Document ByBppVEnRL23OVOEGqO6a2438

PLAINTIFF'S EXHIBIT MBL-102 Update of a Mortality Study of Workers in Petroleum Refineries Samuel D. Kaplan. MD. SM Hyg A large cohort of petroleum refinery workers with loaf duration of employment, loaf lata007, aad relatively young af at biro bad its vital statu* updated tbroufb Ooe 31. 1990. Tbo standardized mortality ratio (SMS) for all cauaos wan 78. Each aooaoopiaatic eauao bad aa SMS bolow 100. laclad(of SMBo of S3 for emphysema and for all lilsooioo of tho genitourinary system aad of 73 for chronic aophritla. Tho SMB for all eancara wan 87. SMBs for 1pacific neoplasms (acludad digosUvo system. 90; loaf, 98; kidney, S8; brain. 89: leukemia. 101; ouitlpla myeloma. 133; nnapocifiod lymphoma, 113; polycythemia vera (four daathn). 488; myelofibrosis (throo daathn). 201; aad banlfa aad nanpaciflod brain neo plasms. 108. Thor* woro ala* daathn from mesothelioma; all sina employees bad atom than 30 yoar* of employment. with an 3MS of 241. In January 1972, Taborshaw-Cooper Associates, Incor porated (TCA), under tbe iponsorship of the Amer ican Petroleum Institute (API), befaa a retrospective cohort mortality study of workers in a sample of United Staten petroleum refineries. Thn original report wan widely distributed by API. Because tha cohort Is not only large but is also one with loaf latency and duratioa of expoeure with early age at hire, this study Is liksly to be a particularly valuable addition to the literature. Therefore, the cohort's vital status has beea updated, and a detailed analysis was carried out. Methods The vital status of the cohort was updated through From SRI International. 303 Revenewood Av*. Moalo Park. CA 94023 (Dr Kaplan. Senior Epidemiologist). Addrvea correspondence to: Department of Hoaith Service*. Adult Haaith Section. 714 P St. Room 430. Secrsmeoto. Ca 93614 (Dr Keplu). ooee-173/M/2ao7i 1 *na oo/o CaernOt *- >7 Asartcaa Oecepeltaeai Medical AaaoeiaUa* S 14 Dee 31, 1980. Data were analyzed by a modified life table program adapted from the one described by Monson* and a computer program from tbe National Institute for Occupational Safety aad Health (NIOSH).* For study oohort members lost to follow-up, personyears were '*m..ioEon until time of loea The stand ardized mortality ratio (SMB) for multipla myeloma (which wan not available la either tha Monson or NIOSH program) was calculated from data provided by Dr Gary Marsh. SMBs for meeotheliotaa were calculated from expected numbers generated by ago- specific rates for white males for tha years 1969 to 1971 from tha Third National Cancer Survey (M. Myers, personal communi cation. 1984) and annual aga-spacifle rates for males from tha SEER Program for tha years 1973 to 1978.1 In all tha abovs analyses, black males wars analyzed in comparison with the mortality experience of nonwhite mains. All othor study oohort members had their mor tality sxperianoa compared with that of whits males, and tha results were combined. Baaed on the distribution by plant of thoae cohort members with known race, we estimate that 99.3% of thoae whoee mortality experience was compared with that of white tnalae were white. TCA collected 90,169 Individual records for this study. Of these. 178 (0.9%) were eliminated from the analysis for the following reasons: 1. Seventy-nine workers from one refinery were elim inated because tha records for men who had quit or were discharged were kept for only 1 year unless tha employee had died on tha Job. 3. Forty-one had worked leas than 1 year between 1962 and 1971. 3. One was female. 4. Fifty-seven were duplicate records, including tea of employees who qualified in mars than oae participat ing refinery. With the elimination of these records, tha total cohort size for follow-up and analysis became 19,991. Petroleum Refinery Woriters/Kaplan 448474 Results Through Dec 31. i960, 13,936 of the cohort were known to bo alive. 3.349 worn known to bo deceased, and 707 (3.5%) worn of unknown vital status. Of tho cohort. 