an increase of 763 whites (80% male), 216 nonwhites (80% male) and 1 male of unknown 'race. Consistent with the previous report, those of unknown race (approximately 5% of the total cohort) were treated as white. This resulted in a cohort 91% male and 83% white (including Hispanic). Selected employment and demographic characteristics of the cohort are also provided in Table I. Ninety-two percent were first employed in their thirties or younger, with only 8% hired at age 40 or older. Approximately half (51%) were hired prior to 1950, and 43% were employed for 30 or more years. The maximum length of potential follow-up was 42 years (January 1, 1946 to December. 31, 1987). The average duration of fotlow-up was 23 years. Fifty-three percent of cohort members had a follow-up of 20 years or more. The increase in person-years from the previous study was 43,072, or 35%, to a total of 166,427. By the end of follow-up (December.31, 1987), 27% of cohort members were still employed, 41% had separated or retired, and 32% were known dead. Information on vital status was 98% complete, with 165 (2%) separated ernpl0yees lost to follow-up. Of the 2,294 cohort members identified to have died between 1946 and 1987, death certificates were obtained for 98%. Overall and Cause-Specific Mortality The number of observed deaths, SMRs and 95% Confidence limits (95% CI) for selected causes for all workers are presented in Table I1. The total number of observed deaths was 2,294 compared with 2,782.5 expected (SMR=821 95% CI: 79-86), indicating a statistically significant deficit of 18% in overall mortality. Mortality from all cancers was slightly lower than expected. The number of observed cancer deaths was 513, versus 559.0 expected (SMR=.92, 95% CI: 84-100), and the deficit was borderline significant. Significant mortality deficits occurred for infective and parasitic diseases, cancer of the buccal cavity and .pharynx, cancer of the rectum, liver cancer, circulatory system diseases, nonmalignant respiratory diseases, digestive system diseases, genito-urinary system diseases, and'~ccidents, poisonings and violence. Significant mortality excesses occurred for the broad category of lymphatic and hematopoietic cancer (SMR=133, 95% CI: 103-170) and senility and ill-defined conditions (SMR=166, 95% CI: 127-212). The elevation for the general category of lymphatic and hematopoietic cancers could be attributed to a borderline significant excess in other lymphatic tissue cancer (SMR=158, 95% CI: 101-235) and a non-significant excess in leukemia (SMR = 139, 95% CI: 92-201). Results for white males (Table III), comprising 75% of the total cohort, showed mortality patterns generally similar to those in the total cohort. However, unlike the total cohort, the deficit in total cancer mortality was larger and.statistically significant. E]orderline significant excess mortality was found for the broad category of lymphatic and hematopoietic cancers (SMR=134, 95% CI: 101-175), whereas mortality excess from the category "other lymphatic tissue cancer" was of comparable magnitude but no longer statistically significant. The mortality excess from senility and ill-defined conditions among white males was statistically significant (SMR=255, 95% CI: 192-333). A total of 52 deaths were observed among female cohort members (not shown). All except 2 deaths occurred among white females. The overall SMR of 95 among white females was not statistically significant. A statistically significant deficit was found for circulatory system diseases (SMR=55, 95% CI: 28-96). Mortality from all cancers was elevated 4 (SMR=145, 95% CI: 92-217), although the increase was not statistically significant. The only statistically significant mortality excess among white females was found for cancer of the large intestine, based on 5 deaths (SMR=320, 95% CI: 104-748). A total of 448 deaths (not shown) were observed among 1,136 nonwhite males, compared with 592.1 deaths expected (SMR=76, 95% CI: 69-83). This 24% deficit was statistically significant and could be attributed primarily to a deficit in diseases of the circulatory system. Other statistically significant mortality deficits occurred for infective and parasitic diseases, nonmalignant respiratory diseases, cirrhosis of the liver, and"~ccidents, poisonings and violence. Deaths from senility and ill-defined conditions were significantly lower than expected (SMR=49, 95% CI: 21-97), in contrast to the significant excess found for white males. No statistically significant cause-specific mortality excesses were observed among nonwhite males. Mortality from all cancers was identical to the expected based on U.S. general population rates (SMR=100, 95% CI: 82-121). In contrast to deficits found in white males, there were slight elevations for mortality from cancer of the respiratory system and stomach cancer among nonwhite males. Mortality by Duration of Employment Table IV shows cause-specific SMRs by duration of employment (<10, 10-29, 30+ years) among all male employees. No analysis by duration of employment was performed for white or nonwhite female employees, since the number of deaths was small. The results for all males indicated that overall mortality remained the same with increased employment duration, whereas total cancer mortality rose slightly. Results for site-specific cancers did not show any consistent pattern. Although SMRs for prostate and brain cancers appeared to increase with duration of employment (primarily due to deficits in the <10 year group), the results for trend tests for both cancer sites were not statistically significant (X2=3.17 and 2.03, respectively). There was no upward trend in mortality from either leukemia or other lymphatic tissue cancer (X2=0.17 and 0.48, respectively). Mortality analysis by duration of employment for white males is shown in Table V. Results are similar to those reported for all males, except that overall mortality decreased slightly with increased employment duration. Also, there was an elevated leukemia SMR in the longest (30+years) length of employment category which was borderline significant (SMR=186, 9'5% 101-311). However, there was no upward trend in mortality among white males from leukemia (X2=0.83). For nonwhite males (not shown), overall mortality rose with employment duration, particularly for diseases of the circulatory system and the category of accidents, poisonings and violence. In general, for the subcohort of nonwhite males, most cancer SMRs decreased among those with more than 10 years of employment, although some of these SMRs were based on small numbers of deaths. Observed numbers of deaths for kidney cancer,. lymphosarcoma and reticulosarcoma, and other lymphatic tissue cancer were small and confined to the longest employment duration category (30+ years). Similarly, nonsignificant mortality excess for lung and prostate cancer occurred among those with 30+ years, of employment. The only statistically significant excess occurred for stomach cancer among nonwhite male workers employed for less than 10 years (SMR=640, 95% C1:132-1871). There was a significant .inverse trend in stomach cancer mortality with increasing duration of employment (X2= 5.57). Mortality by Interval Since First Employment Mortality analysis for all males by interval since first employment (<20, 20-39, 40+ years) is shown in Table VI. As stated above, the number of deaths among white female workers was too small for any subcohort analysis, and no analysis by interval since first employment was performed for white female employees. Mortality for males from all causes and all cancers was marked by increasing SMRs with increasing interval since first employment. For most sitespecific cancers, there was no consistent pattern. Mortality from cancers of the lung, brain & CNS, and lymphosarcoma & reticulosarcoma increased with increased time since first employment. It was interesting to note that prostate, skin, and other lymphatic tissue cancers were each characterized by an inverse trend. No causes of death were significantly elevated for any interval since first employment category. Mortality analysis by interval since first employment for white males is shown in Table VII. Results are similar to those reported for all males except that mortality from cancers of the lung and brain & CNS are unrelated to time since first employment. Also, mortality from the broad category of lymphatic and hematopoietic cancers among white males in the 20-39 years since hire group was significantly elevated (SMR=154, 95% CI: 102-223). This increase was attributed to elevated mortality from leukemia (SMR=177, 95% Cl: 94-302), although this latter increase was not statistically significant. For nonwhite males (not shown), mortality in relation to interval since first employment showed a pattern similar to that among white males. Overall mortality and total cancer mortality increased with increasing number of years since hire. Mortality excess from total cancer was statistically significant among nonwhite male workers in the category of 40 or more years since hire (SMR=137, 95% C~: 106-174). This increase