Document gbmY7O4mmXewa8JM5jBy1ykg9

c c American Journal of Industrial Medicine 11:189-202 (1987) Texaco Mortality Study: 111. A Cohort Study of Producing and Pipeline Workers Barbara J. Divine, PhD, and Virginia Barron, MS While there have been numerous epidemiology studies of refinery workers, no studies have been done on producing and pipeline workers. This is a retrospective follow-up study of all persons who were employed for at least 6 months at a Texaco producing or pipeline location and who worked at some time during the period 1946-1980. Of the 11,098white men in the cohort, 8,964 were alive, 1,886 were known to be dead, and - the vital status of the remaining 248 as of December 31, 1980 was unknown. The standardized mortality ratio (SMR)of 63 for all causes was significantly low, on the basis of 2,976 expected deaths. Statistically significant deficits also were seen for all major causes of death. Mortality patterns were also examined for the major job categories in these departments. Similar patterns of mortality were seen, although there was a significant excess of thyroid cancer in those employed as pumper-gaugers. However, it was based on only four cases. Key words: carcinogens, environmental, longitudinal studies, mortality, neoplasms, crude oil, occupational diseases INTRODUCTION The petroleum ration, production, industry is divided into five major operational segments: explotransportation, manufacturing, and marketing. While refinery , . workers (ie, manufacturing) have been the subject of many epidemiologic studies, there are little epidemiologic data on production and pipeline workers whose primary exposure is to crude oils, volatile light ends, and produced gases. Crude oils contain, in addition to other hydrocarbons, varying amounts of polycyclic aromatic compounds and dissolved gases. Certain crude oils are rich in sulfur-containing coinpounds (sour crudes) and may have high concentrations of hydrogen sulfide. Additionally, for the average worker, petroleum production may also include rare exposures to other potentially hazardous substances, such as form- aldehyde, asbestos, and mercury, as well as more frequent contact with caustics. There are also physical hazards such as explosive vapors, radioactive devices, and noise. Tcxaco Inc.. Houbton, Texis (B.J.D., V.B.). hls. Barron's currcnr addrcss i s EDS COT.. 7171 Forest L n e , One Forest Plaw. Suite 210. Dallas. TX 75230. Address reprint requests 10 Dr. Barbara Divine. Tcxaco Inc.. P.O. Box 1404. Housron, TX 77251. Acceprcd for publrca~ronJuly 8. 1986. 0 1987 Nan K. Liss, Inc. 190 Divine and Barron A number of studies have demonstrated that either high- or low-sulfur crude oil will produce skin tumors on rodents following repeated skin applications. Such studies have shown the effect both for crude oil and for fractions boiling at greater than 650F. A recent toxicology study found certain crude oil fractions to be somewhat to moderately carcinogenic (American Petroleum Institute, unpublished study). In this mouse skin bioassay, 15-50% of the animals developed skin tumors, most of which were malignant. However, an earlier study in humans found no excess incidence of skin cancer in workers potentially exposed to the high-boiling, catalyticallycracked oils that are carcinogenic in toxicology testing w a d e , 19631. The fact that no adverse effect was seen may reflect industrial hygiene and personal hygiene practices, and the study may not be completely relevant to producing workers exposed only to crude oil and not to cracked crudes. Although no detailed mortality studies have been reported for workers in the production area of the petroleum industry, a number of studies have been conducted on workers in other segments. Recent epidemiology studies of refinery workers have shown elevated mortality rates for various causes of death including brain tumors, leukemia, stomach cancer, and pancreas cancer [Thomas et al, 1982; Hanis et al, 1982; Divine et al, 19851; however, rates for the major causes of death were reduced. A study of marketing workers in Great Britain showed elevated mortality for kidney cancer (L. Rushton and M.R. Alderson, unpublished). However, owing