Document 82g4qR1grJGz1B3ddoGVmDNLy

AR226-1308-2 s-c* ? M0 3 Final Report Study No.: Study Title: S-088 A Case-Control Study of Leukemia at the Washington Works Site Principal Investigator: m . Elizabeth Karns, m .p.h . E10p0id7eMmaiorkloegtiSsttreet Du Pont Co. Co-Investigator: William E. Faycrweather, Ph.D., M.P.H. Manager, Epidemiology Section 1007 Market Street Du Pont Co. Study Sponsor: Polymer Products SSttuuddyy ICnoitmiaptiloentioDnaDtea:te: 0152//0311//8919 Page 1 of 13 EID1S1953 \ft RL002162 Page 2 of 13 Final Report - Study No.: S-088 Table of Contents Page No. I. Review and Approval.................--.-- ...................................................................3 oo PCroi-nIcnipveasltIingvaetostri(gsa)t--or.................................................................................................................................................................................33 IT. Introduction................................................................................................................ 4 o Abstract...............................................................................................................4 o Purpose............................ 4 o Rationale.............................................................................................................4 HI. Research Methods.......................................................................................................5 o Study Design.......................................................................................................5 o Exposure Assessment........................................................................................ 5 o Data Analysis..................... 5 IV. Results................................................................... 6-11 V. Discussion.................... ...........................................................................................12 VI. Conclusion............................................................................................................... 12 VH Acknowledgements.............. .................................................................. - .............. 13 VU. References........ ..................................................................-................................... 13 EID1S1M4 If RL002163 Page 3 of 13 Final Report - Study No.: S-088 REVIEW AND APPROVAL I attest that, to the best of my knowledge, this report is scientifically sound, describes the study and procedures used, reflects the data, and that the study was conducted in compliance with the appropriate regulatory requirements, Good Epidemiology Practices, the research protocol, and any amendments thereof. Submitted: _ . T S. ' .Principal Investigator /f ( O U ' t\ .{ -j _ _ _ _ _ _ _____________ (Add additional siyiature lines as needed forco-investigatots) Date: C l-n Date: t v" 3 \ , i V[\ j ^ I } )j | Approved: Section Manager 'I Date: Oui<.v, i t . I f ! j EID1S19S5 cI R L 0 0 2 1 64 Page 4 of 13 ' Final Repon - Study No.: S-0S3 INTRODUCTION ABSTRACT A case-control study (Study No. S-088) was initiated to determine whether the risk of developing leukemia is related to workplace exposures at the Washington Works plant. The study found no association between employment at Washington Works in any particular job or area grouping and the risk of developing leukemia. A statistically significant excess of leukemia was observed in those who worked as custodians or chemists prior to working at the Washington Works. These findings are based on few cases so that considerable variability was observed in the estimated odds ratios. PURPOSE The purpose of the study was to determine whether the risk of developing leukemia is related to workplace exposures at the Washington Works plant. RATIONALE The study was started in response to employee concerns regarding an apparent cluster of leukemia at the Parkersburg, West Virginia site. Prior to the present study, a cluster investigation had been carried out by Walrath and Burke (1989). Their investigation was initiated after an elevated standardized incidence ratio (SIR = 1.47 with 6 observed cases versus 4.08 expected cases) was detected during routine cancer surveillance. In addition, employee concerns were also raised regarding new cases of leukemia. No commonality of exposure was found when work histories were reviewed. This study was undertaken to determine whether the leukemia cluster was related to the work environm ent. EID151956 I7 RL002165 Page 5 of 13 Final Report - Study No.: S-08S RESEARCH METHODS STUDY DESIGN A case-control study design was used. Cases were identified through the Du Pont Cancer