could primarily be attributed to a statistically significant excess of respiratory cancer mortality among nonwhite male workers in this interval subgroup (SMR=196, 95% CI: 127-290). Another significant excess was found for stomach cancer among nonwhite male workers in the shortest interval of time since hire (<20 years; SMR=395, 95% CI: 108-1011). Stomach cancer mortality showed a pattern of decreasing mortality with increasing time since first employment. Mortality from Lymphatic and Hematopoietic Cancers Observed and expected numbers of deaths and SMRs for leukemia and other lymphatic tissue cancer by age at death (<65, 65+) and decade of death among white and nonwhite males are shown in Table VIII. For leukemia deaths among white males, significant excesses were found in older workers (age 65+) during the 1960s and 1970s only. Tabte VIII indicates that leukemia mortality appeared to have peaked dudng the 1960s, when it was significantly elevated among white males (SMR=263). Thereafter, leukemia mortality declined and was no longer elevated p~st-1980. Mortality from leukemia among nonwhite males was difficult to interpret due to the small number of deaths (N=4). Nevertheless, it appeared that leukemia mortality was highest in the 1970s and decreased thereafter. For other lymphatic tissue cancer among white males, mortality was significantly elevated only among younger workers (< age 65) during the 1970s. Only four other lymphatic tissue cancer deaths were observed among nonwhite males. Analysis of mortality from lymphatic and hematopoietic cancers by year of first employment has been found to be informative by other studies of petroleum workers. In 1947, the recommended standard for benzene was reduced from 100 ppm to 50 ppm, which was further reduced to 35 ppm in 1948. In general, benzene exposure levels in the petroleum industry after 1948-1950 were much lower than before. Table IX shows mortality from lymphatic and hematopoietic cancers by year of first employment (<1950, 1950+'). Considering the two separate employment periods, the only statistically significant increase occurred for all lymphatic and hematopoietic cancers among all male workers hired before 1950 (SMR=135, 95% CI: 102-176). With regard to subcategories, both leukemia and other lymphatic tissue cancer were elevated, but neither was statistically significant. In contrast, mortality from both categories of all lymphatic and hematopoietic cancers and leukemia was similar to the expected among workers hired in ~'r ~ 9~ There was an increase of other lymphatic tissue cancer in the latter group, but the increase was based on only 5 deaths, and was not statistically significant. Detailed work histories of leukemia and other lymphatic tissue cancer decedents were examined to determine whether there were common work assignments. A review of these records showed no evidence of any clustering of work assignments or locations. The most frequently reported job departments were pipe, cross still and labor. For the total male cohort the most frequently reported job departments were pipe, labor and lube plant. The subcategories of lymphatic and hematopoietic cancers used in the OCMAP program are based on the statistical classifications compiled by NCHS and do not permit a specific analysis for major cell-type specific leukemias, NHL or MM, which are more appropriate classifications from the biological point of view (Liner, 1985; Heath, 1982; Wintrobe et al., 1981). Therefore, analyses for these specific subcategories of lymphatic and hematopoietic cancers among all male employees were carded out separately. Female cohort members . were not included in these analyses since they contributed fewer than ..one expe.C.t.ed death to these subcategories each. Table X shows that there were 8 observed AML deaths, compared to 5.9 expected (SMR=136, 95% CI: 59-268). For CML, the number of observed deaths was similar to expected (SMR=115, 95% CI: 24-336); whereas non-significant deficits were observed for both ALL and CLL. For NHL among all males, the observed mortality was higher than expected (22 observed versus 15.7 expected deaths, SMR=140, 95% CI: 88-211), but the increase was not statistically significant. For MM among all males, SMR was slightly elevated (9 observed versus 7.5 expected deaths, SMR=121, 95% CI: 55-230). Mortality from Asbestos-related Disease Because of the use of asbestos at the refinery in the past, mortality from asbestos-related diseases (asbestosis, lung cancer, and mesothelioma) among cohort members was examined. Although the OCMAP program provided an analysis for lung cancer, standard rate sets did not analyze mortality data specifically for asbestosis or mesothelioma. Analyses for asbestosis (pulmonary fibrosis) and mesothelioma were performed separately. The term asbestosis refers to pulmonary fibrosis caused by exposure to asbestos, or pulmonary fibrosis in persons with a documented asbestos exposure history. The ICD code (8th revision) for asbestosis is 515.2, and is, therefore, part of the broad category "non-malignant respiratory disease" (ICD 460-519). All death certificates within this broad category among Beaumont refinery workers were reviewed, and no asbestosis death was found. However, since certifying physicians might not be aware of the decedents' asbestos exposure (if any), such deaths could have been coded simply as pulmonary fibrosis. Therefore, the category "pulmonary fibrosis" was analyzed. For pulmonary fibrosis (ICD 515-517), there were 4 observed deaths among all male cohort members, compared to 12.0 expected. The corresponding SMR for pulmonary fibrosis was 33 (95% CI: 0-84), which was statistically significant. As stated above, lung cancer results were provided by the OCMAP program. The results for the overall cohort is presented in Table II and the lung cancer SMR for all males was 92 (95% CI: 78-108). In the 8th revision of ICD, mesothelioma could be coded either as a malignant neoplasm of the peritoneum and retroperitoneal (ICD 158), a respiratory cancer (ICD 162 or 163), a malignant neoplasm of connective and other soft tissue (ICD 171), a malignant neoplasm with unspecified site (ICD 199), or a benign digestive system (ICD 211) or respiratory tumor (ICD 212, 215, or 228). Death.certificates with the above ICD codes among Beaumont refinery workers were reviewed to identify all malignant or benign mesotheliomas. One death listed as "mesothelioma" (ICD 163.0) was identified. Based on national mortality rates [Connelly et al., 1987], 3.1 deaths from mesothelioma were expected (SMR=32, 95% CI: 0-179). Mortality by Job Category Mortality analyses by job category are presented in Tables XI and XII. A total of 1,240 male employees were classified as operators. The SMR for overall mortality among these operators was 66 (95% CI: (]0-72), and that for all cancers was 76 (95% CI: 63-91); both deficits were statistically significant (Table XI). Among operators, no causes of death indicated a significant increase. In contrast, mortality from several malignant and nonmalignant diseases was significantly lower than expected. Analysis of cause-specific SMRs for the 5,360 maintenance craftworkers is shown in Table XII. The SMR for overall mortality was 91 (95% CI: 87-96), and that for all cancers was 113 (95% Cl: 101-125); both were statistically significant. Unlike findings among operators, mortality for maintenance craftworkers was significantly higher than expected for prostate cancer (SMR=145, 95% Cl: 107-194), leukemia (SMR=179, 95% C1:111-273) and the heterogeneous category of other lymphatic tissue cancer (SMR=233, 95% CI: 138-368). Lung cancer mortality was slightly elevated (SMR=120, 95% Cl: 99-145), with borderline statistical significance. Also, mortality from senility and ill-defined conditions was significantly elevated (SMR=184, 95% CI: 135-246). Further analyses for selected causes of death (Table XIII). showed significant elevations for the above malignant causes were restricted to maintenance craftworkers hired before 1950; the SMR for prostate cancer was 140 (95% Cl: 102-189), and that for leukemia was 183 (95% CI: 110-285). For other lymphatic tissue cancer, a higher significant excess occurred among workers hired during or after 1950 (SMR=443, 95% I: 121-1134), although it was based on only 4 observed deaths. Table XIV shows the mortality analyses for specific subcategories of hematopoietic cancers among maintenance craftworkers. There were 7 AML deaths, compared to 3.6 expected (SMR=196, 95% Cl: 78-408). When the AML data were analyzed by duration of employment in maintenance jobs, no upward trend was detected (X2 =0.33). There was also a non-significant increase in NHL (16 observed v. 9.5 expected, SMR=168, 95% CI: 96-273). The highest NHL SMR occurred among workers with-:the shortest duration of maintenance jobs. No upward trend in NHL by duration of maintenance jobs was found (X2 =0.70). There were 6 MM deaths, slightly more than the 4.4 expected. The corresponding MM SMR of 137 was not significant, and no upward trend by duration of maintenance jobs was observed (X2. =0.78). Because maintenance