31.3% wora hirod between 1940 and 1954, aod 17.2% wore hired boforo 1940. Aa of this update, more than two-thirds of tho cohort had boon employed In participating refineries for more than 20 years. Tho mean age at hire for the cohort was 26.2. Tho Table shows the causes of death from tho Monson program analysis, tho observed and expected deaths, tho ShfRs, and the 95% confidence intervals. Informa tion on cause of death was not available for 80 deaths (2.4% of the total), and SMRa were not adjusted for these unknown deaths. If they were, the SMRs would bo between 2% and 3% higher for each cause. For all causes, tho number of observed deaths is less than four-fifths of ths numbsr expected. The SMRs are even lower for many specific causes, including cancer of the kidney, emphysema, and diseases of the genitouri nary system. In addition, because of the large numbers involved, some SMRs are significantly low even though they are close to 100, including all malignant neoplasms and lung cancer. Of the 30 deaths from cancer of other lymphatic tissue, 16 were from multiple myeloma. The number of expected deaths was 12.99, giving an SMR of 123 for this cause and a 95% confidence interval of 70 to 200. Of the 14 other deaths from this cause, four were from polycythemia vara. and the other ten were from other neoplasms of lymphoid tissue (Code 208, International Classification of Diseases. Eighth Revision). There was one benign brain neoplasm, with 1-87 expected, for an SMR of 35. For the unspecified brain neoplasms, there were seven deaths, with 3.51 expected, for an SMR of 127 and a 95% confidence interval of 51 to 262. Nine mesotheliomas were found, all with more than 20 years duration and latency. In addition, we know of one other employee who was alive at the end of the study but who died In 1982 and whoee death certificate listed mesothelioma as the cause of death. That employee also had a long duration of employment and a long interval between start of employment and death. Of the nine deaths from mesothelioma by the closing date of the study, six clearly were indicated to be pleural meso thelioma on the death certificate. For the three others, it is not clear whether they were pleural or peritoneal. Of the occupations recorded on the death certificates, three were machinists, two were welders, and one each was an insulator, maintenance supervisor, "Utility Di vision," and pipe fitter. One death certificate listed asbestosis as a secondary cause. The expected number of mesotheliomas in the study cohort was 3.73 for an SMR of 241, with a 95% confidence interval of 110 to 456. Discussion The cohort is at lower risk for all deaths snd for many TA0LE Observes md Ejected Oeams am Summed uwtwty Rato. iSmr,, i *dr Maw Enveyeae of Pevcawjn rteHnenaa Throve Oec Ji. AJ infective and parasitic cisease Tuberculosa Malignant neoplasms Cancer of buccal cavity and pnarynx Cancer of cigesave organs and pentoneum Cancer of esopnagus Cancer of stomacn Cancer of Urge ntesone Cancer of rectum Cancer of Ww Cancer of pancreas Cancer of respratory sys tem Cancer of larynx Cancer of trig Cancer of bane Cancer of skin Cancer of prostata Cancer of tasks Cancer of biaddar Cancer of ludney Cancer of tyw Cancer of brain and other parts of nervous sys tem . Cancer of thyroid Neoplasms of tymphaoc and hematopONtic tis sue Lymphosarcoma and taJun cel sarcoma Hodgkin's dsease Loitema Cancer of other lym phatic tissue Cancer of other and un specified sites Observed Expected S% OeeOie Oeedte SMfl Confidence internal 3.349 4.280 35 78 76-31 IS 35.07 39 22-65 3 18.02 17 3-49 793 915 59 87 81-93 ta 30 37 59 35-94 223 246.91 90 79-103 20 2651 75 46-117 <3 44 06 98 71-131 73 77.76 94 74-118 27 25.78 105 69-152 13 16.04 81 43-139 44 5125 86 62-115 278 335.19 83 73-33 a 14.63 55 24-108 270 317.72 35 75-96 4 3.35 <04 26-266 15 14.05 107 60-176 60 58.81 102 78-131 4 4 08 98 26-251 15 25.46 59 33-97 15 22.09 68 35-112 0 0.72 0 0-509 22 24.63 39 56-135 2 1 64 122 14-440 88 32.50 107 85-131 16 17.78 90 51-146 7 8.73 80 32-165 32 31.68 101 69-143 30 22.82 131 39-188 50 65 30 77 57-101 Senign neoplasms Endocme. nutritional, and metaooic dsaases Oiabetes meakus 14 10.54 133 79-223 45 78.19 59 43-79 37 64 58 57 40-79 Diseases of Mood and blood- 8 8.93 90 39-177 tonmg organs Mental dberders 3 27.06 11 2-32 Diseases of nervous system and sense organs 19 34.20 56 33-87 Diseases of the cmiatory system Chrorec rheumatic heart disease iscnemc heart dsease Cerebrovascular disease Diseases of respratory sys tem Pneumoraa Emphysema Asthma Diseases of digestive system Peptic ulcer Grmosa of Svw 1.727 2.135 31 79 75-83 26 34.79 75 49-110 1 267 