Registry. The Cancer Registry is limited to cancer cases diagnosed among active employees after 1955 (Pell.et al., 1978 and O'Berg et ai,, 1987). Cases were restricted to those diagnosed with leukemia while working at the Washington Works site during the period 1956 to 1989. Four controls were selected for each case from the site's employee roster that corresponded to the case's diagnosis year. Each control was matched on gender, year of birth (within three years), and payroll status at the time of diagnosis. EXPOSURE ASSESSMENT Occupational exposures were determined through job title and work area matrices. Duration of employment was calculated in each job title and area combination. Although an attempt was made to utilize the chemical-specific information, little data were available for the majority of subjects. DATA ANALYSIS Conditional logistic regression models were used to assess possible relationships between exposure and the risk of developing leukemia (Rothman, 1986). The Egret software package was used to derive the estimates of relative risk (SERC, 1990). The odds ratio was calculated as an index of the strength of the relationship. Ninety percent confidence intervals were calculated for each odds ratio as an index of variability. Any odds ratio whose lower 90% confidence ratio exceeded 1.0 was defined as statistically significant. - EID1S1M9 IS R L 0 0 21 * Page 6 of 13 Final Report - Study Mo.: S-088 RESULTS The nine cases were distributed over a range of leukemia types. Five cases were acute non-lymphocytic leukemias, three were chronic myelocytic leukemias, and one was an acute lymphocytic leukemia. These are distinct leukemia types which are not thought to come from a common exposure (Linet, 1985). A comparison of demographic and work experience characteristics between cases and controls is in Table 1. All subjects were male. The matching of controls to cases was successful: no significant differences were observed between cases and controls regarding age or payclass. Table 2 compares the work experience between cases and controls. Table 1 Washington Works Leukemia Study: Characteristics of Cases and Controls Variable Name Birth year 1902-33 1934-36 1937-42 1943-54 Pay class: Salary Wage Cases n=9 % of cases 11 33 33 22 33 66 Controls n=36 % of controls 22 25 25 28 33 66 E1D1S19S R L002167 Final Report - StudyPaNgoe.:7So-f08183 Table 2 Washington Works Leukemia Study: Employment Characteristics of Cases and Controls Variable Name Cases n=9 mean (SD) Controls n=36 mean (SD) Age start at WW 29(10.3) 27 (8.4) Total years at WW 15 (8.8) 18(7.6) Year started at WW Years in prior jobs 65 (4.8) 8 (8.8) 63 (8.8) 7 (6.1) Occupations prior to coming to Washington Works were categorized into nine groups (Table 3). Custodians and engineers had statistically significantly elevated odds ratios (OR) of 8.0 (90% Cl: 1.1, 60) and 7.9 (90% Cl: 1.0,76), respectively. Chemists, construction, office, and other workers had elevated but not statistically significant odds ratios. Industry experience prior to Washington Works was categorized into nine groups (Table 4). Those who worked in the service or retail sectors had an elevated odds ratio of 4.1 (90% Cl:0.9, 17.5); however, this was not statistically significant EI01S19M RL002168 Page 8 of 13 ' Final Report - Study No.: S-088 Prior Occupation Chemist/ Lab Construction Table 3 Washington Works Leukemia Study: Matched Odds Ratio and 90% Confidence Interval on Prior Occupation Group Number of cases Number of controls Matched Odds Ratio 90% C.L on the Odds Ratio I 2 4 2.7 0.5, 15.4 4 13 1.4 0.4, 4.8 Custodian 2 1 8.0 1.1,60.0 Engineer 3 2 7.9 1.0, 76.0 Automobile 0 12 - - Mfg. 5 20 1.0 0.3, 3.7 Military 0 3 -- Office 4 12 1.7 0.5, 6.5 Other 5 11 2.9 0.8, 10.4 EID151960 R L .002169 Final Report - StudyPaNgoe.:9So-f08183 Table 4: Washington Works Leukemia Study: Matched Odds Ratio and 90% Confidence Interval on Prior Industry Group Prior Industry Number of cases Chemical, Metal Construction Manufacturing Military Other Du Pont Petroleum Service, Retail Transportation Other 3 1 2 1 2 1 6 1 1 Number of controls 14 4 9 12 9 12 14 7 3 Matched Odds Ratio 0.8 90% C.L on the Odds Ratio 0.2, 2 8 1.0 0.9 0.3 * 0.7 0.2,6.3 0.2.3.5 <.01,1.6 <.01. 