craftworkers were likely to have been exposed to asbestos at the refinery in the past, an analysis of asbestos-related diseases was performed for this subcohort. No death was attributed to asbestosis. For the more general category of pulmonary fibrosis (ICD-8, 515-517), there were 4 observed deaths, compared to 6.43 expected (SMR=63, 95% CI: 17-160). One death was attributed to mesothelioma, whereas 1.84 were expected. For lung cancer, there were 109 deaths among the maintenance craftworker subcohort, compared to 90.8 expected. The corresponding lung cancer SMR of 120 was borderline significant. (95% Cl: 99-145). A detailed analysis of lung cancer by duration of employment in maintenance jobs 1o and by first year in a maintenance job is presented in Table XV. A significant increase in lung cancer was observed for workers with 20-29 years of employment in maintenance jobs (46 observed v. 29.8, SMR=154, 95% CI: 113-206). However, there was no upward trend of lung cancer mortality by duration of maintenance jobs (X2 =1.85). In terms of first year in a maintenance job, only those workers hired between 1940-49 showed a significant increase of lung cancer (43 observed v. 30.5 expected, SMR=141, 95%C1: 102-190). In contrast, there were only 3 lung cancer deaths among those who were hired in or after 1950, compared to 9.3 expected, and the deficit was significant (SMR=32, 95% CI: 7-95). DISCUSSION In the discussion below, we w~first compare some major results in the updated study to those previously reported for the original Beaumont cohort by Morgan and Wong [1984]. We ~ then discuss several specific findings in the updated study which are of particular interest in epidemiologic studies of refinery workers. Extended enrollment and follow-up in this update on Beaumont refinery workers added 980 workers (+16%), 43,072 person-years (+35%) and 712 additional deaths (+45%). In the previous analysis 26% were deceased compared to 36% in this update. Updated results for the total cohort were consistent with the earlier report which demonstrated a favorable overall mortality experience of the Beaumont refinery workers. Based on national rates, the updated SMR of 82 for all causes showed a significant deficit similar to that previously observed (SMR = 81, 95% CI = 77-85). As in the previous study, the favorable overall mortality experience of these refinery workers might be attributed to the "healthy worker effect," whereby working populations exhibit decreased mortality due to initial selection into the workplace and maintenance of this healthier status through benefits derived from employment. The earlier report showed statistically significant deficits for a variety of non-cancer categories: infective and parasitic diseases, circulatory system diseases, nonmalignant respiratory disease (specific,~lly, pneumonia), diseases of the digestive system, diseases of the genito-urinary system, an~'accidents, poisonings and violence. These significant deficits were also observed in updated study. While the earlier study reported an elevated nonsignificant SMR of 145 (95% CI: 90-222) for senility and ill-defined conditions among white males, the update showed a significant SMR of 255 (95% CI: 192-333). This finding is consistent with elevations reported for this category in other industrial studies in this region of the U.S. and is explainable by local death certification practices [Bond et al., 1985; Olsen et al., 1994]. The updated SMR of 92 for all cancers for the total cohort was slightly lower than that originally reported (SMR=96, 95% CI: 86-107), and neither was statistically significant. A significant deficit for liver cancer was reported in the update. Mortality from lymphatic and hematopoietic cancers was the only significantly elevated cancer finding in both reports. However, in the earlier report, the elevation was primarily attributed to a significant excess of leukemia (SMR=173, 95% CI: 109-260), whereas the updated results showed a significant excess only in the subcategory of other lymphatic tissue cancer (SMR=158, 95% CI: 101-235). SMRs were lower in the update for the two subcategories of lymphosarcoma and. reticulosarcoma and leukemia. In particular, the excess for lymphosarcoma and reticulosarcoma reported in the earlier study was no longer present in the update. Furthermore, mortality from leukemia in the update was no longer significantly elevated. In comparing the results in the updated study of the Beaumont refinery workers to results from other previous studies of refinery workers, we will concentrate on a number of cancer sites of interest. Some of the previous epidemiologic studies of refinery workers have focused attention on the following cancers: skin, brain or central nervous system, stomach, pancreas, kidney, bone, lung, colon, prostate and lymphopoietic cancers (especially leukemia). Updated results for Beaumont refinery workers indicated that their mortality from skin, brain, pancreas~- kidney, bone, and colon cancers was similar to that in the general population. In addition, there was no relationship or pattern between mortality from these cancers and duration of employment or interval since hire. These findings from the updated Beaumont refinery study were consistent with the conclusion regarding these cancer sites in a comprehensive review conducted by Wong and Raabe [1989]. With regard to prostate cancer, a statistically significant increase was reported among male maintenance craftworkers (SMR=145, 95% CI: 107-194), with SMRs of comparable magnitude reported for white males (SMR=147, 95% CI: 97-209) and nonwhite males (SMR-143, 95% CI: 82-233). However, a formal trend analysis showed no relationship between prostate cancer and employment duration among white males (X2=1.44) or nonwhite males (X~=0.57). This finding together with reports of no apparent prostate cancer excess in other refinery studies of maintenance workers [Divine and Barron, 1986; Wong et al., 1986; Tsai et al., 1996] suggests that the excess is not likely related to employment in maintenance 12 jobs. It should be noted that a number of risk factors for prostate cancer have been suggested by epidemiologic studies including work in the cadmium industry, dietary fat intake, hormones, and sexual habits (Workshop Conference on the Role of Metals in Carcinogenesis, 1981; Heshmat et al. 1985; Kolonel et al. 1988; Mettlin et al. 1989; Hill et al. 1982; Ross et al. 1987; Steele et al. 1971; Krain 1974]. Unfortunately, no information on these potential risk factors was available in the present investigation. With regard to stomach cancer, a statistically significant increase was reported among nonwhite males who were employed for less than 10 years at the Beaumont refinery. This increase was based on only 3 observed deaths (SMR=640, 95% CI: 132-1871). On the other hand, among nonwhite males with more than 30 years of employment at the refinery, mortality from stomach cancer was less than expected (SMR=78 based on 4 observed deaths). The lack of a relationship between stomach cancer mortality and duration of employment argued against an interpretation that the excess among nonwhite males was related to employment at the refinery. Furthermore, the absence of any increase in mortality from stomach cancer among white males supported our interpretation, which was also consistent with the conclusion regarding stomach cancer in a comprehensive review by Wong and Raabe [1989]. For all males in the updated Beaumont refinery study, a borderline significant mortality increase in lymphatic and hematopoietic cancers was found (SMR=135, 95% C/: 102-176). This excess could be attributed to elevations in leukemia (SMR=143, 95% CI: 95-207) and in other lymphatic tissue cancer (SMR=156, 95% CI: .9_9-234), which were of borderline significance. This was also the case for male maintenance craftworkers, who experienced a significant increase for lymphatic and hematopoietic cancers (SMR=160, 95% CI: 117-215). This excess was attributed to significant increases in leukemia (SMR=179, 95% CI: 111-273) and the heterogeneous group of other lymphatic tissue cancer (SMR=233, 95% CI: 138-368). However, the data for all males and the subcohort of craftworkers indicated that the overall leukemia excess peaked during the 1980s and 1970s, and was no longer elevated post-1980. This temporal pattern suggested that the excess observed prior to 1980 was among workers hired in earlier years (prior to 1950), when historical exposures at the refinery might be different from exposures in later years. Substantially lower leukemia mortality was detected among workers hired in or after 1950. A similar observation was made in a comprehensive review of refinery workers [Wong and Raabe, 1989]. Although the present study stops at the end of 1987, this trend of reduced leukemia mortality inthe 1980s has been 13 noted in other studies of workers in the petroleum industry (Honda et al., 1995). With regar~l to specific job categories at refineries, other studies reported essentially no association between leukemia and maintenance jobs [Divine and Barron, 1986; Wong et al., 