Document nmb0xMBOvy54R6g5jQmO6OZm

fj .foPli tun l AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 36:70-74 (1999) Occupational Risk Factors for Cancer of the Central Nervous System (CNS) Among US Women Pierluigi Cocco, md,1* Ellen F. Heineman, PhD,2 and Mustafa Dosemeci, PhD2 Background: In a recent report, wefound an elevated risk of cancer of the central nervous system (CNS) in several occupations and industries, and a modest association with exposure to solvents and to contact with the public. Methods: Tofurther explore the occupational risk ofCNS cancer among women, we extended the analysis of the previous death certificate-based case-control study, including 12,980 female cases (ICD-9 codes 191 and 192) in 24 US states in 1984-1992 and 51,920 female controls who died from diseases other than malignancies and neurological disorders. We applied newly designedjob-exposure matricesfor 11 occupational hazards, previously reported as brain cancer riskfactors, to the occupation and industry codes in the death certificates. We also conducted a separate analysis of 161 meningioma cases (ICD-9 codes 192.1 and 192.3), a tumor morefrequent among women, particularly in the postmenopausal age group. Results: Overall, CNS cancer risk showed a 20-30% increase among women exposed to electromagnetic fields (EMF), methylene chloride, insecticides andfungicides, and contact with the public. Riskfor meningioma was elevated among women exposed to lead (OR = 1.9; 95% Cl 1.0, 3.9). CNS cancer did not show a clearpattern ofrisk increase by probability and intensity of exposure to any of the explored riskfactors. Cross-classification by probability and intensity of exposure did not reveal any significant trend. Cases were too few to explore trends ofmeningioma by probability and intensity ofexposure to lead. Conclusions: We did notfind evidence ofa strong contribution of11 occupational hazards to the etiology of CNS cancer. However, limitations ofthe occupational information might have reduced our ability to detect clear patterns of risk. Am. J. Ind. Med. 36:70-74, 1999. 1999 Wiley-Liss. Inc. KEY WORDS: neoplasms; central nervous system; electromagnetic fields; solvents; pesticides; epidemiology;job-exposure matrices; women INTRODUCTION In a recent article, we investigated the risk of mortality from central nervous system (CNS) tumors associated with 'Institute of Occupational Medicine, University of Cagliari, Cagliari, Italy Occupational Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland, USA 'Correspondence to: Pierluigi Cocco, M.D., Istituto di Medicina del Lavoro, University di Cagliari, via S. Giorgio 12, 09124 Cagliari, Italy. E-mail: coccop@pacs.unica.it Accepted: 1 March 1999 occupation and industry in the USA by gender and race [Cocco et al., 1998a], Our findings revealed an increased CNS cancer risk among education- and health-related occu pations and industries, and a modest association with exposure to solvents and to contact with the public. Expo sure to these risk factors was assessed using simple jobexposure matrices. To further explore the association be tween occupation and CNS cancer in U.S. women, we designed more detailed job-exposure matrices for 11 occupa tional hazards that may be associated with CNS cancer based on our previous study and others [Inskip et al., 1996], We applied these job-exposure matrices to occupations and industries appearing in the death certificates of women who died from CNS cancer in 24 US states in 1984-1992. 1999 Wiley-Liss, Inc. CNS Cancer and Occupation in US Women 71 METHODS Case and Control Selection Details on the 24 US states database and the casecontrol study design are reported elsewhere [Cocco et al., 1998a]. Briefly, we identified 12,980 cases of cancer of the brain and other parts of the CNS (ICD-9 codes 191 and 192) among women who died in 24 states in 1984-1992. They included 161 cases of meningioma (ICD9 codes 192.1 and 192.3), who were the subject of a separate analysis. For each case, we selected four controls among women who died from nonmalignant diseases, excluding neurological disor ders, frequency-matched by state, race, and 5-year agegroup. Exposure Assessment We designed new job-exposure matrices for 11 occupa tional hazards: electromagnetic fields (EMF), solvents, chlo rinated aliphatic hydrocarbons (CAHs), methylene chloride, benzene, lead, nitrosamines, polycyclic aromatic hydrocar bons (PAHs), insecticides and fungicides, herbicides, and contact with the public. An estimate of intensity level of exposure (none = 0, low = 1, medium = 2, high = 3) and probability of exposure (none = 0, low = 1, medium = 2, high = 3) to each hazard was developed by two authors (M.D. and P.C.) for each 3-digit occupation and each 3-digit industry U.S. Census code. Afinal intensity level score and a final probability score were then developed for each occupa tion/industry combination appearing in study subjects' death certificates. Intensity level of exposure was estimated based on industrial hygiene and occupational health textbooks [Parmeggiani, 1983; Zenz et al., 1994], computerized expo sure databases (OSF1A files, NIOSH inspections database), unpublished industrial hygiene reports, and professional experience. The