Document 7Oor0ddNrB7o96zn0MOGkzowB

PLAINTIFF'S EXHIBIT DOW-1561 0 9 fj / . tiJ 0 lS Lung Cancer and the Petroleum Industry in Louisiana Death certificates from the period 1960 to 1975 for Louisiana were surveyed for lung cancer deaths and noncancer deaths among people employed in petroleum min ing and refining to determine if there were specific occupa tions within this industry associated with a higher risk of lung cancer. Workers employed as welders, operators, boilermakers, and painters and oilfield workers had the greatest risks. OH field miners bom outside Louisiana had a higher risk than native-born miners. Occupation as listed on death certificates provides a rapid means of identifying those with high risk, as a large pan of the population can be rapidly screened, covering employees in many in dustries. A study of geographic patterns of cancer in the United States revealed that parishes in Louisiana have some of the highest age-adjusted lung cancer mortality' rates for 1950 to 1969.' Moreover, these parishes are con tiguous in certain areas, particularly along the Culf Coast, the Mississippi River, and urban areas. To determine fac tors influencing lung cancer mortality rales in Louisiana, 19 parishes were selected for a case-control death cer tificate study. National correlation studies suggest that occupational exposures in the petroleum, paper, and chemical in dustries may account for some of the excess in lung cancer mortality rates for certain U.S. counties.1 The petroleum industry is a major industry in Louisiana, employing many of the people currently working in the petroleum industry in the U.S. An increased relative risk for lung cancer was noted primarily in men over 62 years of age in oil production and refinery industries, and in Auocuip Piofrtror o( Medic**- nd I puVmn-,to*y, Tuliiw Umvmitv School or nd PjbJ-c HeJhh Ttoprfjl Medicine. WO TuUne Aw. O*learn. IA 70111. 384 residents of towns where the petroleum industry was the major employer.1 Other industries with occupations also associated with lung cancer included shipbuilding, machinery manufacturing, petrochemicals, transporta tion, and construction. In cancer mortality studies of employees of oil refineries, no significant associations of occupation with cancer risk have been identified, despite the availability of work histories. Since Louisiana has many oil refineries, a population- based mortality study allows the identification of a greater number of exposed individuals, across individual factories, but with occupations in common. Theriault and Goulet/ in comparing the cancer mortality of oil refinery workers to the Quebec population, found lung cancer less frequent than expected, an unusual excess of brain cancer in young people, and an elevated SMR for cancers of the digestive system plus an elevated number of deaths due to non-neoplastic diseases of the digestive system. Hanis. Stavraky and Fowler,1 however, found more than a three-fold risk of esophageal and stomach cancer, and twice the risk of lung cancer in exposed compared to nonexposed petroleum workers. In the exposed group, in creasing risks for both disease categories occurred with increasing duration of employment An examination of deaths with a recorded history of employment in the petroleum industry is presented to identify associations with the elevated risk from cancer of the lung-in this industry and occupation. Methods Parishes included in the Lousiana study were: Acadia. Allen, Ascension, Calcasieu. Concordia. Iberville. Jackson. Morehouse, Plaquemines. St. Bernard, St. Charles. St James, St John, St. Martin. Tensas. Terrebonne. Washing ton. Webster, and West Feliciana. These parishes vs ere selected because of elevated lung cancer death rates among white males for the years 1930 to 1969 or because of their concentration of petroleum, chemical, and paper manufacturing industries, or both. Lung cancer deaths Table 1. -- Reporting ei Occupation en Death Certificates (or Cancer Deaths and ----------------- *T-T1 Non-Cancer Deaths Employed In the Petroleum Indusiry for White end Black Males. White males Black males Total Potioieim Indestry (Codes MS, 377) Cases Occupation Re ported Not Reported 161 12 (6*4) 70 200 Ceetrets Occupation Reported Met Reported 153 12 (7%) 50 170 were identified by ICOA number from a data tape of all Louisiana deaths from 1960 to 1975 compiled by the Loui siana Division of Vital Statistics. Notation of cancer ap pearing anywhere on the death certificate is coded, and retrievable as the underlying cause. An equal number of non-cancer deaths were selected as controls and matched to the lung cancer cases by sex, race, year of death, parish of residence at death, and age of death ( two years). The first control, meeting the criteria, in se quence on the tape was selected. The death certificate for each subject was abstracted to obtain information on other contributing causes of death, occupation and industry. Occupation and industry were coded by the three-digit classifications of the U.5. Census Bureau.