1986; Tsai et al., 1996] Although a borderline significant increase was reported for the heterogeneous category of other lymphatic tissue cancer, the absence of any trend in SMR with increasing duration of employment (X2=1.68 and 2.63 in white and nonwhite males, respectively) and an apparent inverse relationship with time since hire were inconsistent with the hypothesis that the excess was the result of exposures at the refinery. Furthermore, a review of the work histories of the 23 decedents showed no indication of common job assignments. Although the excess of other lymphatic tissue cancer was more pronounced and statistically significant among maintenance craftworkers (SMR=233, 95% CI: 138-368), the. lack of any trend in SMR with increase in duration of maintenance employment does not support a connection with employment in Craft jobs. Also, no other studies of refinery workers have reported excesses for other lymphatic tissue cancer in relation to work in maintenance jobs [Divine and Barron, 1986; Wong et al., 1986; Tsai et al., 1996]. Finally, it is interesting to note that operators had a leukemia mortality experience similar to the general population (SMR=94, 95% CI: 31-220). This suggests that their benzene exposures were controlled below levels at which leukemia risk would be increased. To further analyze mortality data for lymphatic and hematopoietic cancers, analyses for major cell-type specific leukemias (AML, CML, ALL and CLL), NHL and MM were performed. Mortality from AML was elevated (SMR=136, 95% Cl: 59-268, 8 observed deaths versus 5.9 expected), but the increase was not significant. Seven of the 8 AML decedents were hired prior to 1950, including 4 in or prior to the 1930s (7 observed versus 5.0 expected, SMR=140, 95% CI: 56-289). The remaining AML case, who was hired in 1974 and terminated in 1975, died in 1976 at age 23. Based on a consideration of the extremely short latency of two years. alone, this AML case could not have been related to employment at the refinery. Indeed, given the average survival of AML, the diagnosis in this short-term employee was most likely made either before or around the time when he was hired at the refinery. The AML data again supported the conclusion that no increase occurred among workers hired in or after 1950. Observed mortality from ALL and CLL was lower than expected, and for CML the observed mortality was similar to expected. Thus, cell-type specific analyses demonstrated that there was no increase in CML, ALL or CLL among workers at the Beaumont refinery. An increase in mortality was found for NHL among male employees (SMR=140, 95% CI: 88-211), but the increase did not reach statistical significance. Unlike AML, analysis of NHL by hire date did not show any significant difference between those hired prior to 1950 (18 observed versus 13.7 expected, SMR=132, 95% CI: 78-209) and those hired thereafter (4 observed versus 2.1 expected, SMR=193, 95% Cl: 23-697). This pattern appeared to indicate that the NHL increase was not related to presumed higher levels of exposure in the earlier employment period (i.e. , <1950). Other studies of refinery workers, which reported NHL as a separate disease category, did not find any association between NHL and refinery employment. A recently completed study of workers at the Mobil Torrance refinery [Milcarek et al., 1994] reported an SMR of 120 (4 observed versus 3.3 expected, 95% CI: 33-307). A study at the Mobil Paulsboro refinery [Collingwood et al., 1996] reported no significant increase of NHL (15 observed vs. 11.4 expected, SMR=132, 95% C1:74-217). Similarly, a recently completed large-scale study of refinery workers in Port Arthur, Texas, reported an SMR for NHL of 71 based on 30 observed NHL deaths [Satin et al., 1996]. Thus, collectively (total observed = 71, total expected = 72.7) these four studies support the interpretation that the nonsignificant increase of NHL among Beaumont refinery workers was not occupationally related. Epidemiologic studies have identified several potential risk factors for NHL, including autoimmune diseases, viruses, chemotherapy and radiation therapy, ionizing radiation, a history of diabetes, tuberculosis or malaria, working or living on a farm, a family history of cancer, and cigarette smoking [Wilson et al., 1991; Franceschi et al., 1989; Cartwright et al., 1988; Vianna and Polan, 1979; Schumacher, 1985; Schumacher and Delzell, 1988; McDonald, 1987; Hoar et al., 1985; Zahm et al., 1989; Woods et al., 1987; Pearce et.al., 1987]. Unfortunately, no information on these potential risk factors was available for the NHL cases in this study. Mortality from MM among male workers at the Beaumont refinery was slightly, but not significantly, elevated (9 observed versus 7.45 expected, SMR=121, 95% CI: 55-230). This small increase was well within expectation based on U.S. rates, and did not indicate any association between the disease and employment at the refinery. Similarly, other studies of refinery workers, in which MM was analyzed as a separate disease category, did not find any association between MM and refinery employment. The SMRs for MM were 128 (95% CI: 15462) at the Torrance refinery, 74 (95% CI: 20-190) at the Paulsboro refinery, and 98 (95% CI: 59-153) at the Port Arthur refinery [Milcarek et al., 1994; Collingwood et at., 1994; Satin et al., 15 1996]. Collectively, these studies do not support a relationship between refinery employment and MM. Similar to other industrial facilities, asbestos was used at the Beaumont refinery in the past, and detailed analyses of asbestos-related diseases were performed for the overall cohort as well as for the maintenance craftworker subcohort. Although specific exposure measurements were not available, data from the industry as a whole have indicated that asbestos concentrations at refineries were generally extremely low [IARC, 1989]. Among the Beaumont refinery employees, no asbestosis death was reported, and mortality from pulmonary fibrosis was significantly lower than expected. One death was attributed to malignant mesothelioma in the overall cohort, whereas 3.1 deaths were expected. In the overall cohort, lung cancer mortality was similar to expected. Furthermore, analysis by duration of employment among males did not show any upward trend in lung cancer risk. Thus, the data indicated that there was no relationship beb~veen lung cancer mortality and employment at the Beaumont refinery for the overall cohort. A comprehensive review by Wong and Raabe [1989] reached a similar conclusion on lung cancer and employment at refineries. For male maintenance craftworkers, there was a small increase in lung cancer (SMR=120, 95% CI: 99-145), which was borderline significant. Detailed analysis indicated that the significant lung cancer increase was primarily restricted to maintenance craftworkers hired during 1940-49. Several investigators have suggested that workers hired during the same time period (the "war years") were in general less healthy (Doll 1988, Enterline 1988, Choi 1992). Furthermore, trend analyses showed no relationship with duration of employment in maintenance jobs. A complete evaluation of the lung cancer issue requires information on smoking and employment outside the petroleum industry which might entail potential exposure to asbestos and other carcinogens. Unfortunately, such information is not available in the present study. Nevertheless, it should also be pointed out that other refinery studies showed. no association between lung cancer and work in maintenance jobs [Divine and Barron, 1986; Wong et al., 1986; Tsai et al. 1996]. In closing, it should be pointed out that there are some potential limitations in this study. Most are typical of a historical mortality study of industrial populations. First, although both the percentages of individuals with unknown vital status (2%) and the proportion of outstanding death certificates (2%) were low, it was possible that some deaths from causes of interest might have been missed. We do not believe that such low percentages would have any significant impact on our results. Second, the health endpoint in our study was mortality, which is a reasonable surrogate for incidence of most cancers. However, being a mortality study, the investigation inherited the problems associated with death certificates (diagnostic accuracy, for example). No detailed clinical information was available on the deaths in our study. However, it must be pointed out that although detailed information derived from medical records or pathology reports may be more accurate than that based on death certificates, it would be inappropriate to use such information in a historical cohort mortality study. In our study we compared diagnoses based on death certificates to national statistics which were derived from death certificates as well. Furthermore, our analysis was adjusted for calendar time, thus at least partially controlling for changes in survival and diagnostic practices. As in most historical cohort mortality studies, little quantitative exposure data were available during the early part of the study, and