1.575.45 80 76-85 185 287 60 65 56-75 167 262.91 64 54-74 45 88.94 51 37-68 46 72.91 63 46-84 5 6.32 79 26-185 121 211 83 57 47-88 14 24.71 57 31-95 65 119 90 54 42-69 Journal of Occupational Medicine/Volume 23 No. 7/July 1988 515 448475 t < TA 3LE--Continued 202.2) in the Eighth Revision (and for Codes 202 aad Chinas Observed Deems Expected Deems SAW Confidence imams* 203 in the Seventh Revision, which equate to Code 202 in the Eighth Revision) there were one observed death and 1.03 expected deaths, for an SMR of 97. For poly, Diseases of genitourinary system Chronic nephritis Diseases of the skm and sut. cutaneous tissue Diseases of me muscotoskeletal system and con nective tissue 33 52.52 63 43-88 13 1793 73 39-124 3 3.15 95 19-278 3 7 54 40 8-116 cythemia vera there were four observed deaths and 0.88 expected deaths from this cause for an SMR of 433, and 95% confidence interval of 120 to 1,164. The 88 ooserved deaths from neoplasms of lymphatic and hematopoietic tissue are three more than the ob served deaths from each of the components (lymphosar coma and reticulum cell sarcoma. Hodgkin's disease, Symptoms and d-defined conation* 54 63.78 85 64-110 leukemia, and cancer of other lymphatic tissue). The difference is the three deaths from myelofibrosis. The Aeodents. poisonings, and vtotence Ai accidents Motor vehicle accidents Suede AI other causes Unknown 260 375 52 69 61-78 164 237.98 66 59-80 77 106.32 72 57-91 57 81.95 70 53-90 12 19.14 63 32-110 90 expected number of deaths was 1.49. which gives an SMR for that cause of 201 with a 95% confidence interval of 41 to 588. It would appear that the concern about neoplasms of the lymphatic and hematopoietic system arisee primar ily from polycythemia vera, and perhapa from myelofi brosis. There is some question as to whether these are No. of workers: 19.991. No of person-years: 297,591-9. really neoplasms, and the Seventh Revision place* them both under diseases of the blood and blood-forming specific causes of death. The 96.3% follow-up rate com pares favorably with follow-up rates usually achieved in organs. However, the numbers are too small for poly cythemia vera and myelofibrosis to warrant further retrospective cohort mortality studies and fewer than investigation at present. 3.3% of the death certificates were not obtained. Part of the explanation for these low SMBs appears to be the well-known healthy worker effect. Anecdotally, people familiar with petroleum refineries have noted that heavy smokers are not likely to work there because Acknowledgments of the smoking restrictions for safety purposes that are This work u eupported in part by s coatrmct from the Americas Imposed at work. In combination with the usual healthy worker effect, the anecdotally noted low smoking prev alence among refinery workers may explain these find ings. Petroleum Institute. The contributions ef Or William R. GaiTey and Dr Neill K. Wearer to the origiaal study design. the advice aad technical assistance of numerous employees in the petroleum Industry, aad tha unique con tributions to the study at Dr (rviag P. Tabanhav are gratefully The accuracy of the diagnoses and the appropriate acknowledged. ness of the comparison data could have affected the results for mesothelioma. The probable net result Is that tha expected number of deaths from this cause is slightly overestimated, which means that tha SMR La If any References thing. underestimated. 1. Monaoa RfL AajJytu of r*l4tiv* urvtvfti ud proportion*] mor* There wars nine observed deaths and 7.93 expected deaths from ICO Cods 202.2 in the Eighth Revision (other primary malignant neoplasms of lymphoid tissue) for an SMR of ' 13 with a 93% confidence interval of 52 to 216. For t.he remainder of Code 202 (other than UUity. Coasput 9iotnod Rm 1974;7:325-339. 9- WtrirtiUr RJ. 8**umoat JJ. Hnry JA. rt ti: A modified life Ubte kftalyiu system for cohort itudleo. / Occxtp Mod 1963:25:115124. 3. Walker Am. Ltufhlia JZ. FriedUnder ETC tt Projections of sebeoUM-related diM*e. 1983-2009. j'Occvp Mod 1963.25.409-425. sis Petroleum Refinery Workers/Kaplan 448476 TABLE A7 Percent Disagreement far Sub-Cohort Classification Sub-Cohorts Exposed Not Exposed Not at Beaunont, or Location Code Missing Intra (341 pairs) 2.IX 2.32 3.85 Inter (1 82 pair s) 3.3X 4.4J 4.9X 448472 yv> Y 448473 '..IQ .