to the differences in working conditions and in substances handled, results of the above studies may not be applicable to producing workers. Producing workers are exposed principally to crude oil, and the primary route of exposure is by skin contact. Refinery workers, on the other hand, may handle the whole range of crude oil to finished products, including chemicals used in the manufacturing process, whereas marketing workers handle only finished products. For the last two groups, the primary route of exposure is inhalation. This study is a retrospective follow-up study-of Texaco producing and pipeline workers who were employed for at least 6 months and who worked between January 1, 1946 and December 31, 1980. The major objective of the study was to determine the patterns of mortality in these workers. METHODS The study cohort consists of all employees of Texaco Inc. who ever worked for at least 6 months at a US producing or pipeline location and who also worked for at least 1 day between January I, 1946 and December 31, 1980. Employees who met the eligibility requirements were identified by examination of all personnel folders at the division and district offices. Data collected on each employee included name, social security number, race, sex, date of birth, and a complete work history, if available. Vital status information for those employees not known to be either dead or still active was obtained from the Social Security Administration (SSA). For those Texas residents with vital status unknown to the SSA, information was requested from the Texas Department of Public Safety. If no other information was available, vital status from the company records was used. Copies of death certificates were obtained either from company files at the time of data collection or from the states where the deaths occurred. All death certificates were coded for underlying cause of death by a nosologist according to the Texaco Producing and Pipeline Workers 191 Eighth Revision rules of the International Classification of Diseases (8th ICD). No conversion to an earlier revision was made as in an earlier Texaco study [Divine et al, 1985; Divine and Barron, 19861. Complete work histories for each employee were coded. Each job consisted of a unit (eg, maintenance, office), a task (eg, roustabout, pumper-gauger), and the dates that the position was held. Thejob codes were based on the American Petroleum Institute Job Code Classification System of Petroleum Refineries and Selected Petro- chemical Operations [19791 and were developed further in consultation with Texaco industrial hygienists. Job history coding accounted for all periods of time between the cohort mem- ber's initial hire date and final termination date or the study end date, whichever came first. Periods of not working for 1 month or longer were coded as such. When more than one job was held concurrently, all were coded with the same dates, and the person-years were divided evenly among the jobs. Analysis of mortality was performed using Monson's computer program [Mon- son, 19741 to compare the observed and expected number of deaths, with the US white male population as the comparison group. Person-years were accrued beginning when the person met the cohort eligibility requirements or with the study start date, whichever came later. Those persons with vital status unknown were assumed to be lost to study on the date last observed. For subcohon analyses, ie, for those employed for a certain number of years or in a specific job group, person-years were accrued when the person had met the eligibility criteria or on the study start date, whichever came later. Because of the large number of different tasks and units, task and unit categories with similar responsibilities and potential exposures were grouped into larger cate- gories. The groups were chosen by a Texaco industrial hygienist most familiar with producing and pipeline operations. Results are presented for those employed for at least 5 years in the grouped jobs. - -- RESULTS There were 12,894 persons eligible for inclusion in the study cohort. Table I shows the number of subjects according to sex, race, and vital status