analyses by quantitative exposure indices were not possible. In the present study, we relied on duration of employment, period offirst exposure as well as the broad classification of jobs (operators and maintenance craftworkers). Such exposure surrogates might lack both sensitivity and specificity. Although several additional detailed analyses were performed for both white and nonwhite male employees, the number of deaths among white female employees remained small and it was not meaningful to perform similar detailed statistical analyses for the latter in this update. Furthermore, also due to small numbers, no statistical analyses were performed for nonwhite females. Hopefully, by periodically updating and expanding the cohort, the number will be adequate for detailed statistical analyses among female employees in the future. Finally, information on lifestyle as well as on employment elsewhere was not available, although such information would be extremely helpful in interpreting some of the findings in the present study. For example, information on smoking as well as employment elsewhere with potential exposure to asbestos and other carcinogens would be needed in a complete evaluation of the lung cancer issue. CONCLUSION This updated study has confirmed earlier findings of an overall favorable mortality experience of employees at the Beaumont, Texas refinery compared to the U.S. general population. In addition, significant mortality deficits were found for several causes of death. No significant increase was detected in any cause-specific mortality, except for the heterogeneous category of other lymphatic tissue cancer. A non-significant increase in leukemia was also observed. However, detailed analyses indicated that the excess was restricted to workers hired prior to 1950, and that the excess appeared to have peaked in the 1960s. Detailed a~alyses by specific leukemia cell-types indicated a similar pattern for acute myeloid leukemia, but no increase was detected for other leukemia cell-types. No causespecific mortality was found to be associated with duration of employment at the refinery, including several causes which have been reported to be elevated in some previous studies. Subcohort analyses by job category showed mortality patterns in operators which were generally more favorable than the total cohort. On the other hand, statistically significant increases in mortality occurred among maintenance craftworkers for prostate cancer, leukemia, and the heterogeneous category other lymphatic tissue cancer. Also, a small increase in lung cancer was seen. Analyses of the temporal patterns of these SMRs along with findings from other studies of similar cohorts provided evidence that the excesses were most likely to be unrelated to work in maintenance craftsman jobs. REFERENCES Bond GG, Shellenberger R J, Fishbeck WA, Cartmill JB, Lasich B J, Wymer KT, Cook RR (1985): Mortality among a large cohort of chemical manufacturing employees. J Natl Cancer Inst 75:859-869. Breslow NE, Lubin JH, Marek P, Langholz B (1983): Multiplicative models and cohort analysis. J Am StatAssoc 78:1-12. Cartwright RA, McKinney PA, O'Brien C, Richards IDG, Roberts B, Lauder I, Darwin CM,' Bernard SM, Bird CC (1988): Non-Hodgkin's lymphoma: Case control epidemiological study in Yorkshire. Leukemia Research 12:81-88. Choi B (1992): Definition, sources, magnitude, effect modifiers, and strategies of reduction of the healthy worker effect. J ccup Med 34: 979-988. Collingwood KW, Raabe GK, Wong O (1996): An updated cohort mortality study of workers at a northeastern United States petroleum refinery. Arch Occup Environ Health 68:277-288. Connelly RR, Spirtas R, Myers MH, Percy CL, Fraumeni JF (1987): Demographic patterns for mesothelioma in the United States. J Natl Cancer Inst 78:10531060. Divine B J, Barron V (1986): Texaco mortality study: I1. patterns of mortality among white males by specific job groups. AM J Ind Med 10:371-381. Doll R (1988): Commentary in the Report to the Worker's Compensation Board on the Healthy Worker Effect. IDSP Report 3. Toronto: Industrial Disease Standards Panel. Enterline PE (1988): Commentary in the Report to the Worker's Compensation Board on the Healthy Worker Effect. IDSP Report 3. Toronto: Industrial Disease Standards Panel. Franceschi S, Serraino D, Bidoli E, Talamini R, Tirelli U, Carbone A, La Vecchia C (1989): The epidemiology of non-Hodgkin's lymphoma in the north-east of Italy: A hospital-based casecontrol study. Leukemia Research 13:465-472. Heath CW. The leukemias, in Cancer Epidemiology and Prevention (Schottenfeld D, Fraumeni JF; eds.), WB Saunders Company, 1982. Heshmat MY, Kaui L, Kovi J, Jackson MA, Jackson AG, Jones GW, Edson M, Enterline P, Worrell RG, Perry SL (1985): Nutrition and prostate cancer: A case-control study. Prostate 8:7-17. Hill P, Wynder EL, Garbaczewski L, Garnes H, Walker AR (1982): Response to luteinizing releasing hormone, thyrotrophic releasing hormone, and human chronic gonadotrophin administration in healthy men at different risks for prostatic cancer and in prostatic cancer patients. Cancer Res 42:2074-2080. Hoar SK, Blair A, Holmes FF, Boysen CD, Robel R J, Hoover R, Fraumeni JF (1986): Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. JAMA 256:11411147. Honda Y, Delzell E, Cole P (1995): An updated study of mortality among workers at a petroleum manufacturing plant. J Occup Environ Med 37: 194-200. Kolonel LN, Yoshizawa CN, Hankin JH (1988): Diet and prostatic cancer: A case-control study in Hawaii. Am J Epidemiol 127:999-1012. Krain LS (1974): Some epidemiologic variables in prostatic carcinoma in California. Prev Med 3:154-159. International Agency Research on Cancer (1989): IARC monographs on the evaluation of carcinogenic risks to humans, occupational exposures in petroleum refining. Crude Oil and Major Petroleum Fuels, Volume 45. Liner MS. The Leukemias: Epidemiologic Aspects. Oxford University Press, 1985. Marsh GM, Preininger M (1980): OCMAP: A user-oriented occupational cohort mortality analysis program. Am Stat 34:245-246. McDonald JA (1986): A case-control study of non-Hodgkin's lymphoma. Ph.D. dissertation, Department of Epidemiology, University of Washington. Mettlin C, Selenskas N, Natarajan N, Hubert R (1989): Beta-carotene and animals fats and their relationship to prostate cancer risk. Cancer 64:605-612. Milcarek BI, Collingwood KW, Raabe GK (1994): An updated mortality study of workers at the Torrance, California refinery, 1959-1987. Mobil Oil Corporation. Monson R (1987): Memorandum - USDR users. Harvard University School of Public Health, Department of Epidemiology, March 12. Morgan RW, Wong O (1984): An epidemiologic analysis of the mortality experience of Mobil Oil Corporation employees at the Beaumont, Texas refinery. Technical report submitted to Mobil Oil Corporation. Olsen GW, Lacy SE, Cartmill JB, Kravat BA, Camberlin SR, Spadacene NW, Lipps TE (1994): Half-century of cause-specific mortality experience of chemical manufacturing employees. Am J Ind Med 26:203-219. Pearce NE, Sheppard RA, Smith AH, Teague CA (1987): Non-Hodgkin's lymphoma and farming: an expanded case-control study. Int J Cancer 39:155-161. Pickle LW, Mason T J, Howard N, Hoover R, Fraumeni JF (1987): Atlas of U.S. cancer mortality among whites: 1950-1980. DHHS Publication NO. (NIH) 87-2900. Ross RK, Shimizu H, Paganini-Hill A, Honda G, Henderson BE (1987): Case-Control studies of prostate cancer in blacks and whites in Southern California. JNCI 78:869-874. Satin KP, Wong O, Yuan LA, Bailey W J, Newton KL, Wen CP, Swencicki RE (1996): A 50year mortality follow-up of a large cohort of oil refinery workers in Texas. J Occup Environ Med 38:492-506. Schumacher MC (1985): Farming occupations and mortality from non-Hodgkin's lymphoma in Utah. J Occup Med 27:580-584. Schumacher MC, Delzell E (1988): A death-certificate case-control study of non-Hodgkin's lymphoma and occupation in men in North Carolina. Am J Ind Med 13:317-330. Selvin S, Levin LI, Merrill DW, Winkelstein W (1983): Selected epidemiologic observations of ceil-specific leukemia mortality in the United States, 1969-1977. Am J Epidemiol 117(2):14052. Steele R, Lees RE, Krauss AS, Rao C (1971): Sexual factors in the epidemiology of cancer of the prostate. J Chron Dis 24:29-37. Tsai SP, Waddell LC, Gilstrap EL, Ransdell JD, Ross CE (1996): Mortality among maintenance employees potentially exposed to asbestos in a refinery and petrochemical plant. Am J Ind Med 29:89-98. Vianna N J, Polan A (1979): Lymphomas and occupational benzene exposure. Lancet 1:1394-1395. Wilson JD, Braunmald E, Isselbacher K J, Petersdorf RG, Martin JB, Fauci AS , Root RK (1991): Harrison's principles of internal medicine (Twelfth Edition): McGraw-Hill, Inc., New York. Wintrobe MM, Lee GR, Boggs DR, Bithell TC, Forester J, Athens JW, Lukens JN. Clinical Hematology (8th ed.), Lea & Febiger, 1981. Woods JS, Pollssar L, Severson RK, Heuser LS, Kulander BG (1987): Soft-tissue sarcoma and non-Hodgkin's lymphoma in relation to phenoxy herbicide and chlorinated phenol exposure in western Washington. JNCI 78:899-910. Wong O, Morgan RW, Bailey W J, Swencicki RE, Claxton K, Kheifets L (1986): An epidemiological study of petroleum refinery employees. Br J Ind Med 43:6-17. Wong O, Raabe GK (1989): Critical review of cancer epidemiology in petroleum industry employees, with a quantitative me{a-analysis by cancer site. Am J Ind Med 15:283-310. Workshop conference on the role of metals in carcinogenesis, 1980 (1981): Problems of epidemiological evidence. Environ Hlth Perspect 40:11-20. 