* Analysis for relative risk with 95% con fidence intervals is by the modification of RothmanBoice.' Within the petroleum industry, occupational categories were compared for cases and controls, and for age at death. Results The 1970 census for 19 parishes included 346.041 males. A total of 2.803 cases of lung cancer among males were ascertained for the study period. For this group, 2,161 were white males and 642 were black. Information was sufficient to classify the decedent's usual industry of employment into one of 22 industrial categories for 87% of white males and 60% of black males. There was no significant difference between the percentage of controls by reported industry for either racial group. Occupation could be coded from death cer tificate information for 92% white males and 92% of black males. Occupation was reported for a greater number of cases than controls for both whites and blacks. However, for the petroleum industry, there was no signifi cant difference in reporting of occupation for cases and controls (Table 1). Since there were only 12 black males in the study population who were employed in the petroleum industry, the following results are presented for black and white males combined. A total of 200 cases were employed in the petroleumC/) industry, 125 in mining and 75 in refining There were 170"H controls employed in the petroleum industry, 112 in mirvQ ing and 58 in refining. For purposes 3f analysis the mining" and refining data were combined under the heading of~ "petroleum industry" as many occupations were similar--I in both categories, and it was assumed that exposure toP" hydrocarbons was the major factor in both industrial subOh categories. The distribution of occupations for males employed in the petroleum industry is given in Table 2. The odds ratio for cancer for employment in the petroleum industry, 1.19, does not include 1.0 in the 95% confidence limits. Moreover, several occupational categories within the petroleum industry are noted to demonstrate greater risks for lung cancer mortality. For workers and oilfield workers, welders, operators, boilermakers, and painters, the odds ratio is 2.33. This increase is mainly in those over age 60 at death, as noted in Table 3, where the data are presented for two age-at-death groups. < 60 years and > 60 years. A larger proportion of controls than cases died at less than 60 years of age, particularly in the managerial categories. No increase in chronic obstructive pulmonary disease was reported in deceased petroleum workers compared to non-petroleum workers. This disease was recorded for Table 2. -- Rick ol Lung Cancer Mortality by Specific Occupations in the Petroleum industry. Petroleum Worker. oiltielS worker Welder Skilled, pumping and refining Operators Boilermakers Painters Subtotal Other Cooks, chefs, siewards Managerial, professional, sales Skilled, construction Transportion. guards Not available Total ' ' Mining and Refining Cases Controls 26 15 82 28 13 40 30 69 30 20 22 42 37 48 48 44 10 11 12 13 200 170 ROwnn 9SK Confidence Inlervtl 2.33 (1.42-3.Eo) 1.19 (1.05-1.35) Journal of Occupational Medicine/Vol. 22, No. 6/June 1980 385 29*1/.`l I 01S Table 3. -- Distribution of Cancer and Non-Lancer Deaths by Occupation and Ape at Death ________cl tr.ales Employed in the Petrcleum Industry (Industrial Codes 049. 377).________ Occupation <60 Cases Controls Age at Death (in Years) i.60 OR Cites Controls OR Workers. otifieio workers 6 S 1.36 (0.24.22.48) 20 10 1.66 (0.91. 3 8 ) Welder 5 1 1.89 <0.48, 7.37) 3 1 0.93 (0.25. 346) Cooks, chefs, stewards Managerial, professional sales Skilled construction Skilled pumping and refining Transporuicn, guards Total petroleum workers Total lung cancer 12 30 17 34 20 14 23 18 25 26 29 26 1.32(0.92. 190) 25 10 2.43 ft.01. 5.66) 2 8 0.28 (0.23. 0.35) 8 3 2.48 (0.49,12.50) 46 67 87 95 113 71 1005 1137 1798 1666 Table 4. -- Chronic Obstructive Pulmonary Disease (COPD) in Petroleum Mining and Refining Workers and Matched Controls. Occupational Group Petroleum-related Non-petroleum-related Relative odds (95% conSdence interval) Cause el Death COPD Non-CGPD No. % No. 11 6.5 13 8.0 159 155 0.82 0.58-M7 ten and fibrosis in one of the 170 petroleum workers dy ing from causes other than cancer [6.5%). Twelve cases of chronic obstructive lung disease and one case of fibrosis were recorded for the 168 non-cancer deaths in subjects not working in the petroleum industry matched to the petroleum workers dying from lung cancer (8%) (Table 4). The location of the lung cancer is indicated in 27% of the cases. Particulate matter is preferentially expected in the right lung. Among the 200 lung cancer deaths in petroleum workers, the right lung was indicated for 32 cases and the left lung was indicated for 22 cases. The dif ference was noted largely in those categorized to petroleum mining (Table 5). However, interpretation is dif ficult due to the large number with unspecified location. The mean age of death of the cases was 60.5 years. A large number (58%) of petroleum worker controls died at 59 years of age or younger, and more than half of these deaths were due to trauma or myocardial infarction. The distribution of cases and controls by their ages in 1945 (around the time of the offshore oil boom) is given in Table 6. The controls were generally younger than the cases; 59% of controls were 34 or younger compared to 35% of cases. This again reflects premature death from other causes in the controls. A chi-square analysis of the distribution of cases and controls born m Louisiana and out-of-state reveals an excess of cases in the