as of the study end date, December 31, 1980. Those persons with unknown race (n = 563) were assumed to be white. White males made up 86.1% of the total cohort and 96.9% of the decedents. All results reported hereafter are for the cohort of white males. Of the 11,098 total white male cohort, 8,964 (80.8%) were alive at the study end date, 1,886 (17.0%)were deceased, and 248 (2.2%) had vital status unknown. TABLE I. Texaco Producing and Pipeline Workers 1946-1980 by Race. Sex. and Vital Status Race and sex Alive Vital status: No. (percent) Dead Unknown Tola1 White males Nonwhite males White females Nonwhite females Total 8,964 (80.8) 420 (97.2) 971 (74.9) 64 (95.5) 10.419 1,886 (17.0) 12 (2.8) 48 (3.7) 0 1.946 248 (2.2) 0 278 (21.4) 3 (4.5) 529 11,098 432 1,297 67 12.894 192 Divine and Barron Y TABLE 11. Texaco Producing and Pipeline Workers: SMRs for Selected Causes of Death by ' Length of Employment for White Males, 1946-1980' Cause (ICD 8th) Employed 6 months or more Observed deaths Expected deaths SMR 95% C.I. All causes (001-998) All infective and parasitic diseases (OOO- 139) All cancer (140-209) Cancer of digestive system (150-159) Cancer of stomach (151) Cancer of large intestine (153) Cancer of pancreas (157) Cancer of lung (162) Cancer of prostate (185) Cancer of kidney (189) Cancer of bladder (188) Cancer of skin (172,173) Cancer ofbrain (191,192) Cancer of thyroid (193) Lymphosarcoma (200) Hodgkin disease (201) Leukemia (204-207) Cancer of other lymphatic tissue (202,203,208) Benign neoplasms (210-239) Endocrine, metabolic diseases (240-279) Diseases of circulatory system (390-458) Arteriosclerotic heart disease (410-413) Vascular lesions of CNS (430-438) Respiratory diseases (460-519) Emphysema (492) Diseases of digestive system (520-577) Cirrhosis of liver (571) Discases of G-Usystem (580-629) External causes of death (800-998) Accidents (800-949) 1.886 2,976.4 63 61-66 17 42.9 40 23-63 393 581.1 68 61-75 105 168.7 62 51-75 13 33.0 39 21-67 40 54.1 74 53-101 27 32.0 &I 56-123 109 179.0 61 50-73 41 43.7 94 67-127 5 14.1 35 11-83 16 19.0 84 48-137 7 9.8 72 29-148 11 15.9 69 34-124 4 1.2 342 92-876 7 11.8 59 24-122 4 6.6 60 16-154 25 22.7 110 71-163 12 13.5 89 46-155 11 7.7 I42 71-255 13 48.7 27 14-46 1.041 1.580.3 753 - 1,085.5. 66 62-70 69 65-75 160 218.5 73 62-86 103 184.9 56 45-68 29 46.1 63 42-90 50 136.2 37 27-48 13 64.7 20 11-34 17 44.0 39 23-62 123 260.4 47 39-56 91 177.6 51 41-63 *C.I., confidence interval. Over one-third were active employees, and nearly one-haif had either retired or resigned by the end of the study period. The cohort is fairly young, with 26%hired in the last decade of the study. Most of the deaths occurred in older employees, 70% with 20 years or more of service. There were 220,414 person-years of observation. Death certificates were obtained for all but 65 (3.4%)of the decedents. These 65 deaths are included in the all- causes standardized mortality ratio (SMR) but not in the cause-specific SMRs. Only SMRs 2 110 and based on four or more deaths will be mentioned in the text. Table I1 shows the observed and expected numbers of deaths, the SMRs, and their 95%confidence intervals for the total cohort, for those employed for 5 years or more with 20 years latency (n = 5,637). and for those employed for 20 years or more (n = 4,575). There were 2,976.4 deaths expected for the total cohort, giving Texaco Producing and Pipeline Workers 193 c TABLE 11. Texaco Producing and Pipeline Worken: SMRs for Selected Causes of Death by Length of Employment for White, 1946-1980* (continued) Employed 5 years or more with 20 years latency Observed Expected deaths deaths SMR 95 % C.I. Employed 20 years or more Observed Expected deaths deaths SMR 95 % C.I. 1,641 2.4 13.8 68 65-71 1,309 2,050.8 64 60-67 14 26.4 53 29-89 14 22.3 63 34-106 354 493.1 72 65-80 280 421.8 66 59-75 97 144.1 67 55-82 74 123.0 60 47-76 13 27.3 48 25-81 10 23.3 43 21-79 36 47.1 76 54-106 29 40.2 72 48-104 27 27.7 97 64-142 20 23.7 84 52-130 99 156.2 63 52-77 77 134.5 57 45-72 40 41.5 96 69-131 35 35.0 100 70-139 5 11.8 42 14-99 4 10.1 40 11-101 15 17.2 87 49-144 11 14.7 75 37-134 5 7.0 71 23- 166 3 5.9 50 10-147 7 11.0 64 26-132 5 9.4 