21 World Health Organization (1967, 1969): "Manual of the international statistical classification of diseases, injuries and cause of death." Based on the recommendations of the Eight Revision Conference, 1965, and adopted by the Nineteenth World Health Assembly, Vols. I, II. Zahm SH, Weisenburger DD, Babbitt PA, Saad RC, Vaught JB, Cantor KP, Blair A (1990): A case-control study of non-Hodgkin's lymphoma and the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) in eastern Nebraska. Epidemiology 1:349-356. 22 Table I. Characteristics of the Beaumont, Texas Cohort Total Employees Total Person-years Race White Nonwhite Unknown Sex Male Female Year of hire Before 1950 1950+ Duration of employment (yr) <10 10-29 30+ Age at hire <20 20-39 40+ Age at death <50 50-59 60-69 70+ Total Deaths N 7,119 166,427 5,544 1,204 371 6,482 637 3,647 3,472 2,409 2,569 2,141 959 5,613 547 205 342 677 1,070 2,294 100.0 100.0 77.9 16.9 5.2 91.1 8.9 51.2 48.8 33.8 36.1 30.1 13.5 78.8 7.7 8.9 14.9 29.5 46.6 100.0 Table II. Observed arid Expected Deaths by Cause for the Beaumont Refinery Cohort Number of Persons: 71119 Cause of Death (8.th ICDA~ All Causes (001*g9g) Infective & Parasilic Dlseases (001-13g) All Cancers (140-209) Cancer of Buccal Cavity and Pharynx (140-149) Cancer of DigeslJve System (150-159) Cancer of Esophagus (150) Cancer of Stomach (151) Cancer of Large Intestine (153) Cancer of Rectum (154) Cancer of Uver (155-156) Cancer of Pancreas (157) Cancer of Respiratory System (160-163) Cancer of Larynx (161) Cancer of Lung (162-163) Cancer of Bone (170) Cancer of Skin (172-173) Cancer of Breast (174) Cancer of Prostate (185) Cancer of Testis (186-187) Cancer of Bladder (188) Cancer of Kidney (189) Cancer of Brain and CNS (191-192) LymphalJc & HematopoietJc Cancer (200-209) Lymphosarcoma and Reticulosarcoma (200) Hodgktn's Disease (201) Leukemia & Aleukemia (204-207) Other Lymphatic Tissue Cancer (202, 203, 208} Benign Neoplasms (210-239) Diabetes Mellitus (250) Diseases of Blood (280-289) Diseases of Nervous System (320-389) Diseases of Circulatory System {390-458) Chronic Rheumatic Heart Disease (393-398) Arteriosclerotic Heart Disease (410-413) Vascular Lesions of CNS (430--438) Nonmalignant Respiratory Disease (460-519) Pneumonia (480-486) Emphysema (492) Diseases o1 Digestive System (520-577) Cirrhosis of Liver (571) Diseases of Genito-Urinary System (580-629) Senility & III-Detined Conditions (780-799) Accidents, Poisonings, & Violence (800-998) Accidents (800-949) Motor Vehicle Accidenls (810-827) Suicide (950-959) Observed Deaths 2,294 18 513 4 129 10 32 40 5 5 31 158 3 155 0 9 7 53 1 11 8 12 65 9 3 28 24 5 32 4 23 1,246 12 815 191 101 35 37 70 20 25 63 123 82 32 28 Expected Deelhs 2,782.5 51.5 559.0 15.8 163.7 15.8 33.8 50.9 16.2 13.3 29.7 175.3 7.9 165.7 2.2 8.1 3.8 54.6 2.3 17.2 12.1 12.0 48.7 8.6 4.5 20,1 15.2 7,1 41.4 7.0 24.9 1,454,9 20.7 926.1 228.3 184.8 74.4 33.5 112.1 50.8 54.5 38.0 206.1 135.8 58.6 41.3 SMR 82 35 92 25 79 63 95 79 31 38 105 90 38 94 0 111 185 g7 44 64 66 100 133 105 66 139 158 70 77 57 93 86 58 88 84 55 47 111 62 39 46 166 60 6O 55 68 Person Years: 166~427 Confidence Umll~ Lower Upper 79 86 21 55 7 65 66 94 30 116 65 134 56 107 10 72 12 88 71 148 8 111 79 110 51 211 74 382 73 127 1 247 32 115 52 175 103 170 48 199 14 194 92 201 101 235 23 164 53 109 16 146 59 139 81 91 30 101 82 94 72 96 45 66 33 65 78 152 49 79 24 61 30 68 127 212 50 71 48 75 37 77 45 98 * Statistically significant al 0.05 level ** Statistically signiticant at 0.01 level Table III. Observed and Expected Deaths by Cause for White Male Employees in the Beaumont Refinery Cohort Number of Pemons: 5,346 Cause of Death (8th ICDA) All Causes (001-ggg) Infective & Parasitic Diseases (001-I 3g) All Cancers (140-209) Cancer of Buccal Cavily and Pharynx (140-149) Cancer of Digestive System (150-159) Cancer of Esophagus (150) Cancer of Stomach (151 ) Cancer of Large Intestine (153) Cancer of Rectum (154) Cancer of Liver (155-156) Cancer of Pancreas (157) Cancer of Respiratory System (160-163) Cancer of Larynx (161) Cancer of Lung (162-163) Cancer of Bone (170) Cancer of Skin (172-173) Cancer of Prostate (185) Cancer of Testis (186-187) Cancer of Bladder (188) Cancer of Kidney (189) Cancer of Brain and CNS (191-192) Lymphatic & Hematopoietic Cancer (200-209) Lymphosarcoma and Reticulosarcoma (200) Hodgkin's Disease (201) Leukemia & Aleukemia (204-207) Other Lymphatic Tissue Cancer (202, 203, 208) Benign Neoplasms (210-239) Diabetes Mellitus (250) Diseases of Blood (280-289) Diseases of Nervous System (320-389) Diseases of Circulatory System (390-458) Chronic Rheumatic Heart Disease (393-398) Arteriosclerotic Heart Disease (410-413) Vascular Lesions of CNS (430-438) Nonmalignant Respiratory Disease (460-519) Pneumonia (480-486) Emphysema (492) Diseases of Digestive System (520-577) Cirrhosis of Liver (571) Diseases of Genito-Urinary System (580-629) Senility & Ill-Defined Conditions (780-799) Accidents, Poisonings, & Violence (800-998) Accidents (800-949) Motor Vehicle Accidents (810-827) Suicide (950-959) Observed Deaths 1,794 13 381 3 93 9 19 30 3 3 24 120 3 117 0 8 37 0 9 7 10 54 7 3 24 19 4 22 3 19 999 9 677 142 80 26 31 55 18 13 54 100 68 25 26 Expected Deaths 2,135.4 29.8 433.7 12.4 124.1 10.1 23.3 41.5 13.3 9.7 23.1 142.5 6.1 134.9 1.8 7.3 37,6 1.8 14.4 10.3 10.6 40.2 7.4 3.8 16.9 11.9 5.4 30.6 5.3 19.3 1,140.2 16.5 774.6 158.3 146.1 53.8 29.7 88.3 41.0 33.6 21.2 153.1 103.8 45.8 37.2 SMR 84 ** 44 ** 88 * 24 * 75 ** 89 82 72 23 ** 31 * 104 84 49 87 0 ...... 110 98 0 ...... 63 68 95 134 * 95 79 142 160 74 72 56 99 88 ** 55 87 ** 90 55 ** 48 ** 104 62 ** 44 ** 39 ** 255 ** 65 ** 66 * 55 ** 70 Pemon Years: 129,793 Confidence Umlts Lower Upper 80 88 23 75 79 97 5 71 61 92 41 170 49 128 49 103 5 66 6 90 67 154 70 101 10 143 72 104 48 217 69 136 29 119 27 140 45 174 101 175 38 196 16 230 91 211 96 249 20 188 45 109 12 165 59 154 82 93 25 104 81 94 76 106 43 68 32 71 71 148 47 81 26 69 21 66 192 333 53 79 51 83 35 81 46 103 * Statistically significant at 0.05 level ** Statistically significant at 0.01 level Table IV. Observed Deaths, SMRs and 95% CI for Selected Causes of Death Among All Male Employees in the Beaumont Refinery Cohort by Length of Employment Cause of Death (8th ICDA) All Causes All Cancers (140-209) Cancer of Digestive System (150-159) Cancer of Esophagus (150) Cancer of Stomach (151) Cancer of Large Intestine (153) Cancer of Rectum (154) Cancer of Liver (155-156) Cancer of Pancreas (157) Cancer of Respiratory System (160-163) Cancer of Lung (162-163) Cancer of Skin (172-173) Cancer of Prostate (185) Cancer of Bladder (188) Cancer of Kidney (189) -Cancer of Brain and CNS (191-192) Lymphatic and Hematopoietic Cancer (200-209) Lymphosarcoma and Reticulosarcoma (200) Hodgkin's Disease (201) Leukemia& Aleukemia (204-207) Other Lymphatic "Rssue Cancer (202, 203, 208) Benign Neoplasms (210-239) Diseases of Blood (280-289) Diseases of Circulatory System (390-458) Arteriosclerotic Heart Disease (410-413) Nonmalignant Respiratory Disease (460-519) Pneumonia (480-486) Emphysema (492) Diseases of Digestive System (520-577) Cirrhosis of liver (551) Accidents, Poisonings, & Violence (E800-E998) Accidents (800-949) Motor Vehicle Accidents (810-827) Suicide 1950-959) Number of Persons at Risk Number of Person-Years Statistically significant at 0.05 level Statistically significant at 0.01 level Obs 300 60 20 1 6 4 1 1 6 18 18 3 1 0 1 0 9 0 0 5 4 0 2 133 89 9 3 2 9 7 34 24 11 7 5,054 57,887 <10 SMR 83 ** 89 113 61 181 70 57 69 181 80 84 190 28 0 60 0 119 0 0 165 204 0 196 89 89 46 * 40 66 51 * 72 49 ** 55 ** 50 * 55 (95% CI) (74-93) (68-115) (69-175) (2-175) (67-395) (19-180) (1-317) (2-382) (66-393) (47-126) (50-133) (39-554) (1-153) ~ (2-333) ~ (55-227) --- (54-385) (56-521) -(24-708) (74-105) (72-110) (21-88) (8-116) (8-238) (23-96) (29-148) (34-69) (35-82) (25-89) (22-113) Length of EmpIo~/ment ~ears) 10-29 Obs SMR (95% CI) 886 185 37 4 11 8 3 1 9 64 63 4 16 7 3 6 25 5 2 8 9 3 1 470 320 46 12 21 37 7 55 36 11 11 4,444 64,849 81 *" 87 58 -" 62 78 42 ** 46 19 79 95 100 126 82 108 63 121 134 142 104 105 165 105 36 82 " 88 " 66 *" 42 ** 167 * 76 31 *" 65 ** 65 - 48 *" 64 (75-86) (75-100) (41-8o) (17-160) (39-139) (18-84) (10-135) (1-1o6) (36-149) (73-122) (77-127) (34-324) (47-133) (43-222) (1 3-184) (44-262) (87-198) (46-332) (13-376) (45-207) (75-313) (22-306) (1-202) (75-90) (79-99) (48-88) (21-73) (104-256) (54-105) (12-63) (4985) (46-90) (24-86) (32-114) 30+ Obs SMR 1056 245 66 5 15 23 I 2 16 72 70 2 36 4 4 5 29 3 1 15 10 2 i 630 401 45 19 14 23 6 33 22 10 9 2,104 32,356 84 ** 93 85 66 94 93 13 "* 32 113 87 88 65 115 45 74 123 137 85 80 168 137 71 33 89 *" 89 * 49 ~" 52 ** 80 