high-risk petroleum mining occupations (welders and oilfield workers) born out-of-state (Table 7)..JRO 227 (95% con fidence interval: 0.58, 8.83)]. This association did not ap pear for refining workers. Discussion The elevated risk of lung cancer mortality' among petroleum workers in these selected Louisiana parishes is attributable largely to several occupational groups: welders, oilfield workers, operators, boilermakers, and painters. An excess in cases among boilermakers and painters was observed for the entire data set (across all in dustry categories) as well as for the petroleum industry alone. However, the excess among cases who were welders, which is apparent in the petroleum industry, was not observed across all industry categories. It is possible that different factors are operating in the etiology of lung cancer for these three occupational groups. The operator and worker categories are specific to the petroleum in dustry. The association between chronic obstructive pulmo nary disease (COPD) and petroleum-related occupations was also examined. This can be considered a measure of the relative amount of cigarette smoking by workers in this industry. To measure this association, the matched controls (non-cancer) of the cases working in a non petroleum-related industry were compared with all con trols who were reported to have worked in the industry. The resulting percentage of COPD deaths for petroleumrelated occupations, and non-petroleum-related occupa tions was 6.5 and 8.0. respectively (Table 4). This indicates the accuracy and similarity of diagnostic reporting in the Table 5. -- Laterality o) Lung Cancer in Petroleum Mining and Relining Workers and Matched Non-Petroleum Workers. Oceupattons Peuoreum mining Petroleum relinmg Non- petroleum-related Right No. laterality Specified Lett % Ke. 20 63 12 12 55 10 25 54 21 % 38 45 46 Unspecified No. 93 53 124 286 Lung Cancer and the Petroleum Industry/Gottlieb Table 6. - Petroleum Industry Cohort Age in 1945. ' Ago 0-19 20-34 35 + Age in 1945 not eniteWe Total Ceus NO. % 10 5 59 30 130 ES Centtsls No. % 21 12 79 47 70 41 1 0 200 170 two groups, and the security that all cases of lung cancer were likely identified and reported, or one might have ex pected to have identified more non-specific lung disease. The similarity in the proportion of petroleum-related controls with COPD compared to that from non petroleum-related controls suggests that cigarette smok ing does not differ in the two groups, and that the higher rate of lung cancer associated with the petroleum in dustry is not likely to be related to cigarette smoking alone. The reported rate of 6% COPD is lower than that found in a similar comparison of workers related to the shipbuilding industry.* The laterality of lung cancer was also investigated among occupational groups, since particulate substances lend to be trapped preferentially in the right lung due to the configuration of the right bronchus. To measure this association, the cases not working in the petroleum in dustry who were matched to the controls working in these occupations were compared with cases working in petroleum industry occupations. The resulting proportion of right lung cancer in the petroleum industry occupa tions was not increased significantly above that of lung cancer cases in the non-petroleum industry occupations. However, information on laterality was specified in only 27% of the cases. The mean age at death for cases was several years older than that for controls. Similarly, the median age of death for cases*' was substantially greater than that for controls1' considering that cases and controls in the original data set were age-matched two years. (For all industries combined the median age at death was 62.8 for cases and 61.4 for controls.) These findings may be ex plained by the fact that many of the deaths for controls in the petroleum industry were accidental and occurred in very young men, who might have become cases, if they had survived. This indicates that risk of lung cancer for this group may be underestimated in this study. A revised odds ratio was estimated after excluding accidental.' traumatic or violent deaths, chronic obstructive / pulmonary disease, and those whose age was lesvsJian-40 years at time of death from the control group. It was T29 D-11 to 1.50) for the petroleum industry. For the high-risk occupation group in this industry, the corrected odds ratio is 3.03 [1.45 to 634). a further elevation. This rein forces the significance of the risk of these occupations. The ages of cases and controls relative to a reference date. 1945, were examined to see whether differences in the ages of the two cohorts might result in differing tem poral exposures for the two groups. More controls were under 35 years of age and more cases were 35 or over. The older cases probably reflect a longer period of ex posure to the industry. The observation that petroleum mining cases were more frequently bom out-of-state than controls is prob-^ ably related to migration patterns involved with employ-__ | ment in the industry. Those mining workers who relocate,-^ frequently at new mining sites are probably long-term_ employees in the industry and therefore long-term ex-_- posure, high-risk employees. The odds ratio for these__j workers is elevated, but due to thermal! sample size, wasXT' not significant. Refining workers, on the other hand, were^n not more likely to be bom out-of-state. r .