53 17-124 4 1.o 423 114-1,083 4 0.8 496 133-1.270 6 9.1 66 24-144 4 7.7 52 14-132 3 3.6 84 17-245 3 3.0 99 20-290 19 17.6 108 65-168 14 14.9 94 51-157 11 11.8 94 47-168 11 10.0 110 55-196 9 5.6 162 74-308 6 4.7 127 46-277 10 40.9 24 12-45 8 34.7 23 10-45 937 1,358.4 69 65-74 756 1.153.5 66 61-70 669 935.7 71 66-77 533 796.1 67 61-73 -149 193.9 77 65-90 126 163.3 77 64-92 96 162.9 59 48-72 25 42.6 59 38-87 41 102.8 40 29-54 9 47.0 19 9-36 17 34.2 50 29-80 66 123.5 53 41-68 50 83.4 60 44-79 78 138.4 56' 45-70 22 36.4 60 38-92 34 87.4 39 27-54 6 40.0 15 5-33 14 28.8 49 27-82 44 103.3 43 31-57 35 69.8 50 35-70 an all-causes SMR of 6 3 . This deficit is statistically significant. Large and statistically significant deficits are seen for almost every major cause of death including all cancer (SMR = 68), cancer of the lung (SMR = 62), arteriosclerotic heart disease (SMR = 70),vascular lesions of the central nervous system (SMR = 71), respiratory disease (SMR = 58), and all external causes of death (SMR = 47). For the cohort as a whole, there are a few causes of death in which the SMR is greater than 100; however, in each instance, the lower confidence limit does not exceed 100. The highest SMR (342) is for cancer of the thyroid, with four deaths observed and 1.2 expected. Other causes of death in which the SMR is greatcr than 100 are leukemia (SMR = 110), benign neoplasms (SMR = 142), and symptoms, senility, and ill defined conditions (SMR = 116). The patterns of mortality are essentially the same for the two subgroups, although the SMR for cancer of the thyroid becomes statistically significant. m 194 Divine and Barron mZ Nw mn mn VI 0, 0,o e1 m m 0 d e4 m N WWmb "Wln-r-ro m r- r4- e Texaco Producing and Pipeline Workers TABLE IV.Texaco Producing and Pipeline Workers: Activities Associated With Grouped Job Categories Job category Activities Unexposed Pumper Roustabout Pipeline, field Maintenance Gas plant Drilling Office work, supervision, engineering Responsible for running of wells, pumps, and other oil field equipment; determines oil production of leases and wells Performs general oil field or gas plant construction and repair: entry level position Performs maintenance on pipelines and pump stations, operates pump stations, gauges tanks Responsible for construction, repair, and maintenance of producing and pipeline facilities and equipment Inspects and maintains mechanical equipment at a gas plant Responsible for the operation of equipment on the drilling site 195 Table III shows the observed and expected numbers of deaths, the SMRs, and their 95 % confidence intervals for producing workers (n = 8,943) and for pipeline workers (n = 2,169). There are 14 people (two deaths) who are in both groups. Significant deficits of mortality are again seen for all of the major causes of death. For producing workers, a nonsignificant elevation is seen for benign neoplasms (SMR = 126). A significantly elevated SMR is seen for cancer of the thyroid (SMR = 483), based on four deaths observed and 0.8 expected. For the pipeline workers, elevated SMRs are seen for lymphosarcoma (SMR = 117), leukemia (SMR = 139), and benign neoplasms (SMR = 179). None of these issignifieant. Work History Analysis There were 11,034 white males with work histories available for analysis. The job categories examined were as follows: unexposed, pumper in producing, roustabout in producing, pipeline field personnel, maintenance, gas plant, and drilling. Table IV shows some of the job activities associated with each category. Table V shows the observed and expected numbers of deaths, the SMRs, and their 95% confidence intervals for those employed 5 years or more in the job groups. The allcauses SMRs show statistically significant deficits, as do the major causes of death. For the group of workers with little or no likely exposure, primarily office workers and supervisory personnel, deficits are seen for almost all causes of death. The only elevated SMRs are for cancer of the brain (SMR = 137) and cancer of other lymphatic tissue (SMR = 156). Neither is statistically significant. For producing workers who worked 5 years or more in pumper-gauger-type jobs, slight nonsignificant elevations are seen for cancer of other lymphatic tissue (SMR = 127) and for cancer of the bladder (SMR = 111). All four workers who developed cancer of the thyroid worked for 5 years or more in this job category and the SMR is significantly elevated (SMR = 975). Two of the thyroid cancer decedents also worked for more than 5 years in roustabout-type jobs in -producing (SMR = 556). The roustabout category also shows a nonsignificant SMR for leukemia (SMR = 119). 