53 "" 36 "* 67 * 63 * 81 87 (95% (79-89) (82-106) (65-108) (21-153) (53-156) (59-139) (0-73) (4-115) (64-183) (68-109) (69-112) (8-235) (80-159) (12-114) (20-190) (40-287) (92-197) (18-249) (2-446) (94-277) (66-251) (9-258) (1-185) (82-96) (80-98) (35-65) (31-81) (44-134) (34-80) (13-78) (46-94) (40-96) (39-150) (40-1651 Table V. Observed Deaths, SMRs and 95% CI for Selected Causes of Death Among White Male Employees In the Beaumont Refinery Cohort by Length of Employment Cause of Death (Sth ICDA) All Causes All Cancers (140-209) Cancer of Digestive System (150-159) Cancer of Esophagus (150) Cancer of Stomach (151) Cancer of Large Intestine (153) Cancer of Rectum (154) Cancer of Uver (155-156) Cancer of Pancreas (157) Cancer of Respiratory System (160-163) Cancer of Lung (162-163) Cancer of Skin (172-173) Cancer of Prostate (185) Cancer of Bladder (188) Cancer of Kidney (189) Cancer of Brain and CNS (191-192) Lymphatic and Hematopoietic Cancer (200-209) Lymphosarcoma and Reticulosarcoma (200) Hodgkin's Disease (201) Leukemia& Aleukemia (204-207) Other Lymphatic Tissue Cancer (202, 203, 208) Benign Neoplasms (210-239) Diseases of Blood (280-289) Diseases of Circulatory System (390-458) Arteriosclerotic Heart Disease (410-413) Nonmalignant Respiratory Disease (460-519) Pneumonia (480-486) Emphysema (492) Diseases of Digestive System (520-577) Cirrhosis of Uver (551) Accidents, Poisonings, & Violence (E800-E998) Accidents (800-949) Motor Vehicle Accidents (810-827) Suicide {950-959) Number of Persons at Risk Number of Person-Years * Statistically significant at 0.05 level * Stalistlcally significant at 0.01 level < 10 Obs SMR (95% CI) 273 53 15 0 3 4 1 1 5 17 17 3 1 0 1 0 8 0 0 4 4 0 2 123 84 8 2 2 9 7 30 22 11 7 4,207 50,631 87 * 84 93 0 106 74 61 79 160 79 83 195 30 0 63 0 114 0 0 142 216 0 247 90 88 47 33 68 58 82 56 *" 61 * 58 S0 " (77-98) (63-110) (52-153) --- (22-309) (20-190) (2-341) (2-438) (52-374) (46-127) (48-133) (40-569) (1-169) --- (2-350) --- (49-224) ----- (39-365) (59-553) --(30-892) (75-107) (71-109) (20-93) (4-118) (8-245) (27-110) (33-168) (38-80) (38-92) (29-104) {24-124) Lenc Obs 703 149 27 4 5 7 2 1 8 54 53 3 13 5 3 5 23 5 2 6 9 2 0 374 266 37 10 18 29 6 44 30 9 11 3,569 53,254 10-29 SMR {95% Cl) 84 ** 89 55 ** 101 52 44 * 37 26 88 98 102 103 98 93 73 111 147 163 122 92 206 89 0 84 ** 88 * 68 * 50 * 162 76 32 ** 71 73 51 * 69 (78-90) (75--104) (36-80) (28-259) (17-121) (18-91) (5-134) (1-145) (38-174) (74-128) (76-133) (21-301) (52-168) (30-216) (15-213) (36-258) (93-220) (53-381) (15-442) (34-200) (94-392) (11-323) --(76-93) (77-99) (48-94) (24-92) (96-225) (51-109) (12-70) (52-95) (49-104) (24-98) (35-124) 30+ Obs SMR (95% Cl) 818 179 51 5 11 19 0 1 11 49 47 2 23 4 3 5 23 2 1 14 6 2 1 502 327 35 14 11 17 5 26 16 5 8 1,691 25,908 83 *" 88 86 107 102 94 0 ** 22 101 74 " 75 71 109 53 66 137 132 66 95 186 106 92 42 90 87 * 47 "' 51 ** 71 49 "* 36 69 62 " 54 83 (78-89) (76-102) (64-114) (35-249) (51-182) (56-146) --- (1-121) (50-180) (55-98) (55-100) (9-257) (69-164) (15-136) (14-101) (45-320) (84-198) (8-237) (2-527) (101-311) (39-230) (11-331 ) (1-234) (82-98) (78-97) (33-65) (28-85) (35-126) (29-79) (12-84) (45-101) (35-100) (18-126) (36-164) Table Vl. Observed Deaths, SMRs and 95% CI for Selected Causes of Death Among All Male Employees In the Beaumont Refinery Cohort by Time Since First Employment Cause of Death 18th ICDA) All Causes All Cancers (140-209) Cancer of Digestive System (150-159) Cancer of Esophagus (150) Cancer of Stomach (151) Cancer of Large Intestine (153) Cancer of Rectum (154) Cancer of Liver (155-156) Cancer of Pancreas (157) Cancer of Respiratory System (160-163) Cancer of Lung (162-163) Cancer of Skin (172-173) Cancer of Prostate (185) Cancer of Bladder (188) Cancer of Kidney (189) Cancer of Brain and CNS (191-192) Lymphatic and Hematopoietic Cancer (200-209) Lymphosarcoma and Reticulosarcoma (200) Hodgkin's Disease (201) Leukemia& Aleukemia (204-207) Other Lymphatic Tissue Cancer (202, 203, 208) Benign Neoplasms (210-239) Diseases of Blood (280-289) Diseases of Circulatory System (390-458) Arteriosclerotic Heart Disease (410-413) Nonmalignant Respiratory Disease (460-519) Pneumonia (480-486) Emphysema (492) Diseases of Digestive System (520-577) Cirrhosis of Uver (551) Accidents, Poisonings, &Violence (E800-E998) Accidents (800-949) Motor Vehicle Accidents (810-827) Suicide/950-959) Number of Persons at Risk N umber of P erson-Years * Statistically significant at 0.05 leve ** Statistically significant at 0.01 level Obs 231 40 14 1 6 2 1 1 2 9 9 3 2 0 0 2 8 1 1 2 4 0 0 92 66 3 2 0 10 5 55 37 14 12 5,601 73,429 < 20 SMR 62 ** 71 " 85 65 153 46 55 67 71 56 60 195 106 0 0 75 108 67 62 67 302 0 0 64 *" 74 * 18 ** 26 * 0 48 * 46 63 ** 67 * 50 ** 76 (95% CI) (54-71) (51-96) (47-143) (2-362) (56-333) (6-166) (1-308) (2-373) (9-258) (25-106) (27-113) (40-568) (13-384) -- --(9-270) (47-213) (2-373) (2-347) (8-242) (82-773) ----(52-79) (57-95) (4-54) (3-94) --(23-89) (15-107) (48-82) (47-93) (27-84) (39-133) Time Since First Employment Hears) 20-39 Obs 949 218 50 4 12 9 2 2 16 75 74 5 20 6 4 6 29 3 2 14 9 2 3 514 351 42 9 17 35 10 43 30 9 7 4,078 62,124 SMR 78 "* 84 * 65 "* 47 71 41 26 ** 31 113 84 88 133 105 83 66 98 136 73 103 165 135 63 105 81 "* 85 * 58 ** 33 "' 120 62 *" 35 ** 54 ** 57 ** 42 39 * (95% CI) (73-83) (73-96) (48-85) (13-121) (37-125) (19-77) (3-94) (4-113) (65-184) (66-105) (69-110) (43-311) (64-162) (30-180) (18-169) (36-213) (91-196) (15-214) (13-374) (90-276) (62-257) (8-228) (22-307) (74-88) (77-95) (42-78) (15-62) (70-192) (43-87) (17-65) (39-72) (38-81) (19-81) (16-81) 40+ Obs SMR 1062 232 59 5 14 24 2 1 13 70 68 1 31 5 4 3 26 4 0 12 10 3 1 627 393 55 23 20 24 5 24 15 9 8 2,279 19,539 94 " lO3 89 88 105 32 20 108 104 106 40 92 58 91 109 140 145 148 148 129 36 96 95 59 *T 60 * 120 73 49 69 59 * 111 119 (95% Cl) (88-100) (9o-117) (68-115) (29-205) (66-157) (4-114) (1-11 o) (57-185) (81-131) (82-134) (1-220) (63-131) (1 9-136) (25-234) (23-319) (91-205) (40-372) . (77-259) (71-272) (27-376) (1-199) (89-104) (86-105) (45-77) (38-90) (73-185) (47-109) (16-113) (44-103) (33-97) (51-211) (51-234~ Table VII. Observed Deaths, SMRs and 95% CI for Selected Causes of Death Among White Male Employees in the Beaumont Refinery Cohort by Time Since First Employment Cause of Death (8th ICDA) All Causes All Cancers (140-209) Cancer of Digestive System (150-159) Cancer of Esophagus (150) Cancer of Stomach (151) Cancer of Large Intestine (153) Cancer of Rectum (154) Cancer of Liver (155-156) Cancer of Pancreas (157) Cancer of Respiratory System (160-163) Cancer of Lung (162-163) Cancer of Skin (172-173) Cancer of Prostate (185) Cancer of Bladder (188) Cancer of Kidney (189) Cancer of Brain and CNS (191-192) Lymphatic and Hematopoietic Cancer (200-209) Lymphosarcoma and Reticulosarcoma (200) Hodgkin's Disease (201) Leukemia& Aleukemia (204-207) Other Lymphatic Tissue Cancer (202, 203, 208) Benign Neoplasms (210-239) Diseases of Blood (280-289) Diseases of Circulatory System (390-458) Arteriosclerotic Heart Disease (410-413) Nonmalignant Respiratory Disease (460-519) Pneumonia (480-486) Emphysema (492) Diseases of Digestive System (520-577) Cirrhosis of liver (551) Accidents, Poisonings, & Violence (ES00-E998) Accidents (800-949) Motor Vehicle Accidents (810-827) Suicide/950-959) Number of Persons at Risk Number of Person-Years Statistically significant at 0.05 level * Statistically significant at 0.01 level Obs 188 33 8 0 2 2 1 1 2 9 9 3 2 0 0 2 7 1 1 1 4 0 0 77 56 1 1 0 10 5 45 31 12 12 4,656 62,035 < 2O SMR (95% CI) 66 ** 69 * 61 0 69 52 64 91 84 65 69 206 139 0 0 81 108 75 70 38 363 0 0 67 ~ 71 ** 8 ** 20 0 62 58 70 * 70 * 52 * 85 (57-76) (48-97) (26-120) --(8-249) (6-189) (2-355) (2-506) (10-305) (30-123) (32-131) (42-601) (17-503) ----(10-293) (43-222) (2-416) (2-3g0) (1-214) (99-929) ----(53-83) (54-92) (0-47) (1-112) --(30-114) (19-135) (51-93) (48-100) (27-91) 144-148) Time Since Firsl 20-39 Obs SMR (95% CI) 797 183 38 4 6 8 1 2 12 66 65 4 15 4 4 6 28 3 2 13 9 1 2 430 304 36 7 16 28 9 37 27 8 7 3,443 52,135 84 ** 88 64 ** 75 51 42 = 15 * 42 106 90 93 115 114 66 75 107 154 * 83 120 177 165 40 89 85 * 87 * 63 "* 37 ** 126 62 ** 38 *" 60 -67 48 * 42 * (78-90) (76-101) (45-87) (21-193) (19-112) (18-84) (0-86) (5-152) (55-185) (69-114) (72-119) (31-295) (64.188) (18-169) (21-193) (39-233) (102-223) (17-243) (15-435) (94-302) (76-314) (1-222) (11-320) (77-94) (77-97) (44-87) (15-76) (72-205) (41-89) (17-72) (42-83) (44-98) (21-95) (17-86) 40+ Obs SMR 809 165 47 5 11 20 1 0 10 45 43 1 20 5 3 2 19 3 0 10 6 3 1 492 317 43 18 15 17 4 18 10 5 7 1,825 15,623 90 * 93 92 134 127 106 19 0* 106 82 82 43 87 69 80 80 123 125 0 145 112 162 44 95 92 56 "* 61 * 101 64 46 67 51 * 81 112 (95% CI) (84-97) (79-108) (68-123) (43-312) (63-227) (65-164) (1-106) --- (51-195) (60-110) (60-111) (1-239) (53-134) (22-160) (17-234) (1 0-290) (74-191) (26-365) --- (69-266) (41-244) (33-474) (1-247) (86-103) (82-103) (40-75) (36-96) (56-166) (37-102) (13-118) (40-106) (25-94) (26-188) (45-231/ Table VIII. Observed and Expected Deaths and SMRs by Decade of end Age at Death for Whlte and Nonwhite Males 1950s 1960s 1970s 1980s Age at Deat.