6.1 0.3 <.01,1.6 4.1 0.9,17.5 0.5 <.01, 3.4 1.5 0.1, 15.4 EID151961 RL002170 PFiangael 1R0eopfo1r3t *Study No.: S-088 Washington Works experience was analyzed by job title and by work area. The paucity of data precluded further combinations or stratifications. Job titles were categorized into eight groups (Table 5). Elevated odds ratios were observed in chemists (O R =1.9), engineers (O R =5.1), office workers (O R =2.0), and custodial workers (OR=4.0). All confidence intervals on the odds ratios included 1.0 and were not statistically significant. Work areas were grouped into nine categories (Table 6). The work area analysis was done by determining if the subject had ever worked in the area or never worked in the area. Maintenance, laboratories, and office areas had elevated odds ratios greater than 1.5. However, all confidence intervals included 1.0. Table 5 Washington Works Leukemia Study: Matched Odds Ratio and 90% Confidence Interval on Job Title Group Job Title Chemist Foreman Engineer Handler Mechanic Office Operator Custodial Number of cases 3 0 2 4 6 1 6 1 Number of controls 3 9 3 20 29 2 28 1 Matched Odds Ratio 1.9 - 5.1 0.4 0.5 2.0 0.6 4.0 90% C.L on the Odds Ratio 0.4, 8.1 0.6, 42.6 < .0 1 ,2 .2 0.1,2.8 0.3, 15.0 0.1, a 2 0.4, 41.0 E iO lfttW t RL002171 3 Area of work Butacite Delrin Maintenance Mechanical Nylon/ Acrylics Final Report - StudPyagNeot.:lSo-0f 8183 Table 6 Washington Works Leukemia Study: Matched Odds Ratio and 90% Confidence Interval on Area of Work Group Number of cases 4 2 a 1 2 Number of controls 20 8 30 12 19 Matched Odds Ratio 90% C.L on the Odds Ratio 0.6 0.2, 2.2 1.0 0.2,4.3 2.1 0.2, 21.5 0.3 4> 0.2 < .0 1 ,1 .6 < .0 1 ,1 .0 Teflon Zytel Lab Office 4 4 4 1 19 0.7 0.2,2.5 13 1.5 0.4, 5.7 11 Z1 0.5, 8.3 2 2.4 0.2. 26.9 EID151963 R L002172 . FPiangael 1R2eopfo1r3t - Study No.: S-088 DISCUSSION This study of leukemia at the Washington Works found no association between employment in any particular job or area grouping and the risk of developing leukemia. A statistically significant excess of leukemia.was observed in those who formerly worked as custodians or chemists prior to working at the Washington Works. These findings are based on few cases, thus wide confidence intervals bracket the odds ratios. This study could be missing a true positive association because of a number of factors. Misclassification of exposure as approximated by work history is possible and could lead to a false negative result. However, since exposure was assigned based on job title and without knowledge of the subject's disease status, this is unlikely. Misclassification of the disease status is also possible although unlikely given the high awareness of the disease at the site. Also, several types of leukemia are grouped together in this study. These types of leukemia probably have different etiologies, latency periods, and different age groups at risk. Unmeasured confounders such as cigarette smoking and non-occupational exposures could also play a role. The small number of cases severely limits the statistical power of the study to detect a true association. The positive finding of an excess of leukemia among those who worked as custodians and chemists prior to joining Du Pont is provoking. Both occupations have potential exposures to multiple chemical agents and exposures could be at high levels. Because of lack of specific information on the exposures in their former jobs, and lack of understanding of the selection process that brought them to Washington Works, no conclusions can be drawn. CONCLUSION This case-control study of leukemia found no association between employment in any particular Washington Works job or area grouping and the risk of developing leukemia. The site should be monitored for future cases of leukemia to determine if the excess persists and any patterns emerge with respect to workplace exposures. EID1S1M4 R L 002173 Page 13 of 13 Final Report - Study No.: S-0S8 ACKNOWLEDGEMENTS The authors would like to thank the Washington Works staff who collected the data, the Occupational Health consultants, Amy Berg and Bill Anderson and the Epidemiology Section programmers, Alan Mann and John Geriach, who assisted in the data processing and analysis for this study. REFERENCES Linet, M.S. The Leukemias: Epidemiologic Asoects, New York: Oxford University Press, 1985 O'Berg, M., Burke, C.. Chen, J.. Walrath, J., Pell, S., and Gallie, C., "Cancer incidence and mortality in the Du Pont Company: an update" in Journal of Occupational Medicine. 1987, 29:3, p. 245-252. Pell, S., O'Berg, M., and Karrh, B., "Cancer epidemiologic surveillance in the Du Pont Company" in Journal of Occupational Medicine. 1978, 20:11, 725-740. Rothman, Kenneth, Modem Epidemiology. Little, Brown, and Company, 1986, p. 297298. Statistics and Epidemiology Research Corporation, EGRET Reference Manual. Version 1, Revision 1, 1990. Walrath, J. and Burke, C. "An investigation into the occurrence of leukemia at Washington Works", Internal Report from the Epidemiology Section, Medical Division of the Du Pont Co., revised April, 1989. ----------- EID151965 6 RL002174