196 Divine and Barron -I -~mW l n _ N l - WW om 3 8y -ymy & n- vol m0 -owol-l--o ON N PIN -0-m I n ~ P Nm m N m - m O - N m v l P I v l8w "m U .u-l Texaco Producing and Pipeline Wprkers 197 -2 .-m(A .a 198 Divine and Barron oopg - CVCII Texaco Producing and Pipeline Workers 199 c U Iv?1 .ecn- 200 Divine and Barron w For those employed for 5 years or more as pipeline workers (excluding ofice and supervisory personnel) ,elevated SMRs are seen for cancer of the pancreas (SMR = 121) and leukemia (SMR = 117). Neither is significant. For those employed as maintenance workers in either producing or pipeline for 5 or more years, the only elevated SMR is for leukemia (SMR = 144). For the gas plant job category, no elevated SMRs are seen for any cause with four or more deaths. For the drilling category, elevated SMRs are seen for cancer of the prostate (SMR = 177), vascular lesions of the CNS (SMR = 113), and diseases of the genitourinary system (SMR = 172). Again, none of these is significant. DISCUSSION The most striking result of this study is the significantly low overall SMR that is seen for all the major causes of death and among all the subcohorts examined. This SMR (63) is lower than that found in most of the other cohort mortality studies in the petroleum industry [Hanis et al, 1982; Divine et al, 1985; Rushton and Alderson, unpublished; Wen et al, 19831. Even when the analysis is restricted to those employed for 20 years or more or to those with 20 years latency, the SMRs increase only slightly. In addition to the healthy worker effect, it is believed that the wages and benefits provided to Texaco employees have had beneficial effects on mortality. These effects appear to be more marked for diseases in which medical care and preventive medicine have an impact, eg, infectious disease and heart disease. The highest SMR seen was for cancer of the thyroid; however, it is based on only four cases. The cases are all long-term producing department employees who were employed predominantly as pumper-gaugers. The expected exposures in this job are to crude oil, hydrocarbon vapors, hydrogen sulfide, caustic, water treatment chemicals, and noise; none has been associated with thyroid cancer. The SMR for benign and unspecified neoplasms was aIso elevated. Of the 11 deaths, five were unspecified brain tumors, and one was a benign brain tumor. There were 3.7 expected unspecified brain tumors giving an SMR of 135. There were 1.2 benign brain tumors expected. If ail brain tumors, malignant, benign, and unspecified, are combined, the SMR for the group is 84. If the analysis is restricted to those employed for 5 years or more with 20 years of latency, the benign and unspecified brain tumor SMR increases to 152, but the all-brain-tumor SMR does not change. Therefore, there is overall no brain tumor excess. The brain tumor decedents did not appear to be restricted to any specific job category. No other clusters were seen in the benign and unspecified neoplasms category. Another elevatcd SMR seen for the entire cohort is leukemia. However, the excess is small (SMR = 110). and it disappears when the analysis is restricted to those employed for 20 years or more. While an increase is seen in the SMR for cancer of other lymphatic tissue for those employed for 20 years or more (SMR = 110), it is small and nonsignificant. The SMRs for the 20-year latency group are similar to those for the entire cohort for these two causes of death. Both the producing and pipeline subcohorts show patterns similar to those of the entire cohort, as do each of the job groups. Although some SMRs are elevated in Texaco Producing and Pipeline Workers 201 each group, none of the elevations is statistically