___~h < 65 65 Total <65 65+ Total <65 65+ Total <65 65+ To~I TOTAL Leukemi~a ,O 3 0 3 White Males E 1.71 0.56 2.27 SM.._.~R 175 0 132 3 1.94 7 1.87 10 3.81 155 374 263 2 1.80 8 3.44 10 5.24 111 233 191 0 1.46 1 3.62 1 5.08 0 28 20 24 16.8g 142 NonWhite Males O__ E SMR 0 0.32 0 0.10 0 0.42 0 0 0 1 0.30 333 0 0.33 0 1 0.63 159 1 0.28 357 1 0.56 179 2 0.84 238 0 0.20 0 1 0.52 192 1 0.72 139 4 2.70 148 Decade of Death 1950s 1960s 1970s 1980s Age at Deat.._.~h < 65 65+ Total "Other Lymphatic Tissue" Cancer White Mates O E SM~R 1 0.73 0 0.15 1 0.88 137 0 114 <65 65+ Total 0 0.83 3 0.82 3 1.65 0 366 182 <65 65+ Total 5 1.46 1 2.42 6 3.88 343 * 41 155 <65 65+ Total 4 1.65 5 3.82 9 5.47 242 131 165 TOTAL 19 11.90 160 NonWhite Males O E SM,._~R 0 0.26 0 1 0.02 5000 1 0.28 357 0 0.29 0 1 0.26 385 1 0.55 182 0 0.30 0 1 0.66 152 1 0.96 104 0 0.28 0 1 0.77 130 1 1.05 95 4 2.85 140 * Statistically significant at 0.05 level ** Statistically significant at 0.01 level Table IX. Observed and Expected Deaths from Lymphatic and Hematopoletlc Cancers Among All Males by Year of First Employment (<1950, 1950+) Cause of Death (Sth iCDA) Lympb, opoietic Cancer Lymphosarcoma & Reticulosarcoma Hodgkin's Disease Leukemia Other Lymphatic Tissue Statistically significant at 0.05 level < 1950 Obs Expected SMR 55 40.8 135"(102-176) 8 7.3 .110 (4~-217) 3 3.5 85 (18-249) 25 17.0 147 (95-217) 18 12.7 141 (84-224) Year of First EmpIo),m, ent 1950+ Obs Expected SMR (95%) Obs 6.6 122 (53-241) 63 8 3 2.6 117 (24-343) 28 2.0 248 (80-578) 23 Total 47.3 133" (102-170) 8.3 96 (41-189) 4.4 68 (14-200) 19.6 143 (95-207) 14.8 156 (99-234) Table X. Observed and Expected Deaths, SMRs and 95% CI forLeukemla Cell Types,NonHodgkin's Lymphoma, and Multiple Myeloma Among Male Employees Cause of Death (ICD-8) ALL (204.0) CLL (204.1) AML (205.0) CML (205.1) NHL (200,202) MM (203) Observed 0 1 8 3 22 9 Deaths Expected 1.1 4.0 5.9 2.6 15.7 7.5 SMR (95% CI) 0 --- 25 (1-139) 136 (59-268) 115 (24-336) 140 (88-211) 121 (55-230) Table Xl. Observed and Expected Deaths, SMRs and 95% CI by Cause for Male Beaumont Refinery Workers with > 6 Months Employment In Operator Jobs Number of Persons: 1,240 Cause of Death (Sth ICDA~) , All Causes of Death Al~ Malignant Neoplasms All Infective and Parasitic Disease Cancer of Buccal Cavity and Pharynx Cancer of Digestive Organs and Peritoneum Cancer o1 Esophagus Cancer of Stomach Cancer o1 Large Intestine Cancer of Rectum All Cancer o1 Liver Cancer of Pancreas Cancer of Respiratory System Cancer of Larynx All Cancer of Lung - Pdmary and Secondary Cancer of Bone Cancer of Skin Cancer of Proslate (Males Only) Cancer of Testis (Males Only) Cancer of Bladder Cancer of Kidney Cancer of Brain and Other Central Nervous System All Lymphopoielic Cancer Lymphosarcoma and Reticulosarcoma Hodgkin's Disease Leukemia and Aleukemla Cancer of Other Lymphatic Tissue Benign Neoplasms Diabetes Mellilus All Diseases of Blood and Blood-Forming Organs All Diseases of Nervous System and Sense Organs All Diseases of Circulatory System Chronic Rheumatic Heart Disease Arteriosclerotic Heart Disease, Including CHD All Vascular Lesions of CNS All Respiratory Diseases All Pneumonia Emphysema All Diseases of Digestive System Cirrhosis of Liver AllDiseases of Genito-Udnary System Symptoms, Senility & Iti Defined Conditions All External Causes of Dealh All Accidents Motor Vehicle Accidents Suicide Unknown Cruises (In All Causes Category Only) Observed Deaths 501 113 3 1 23 1 4 10 1 1 5 36 0 36 0 3 17 . 0 1 3 4 12 2 0 5 5 2 6 2 3 301 2 204 39 33 11 13 8 1 5 9 12 8 2 3 1 Expected Deaths 761,3 148.7 14.1 4.1 44.6 4.4 9.4 13.7 4.4 3.6 8.0 45.9 2.1 43.3 0.6 1.9 16.8 0.6 5.0 3.1 2.7 12.7 2.2 1.1 5.3 4.0 1.8 11.4 1.9 6.4 409.5 5.1 258.1 67.3 53.1 22.1 9.6 28.7 11.9 15.5 10.6 48.9 32.7 13.4 9.2 J SMR Person-Years: 34,739 Confidence Umlts Lower Upper 66 ** 76 ** 21 ** 24 52 ** 23 42 73 23 28 62 78 0 ...... 83 0 ...... 156 101 0 ...... 20 97 146 95 91 0 ...... 94 i25 109 53 105 47 74 ** 39 79 ** 58 ** 62 ** 50 * 135 28 ** 8 ** 32 ** 85 25 ** 25 ** 15 ** 33 * 60 63 4 1 33 1 12 35 1 1 20 55 58 32 59 1 20 40 49 11 31 40 13 19 13 10 65 5 69 41 43 25 72 12 0 11 39 13 11 2 7 72 91 62 135 77 126 109 134 127 154 146 109 115 456 162 113 282 374 166 330 220 291 393 115 378 137 82 142 91 79 87 89 231 55 47 75 161 43 48 54 95 * Slatistically Significant at 0.05 Level * Statistically Significant at 0.01 Level Table Xll. Observed and Expected Deaths, SMRs and 95% CI by Cause for Male Beaumont Refinery Workers with > 6 Months Employment In Maintenance Craft Jobs Number of Persons: 5,360 Cause of Death (8th ICDA} All Causes of Death All Malignant Neoplasms All Infective and Parasitic Disease Cancer of Buccal Cavity and Pharynx Cancer of Digestive Organs and Peritoneum Cancer of Esophagus Cancer of Stomach Cancer of Large Intestine Cancer of Rectum All Cancer of Uver Cancer of Pancreas Cancer of Respiratory System Cancer of Larynx All Cancer of Lung - Primary and Secondary Cancer of Bone Cancer of Skin Cancer of Prostate Cancer of Teslis Cancer of Bladder Cancer of Kidney Cancer of Brain and Other Central Nervous System All Lymphopoie~c Cancer Lymphosarcoma and Reticulosarcoma Hodgkln's Disease Leukemia and Aleukemia Cancer of Other Lymphalic Tissue Benign Neoplasms Diabetes Mellilus All Diseases of Blood and Blood-Forming Organs All Diseases of Nervous System and Sense Organs All Diseases of Circulatory System Chronic Rheumatic Head Disease Adedosclerotic Head Disease, Including CHD All Vascular Lesions of CNS All Respiratory Diseases All Pneumonia Emphysema All Diseases of Digestive System Cirrhosis of Uver All Diseases of Genito-Udnary System Symptoms, Senility & III Defined Conditions All External Causes of Death All Accidents Motor Vehicle Accidenls Suicide Unknown Causes (In All Causes Category Only) * Statistically Significant at 0.05 Level * Statistically Significant at 0.01 Level Observed Deaths 1559 358 9 4 86 7 18 28 4 2 21 111 2 109 0 5 46 1 8 6 7 45 4 1 21 18 4 27 2 17 873 9 568 141 80 31 26 44 10 11 46 74 51 19 17 7 Expected Deaths 1706,4 317.7 35.6 9.7 97.5 9.8 21.9 28.3 10.1 7.9 17.4 96.7 4.8 90.8 1.5 4.6 31.6 1.6 10.2 6.9 6.9 28.1 5.4 3.0 11.7 7.7 4.2 24.8 4,3 14.0 902.3 13.9 565.6 148.2 107.5 46.2 20.9 71.3 31.7 35.0 25.0 141.2 93.0 40.5 26.6 SMR Person.Years: 105,370 Confidence Umlts Lower Upper 91 ** 113 * 25 ** 41 88 72 82 99 40 25 * 121 115 42 120 0 ...... 110 145 * 65 78 87 101 160 ** 74 33 179 * 233 ** 95 109 47 121 97 65 1 O0 95 74 ** 67 * 125 62 ** 32 ** 31 ** 184 ** 52 ** 55 ** 47 ** 64 87 101 12 11 71 29 49 66 11 3 75 95 5 99 36 107 2 34 32 41 117 2O 1 111 138 26 72 16 71 90 30 92 80 59 46 81 45 15 16 135 41 41 28 37 96 125 48 106 109 148 130 143 102 92 185 138 151 145 256 194 359 155 190 208 215 189 186 273 368 242 158 170 194 103 123 109 112 93 95 183 83 58 56 246 66 72 73 103 Table Xlll. Selected Cause-Specific Mortality Among Beaumont Males Employed ~ 6 Months in Maintenance Craft Jobs by Hire Pedod All Causes All Cancer Digestive Cancer Lung Cancer Skin Cancer Prostate Cancer Kidney Cancer Brain & CNS Cancer Leukemia Other Lymphatic Tissue Blood Diseases Non-malignant Respiratory Diseases Obs 1428 334 81 106 3 43 5 6 19 14 2 76 "Statistically significant at 0.05 level. ** Statistically significant at 0.01 level. < 1950 Exp SMR 1554.0 92** 89.8 115" 90.7 89 81.6 130' 3.7 81 30.7 140" 6.2 81 5.7 105 10.4 183" 6.8 205* 3.9 ' 52 100.7 76* (95%CI) (87-97) (103-128) (71-111) (106-157) (17-236) (102-189) (26-190) (38-228) (110-285) (112-344) (6-187) (60-94) Obs 131 24 5 3 2 3 1 1 2 4 0 4 1950+ Exp SMR 152.4 86 27.9 86 6.8 74 9.3 32* 0.8 237 1.0 305 0.7 138 1.2 84 1.3 149 0.9 443* 0.4 0 6.8 59 (95% CI) (72-102) (55-128) (24-173) (7-95) (29-856) (63-891) (3-767) (2-465) (18-539) (121-1134) (16-150) Table XIV. Observed and Expected Deaths, SMRs and 95% CI for Hematopoletlc Cancers Among Male Maintenance Craftworkers by Duration of Employment Cause of Death (~CD-8) ALL (204.0) OBS EXP SMR (95%Cl) <10 0 0.2 0 Duration of Employment (Years) 10-29 30+ 00 0.3 0.2 00 Total 0 0.7 0 CLL (204.1) OBS 01 0 EXP 0.2 1.2 0.9 2.3 SMR (95%CI) 0 83 (1-460) 0 43 (1-242) AML (205.0) OBS EXP SMR (95%CI) 1 0.6 167 (2-976) CML (205.1) OBS EXP SMR (95%CI) 0 0.3 0 3 1.9 158 (32-459) 1 0.9 115(2-640) 37 1.1 3.6 273 (55-797) 194 (78-401) Chi-square trend = 0.33 12 0.5 1.6 213(3-1184) 124 (14-449) NHL (200,202) OBS EXP SMR (95%CI) 3 1.2 244 (49-713) MM (203) OBS EXP SMR (95%Cl) 1 0.4 238 (3-1325) 9 5.2 174 (79-330) 1 2.3 43 (1-238) 4 16 3.1 9.5 129 (35-329) 168 (96-273) Chl-square Irend = 0.70 46 1.6 4.4 247 (67-637) 137(50-299) Chi-square Irond = O. 78 Table XV. Lung Cancer Mortality Among Male Maintenance Craftworkers by Duration of Employment and Year of First Employment in Craft Jobs Duration of Employment (years) <10 10-19 2o-29 30-39 40+ Calendar Year of Hire <1920 1920-1929 1930-1939 1940-1949 1950+ "p<0,05 * p<0.01 Observed 4 20 46 30 9 Observed 19 33 11 43 3 SMR (95% CI) 48 (13-122) 109 (67-168) 154"* (113-206) 102 (69-146) 180 (83-343) Chl-square kend = 1.85 SMR (95% CI) 141 (85-220) 108 (75-152) 154 (77-275) 141" (102-190) 32* (7-95)