*> The observation that lung cancer was recorded more frequently for the right lung than the left lung particularly among the miners is compatible with the idea that air borne particulate substances contribute to the etiology of lung cancer. Whether or not differences in cigarette smoking pat terns associated with social class and occupation is the common denominator in the differences between cancer deaths and non-cancer deaths cannot be directly evaluated from this study. Indirectly, however, the high median age at death in cases and the similarity in percent reported COPD for cancer and non-cancer deaths would tend to require more than differences in tobacco smoking patterns as an explanation. Conclusion National correlation studies have suggested an associa tion between employment in the petroleum industry and lung cancer mortality. This death certificate study has demonstrated an elevated risk of lung cancer mortality (odds ratio -- 1.19) for workers employed in the petroleum industry in 19 selected Louisiana parishes. In dividuals with several specific occupations within the petroleum industry (welders, operators, boilermakers, painters, and oilfield workers) appear to be a greatest risk of lung cancer mortality. RO 233 0-42 to 3.85) The petroleum industry has been an important industry in Table 7. -- Distribution of Petroleum Workers by State ol Birth -- Louisiana and Other. Petroleum Industry Occupational Risk Cases Birthplace Louisiana Controls Outside Utrsiene Cases Odds Ratio Controls Mining Odds ratio Reiining All 82 61 44 30 High 24 16 17 5 2.27 [(0.58.8 83)) Ail 51 39 23 In High 20 8 10 3 Journal of Occupational Medicine/Vol. 22, No. 6/June 1980 387 *19*1/.*1101s RitTisiaua for the past 50 years and may explain part of the excess in reported lung cancer mortality rates. The use of death certificates representing a large population base whose employees are at an increased risk of cancer is an alternative to studying the work histories of employees of individual companies. The large popula tion base which enables surveillance of many death cer tificates allows the identification of workers from several companies and a broader identification of those exposed. Only for those individuals with a substantial period of employment is the industry likely to be recorded on the death certificate. This would tend to eliminate casual employees and allow the inclusion into the study of employees who achieved a long period of employment This is true for cancer and non-cancer deaths. Specific job titles were recorded with a high frequency on the death certificate. Although the inherent weakness of death cer tificate information is well-known, it appears to be a valid screening tool for detecting industrial and occupational associations with lung cancer. Ihil wo V was supported SO Contract No MOSS and bv Department erf Health an<j Hufiwn ftevources lou*'a%u References 1 Majon T| and McKay FW: U S Cancer Monalm b. Ceunrv 1950-1969. Department of Health. Education anc Wehare P-fcl No (KtH) 74-615 Washington. DC 19782. Blot W| and Fraumem ]F Jr Geographic pattern! ot lung cancer Industrial correlations. Am I Ipidemiol 103:539-550. 19*6. 3. Gottlieb MS, Pickle L. Blot WJ. and Fraumem JF Jr ; Lung cancer in Louisiana: Death certificate analysis. / Naif Cancer Inst 43:1131-1137, 1979 4. Theriault C and Coulet L: A mortality study of oil refinery workers. I Oczup Med 21:367-370. 1979. 5. Hanis NM. Stavraky KM. and Fowler |L: Cancer mortality in oit refinery workers. / Occup Med 21:167-183.1979. 6. U S. Census of Population. Alphabetical Index of Industries and Occupations. Washington. D.C. U S. Government Printing Office. 1971. 7 Rothman-Boice Moditicationso; Mietunen OS. tstimibiliN and estimation m case reieience studies. Am / Epidemiol 103.225-235. 1976. 8 Gottlieb MS and Sledman RB. Lung Cancer in Shipbuilding and Related Industries in Louisiana. South Med I 711099-1101.19*9. Y Aesculapian Authority Paterson defined medical authority as consisting of three types: sapiential, moral, and charismatic, combined in a particular manner. He called this "Aesculapian authority." By sapiential authority, Paterson means the right to be heard that is derived from knowledge or expertness. Such authority resides in the person and not in any position that he may occupy. A person with this kind of authority may advise, inform, instruct. and direct, but not order. The second ingredient of Aesculapian authority is moral authority, the right to con trol and direct that is derived from the rightness and goodness according to the ethos of the enterprise. The moral authority of doctors, which is expressed in the Hippocratic oath, stems from their doing what is expected of them as doctors and their concern with the good of the patient The third ingredient in Aesculapian authority is charismatic authority, the right to control and direct that is derived from Cod-given grace. This element in Aesculapian authority reflects the original unity of religion and medicine that still exists in many parts of the wofld. -- From 'Cod and th Doctor* bv Humphrey Osmond, in The Efigbnd lounj of Medicine. March 6.19G0 388 Lung Cancer 2nd the Petroleum Industry/Gottlieb