significant except for cancer of the thyroid, and most are based on fewer than five deaths. NO further analysis of any of the causes of death appears indicated. . One surprising finding was the significantly low SMR (61) for cancer of the lung. Similarly low SMRs were seen for all of the subgroups analyzed. Although this was seen in the earlier Texaco study pivine et all 19851, those workers spent most of their working time in areas where smoking is prohibited. Producing and pipeline workers, on the other hand, work in a less restrictive outdoor environment and, therefore, have greater access to smoking opportunities. This deficit is in contrast to an earlier case-control study in Louisiana that found that oilfield workers had an odds ratio of 2.33 for lung cancer as compared with the control group [Gottleib, 19801. While an attempt was made to look for any cluster of deaths in specific job groups, the results from Table V should be interpreted with caution. It is not possible to identify specific chemicals and levels of exposures that are i i e d exclusively with a specific job. Jobs were grouped based on a company industrial hygienist's best estimate of those having similar responsibilities and exposures. It is likely that some of these exposures and work tasks may have changed substantially over time. Many of the people held a number of different jobs throughout their working time with Texaco and may be counted in more than one group. Based on a comparison with the earlier Texaco cohort, it is likely that few of this group had refmery-, research-, or petrochemical-related exposures. There were only 12 people who overlapped between the two cohorts [Divine et al, 1985; Divine and Barron, 19861. The major exposure for these workers is likely to be dermal exposure to crude oil. It has been suggested that this may lead to an increased risk of skin cancer. No excess was seen; however, because of the extremely low mortality associated with skin cancer, this type of study is not well suited to uncover any possible associations. A recent hospital-based case-control study of testicular cancer found a signifi- -- cant association with work in the crude petroleum and natural gas extraction industry, with an odds ratio of 2.29 [Mas et ai, 19841. However, no deaths from testicular cancer were seen in this study, although 3.2 were expected. The upper confidence limit is 115, which is well below the odds ratio seen in the case-control study. In summary, the mortality of Texaco producing and pipeline workers was substantially lower than expected, and large deficits were seen for all major causes of death. A further detailed examination of the cohort by grouped job categories did not uncover any jobs or diseases warranting concern. REFERENCES American Petroleum Institute (1979): "lob Code Classification System Part 1. Petroleum Refineries and Selected Petrochemical Operations." Washington, DC: American Petroleum Institute. Divine BJ. Barron V, Kaplan SD (1985): Texaco Mortality Study. I. Mortality among refinery, petrochemical. and research workers. J &cup Med 27:445-547. Divine BJ. Baron V (1986): Texaco Monality Study. ll. Patterns of mortality among white males by specific job groups. Am J Indust Med 10:371-381. Gottlieb MS (1980): Lung cancer and the petroleum industry in Louisiana. J &cup Med 22:384-388. Hank NM. Holmes TM, Shallcnberger LG. Jones KE (1982): Epidemiologic study of refinery and chemical plant workers. J Occup Med 24:203-212. Mills PK. Newell GR. Johnson DE (1984): Tuticulx cancer associated with employment in agriculture and oil and natural gas extraction. Lancet 1:207-210. 202 Divine and Barron Monson RR (1972): Analysis of relative survival and proportional mortality. comput Biomed Rcs 7~325-332. Thomas TL, Waxweiler RJ, Moure-Eraso R. Itaya S. Fraumeni JF (1982): Mortality patterns among workers in three Texas oil refineries. J Occup Med 24:135-141. Wade L (1963): Observations on skin cancer among refinery workers. Arch Environ Health 6730-735. Wen CP, Tsai SP, McClellan WA, Gibson RL (1983): Long-term mortality study of oil refinery workers. I. Mortality of hourly and salaried workers. Am J Epidemiol 118:526-542. ..* ,