intensity score for contact with the public, a surrogate for exposure to transmissible pathogens, was calculated differently. Five components were identified and scored individually, namely, frequency of contact with the public (episodic or daily), ambient spaciousness in which contact with the public occurs (outdoor, large indoor meet ing room or classroom, standard size room), average number of daily contacts (^20, 21-50, 51-100, 101), average health status of contacted persons (good, possible contact with sick persons, likely contact with sick persons), and contact mainly with elders or children. The individual scores were then combined to derive the public contact intensity score, which was categorized in four levels (none, low, medium, high). The probability index was estimated based on the proportion of workers within a given job title or industry that are typically exposed to the hazard. For each study subject, a final intensity level score and a final probability score was calculated for each hazard by combining the occupation and industry scores in the follow ing ways: 1) If both the occupation and industry involved exposure to the hazard (e.g., driver in a chemical industry), then the final intensity score was equal to the product of the individual intensity scores. The final probability score was that attributed to the industry code alone. 2) If exposure was related only to occupation, regardless of industry (e.g., painter), only the intensity and probability scores related to occupation were used to derive the final scores. Intensity score was squared in these instances to maintain consistency in units. The final intensity and probability scores were then grouped into four levels (unexposed, low, medium, and high). Low, medium, or high probability and intensity of exposure are meant as comparisons within a given exposure and are not comparable across exposures. This procedure for designing job-exposure matrices has been validated in a study of asbestos exposure and risk of peritoneal cancer [Cocco and Dosemeci, 1999] and has been applied in two other studies [Cocco et al., 1998b; 1999]. The categorization of exposure probability and intensity levels in the newly designed job-exposure matrices resulted in greater sensitivity in identifying exposures, particularly in the low probability / low intensity groups. In fact, for each risk factor considered in our previous analysis, a number of occupa tions and industries considered as unexposed in the former "yes/no" exposure assessment were now included in the most sensitive exposure category (low probability / low intensity). Therefore, the total number of people exposed to each risk factor is greater in this study compared to the previous one. Statistical Analysis Odds ratios (ORs) were calculated with logistic regres sion for each workplace exposure, adjusting for marital status (never vs. ever married), socioeconomic status (SES) (three levels, based on Green's Standardized Score for Specific Occupations [Green, 1970]), and age at death (continuous). ORs and 95% confidence intervals (95% Cl) were calculated with the Wald method using the GMBO program in the Epicure software package [Preston et al., 1990]. The statistical significance of the linear trend by increasing intensity and probability of exposure to the 11 CNS cancer risk factors was tested by dividing the regres sion coefficients of the variables assumed as noncategorical by their standard error to generate a Z statistic. Under the null hypothesis, this test behaves as a normal standard deviate [Breslow and Day, 1980]. Two-tailed P values were considered throughout this study. RESULTS Table I shows ORs for CNS cancer and meningioma for each of the 11 workplace risk factors (none vs. any 72 Cocco et al. TABLE I. Risk of Mortality From CNS Cancer and Meningioma Associated With 11 Occupational Hazards: None vs. Any Exposure: Female Deaths in 24 US States 1984-1992 Exposure CNS cancer Meningioma exposed exposed cases/controls OR 95% C.I. cases/controls OR 95% C.l. Electromagnetic fields Solvents Chlorinated aliphatic hydrocarbons Methylene chloride Benzene Lead Nitrosamines Polycyclic aromatic hydrocarbons Insecticides and fungicides Herbicides Contact with the public 2,901/10,567 1,047/4,059 1,585/6,414 867/3,239 711/2,740 366/1,459 289/11,183 364/1,457 210/725 61/240 2,959/9,724 1.2 1.1-1.2 1.1 1.1-1.2 1.1 1.1-1.2 1.2 1.1-1.3 1.1 1.0-1.2 1.1 1.0-1.2 1.1 1.0-1.3 1.1 1.0-1.3 1.3 1.1-1.5 1.2 0.9-1.6 1.2 1.1-1.2 34/10,567 14/4,059 25/6,414 13/3,239 12/2,740 9/1,459 3/1,183 2/1,457 4/725 0/240 39/9,724 0.9 0.6-1.4 1.0 0.6-1.8 1.1 0.7-1.8 1.2 0.7-2.2 1.3 0.7-2.5 1.9 1.0-3.9 0.7 0.2-2.2 0.4 0.1-1.6 1.6 0.6-4.3 ---- 1.4 0.9-2.1 exposure). As observed in our previous study, there was a modest, but statistically significant, increase in risk of mortality from CNS cancer (10-30%) among women ex posed to solvents, insecticides and fungicides, and contact with the public. Unlike the previous study, risk associated with exposure to EMFs was significantly increased, and risk associated with exposure to herbicides was no longer statistically significant. Among the other risk factors, only CAHs and methylene chloride showed a statistically signifi cant increase in CNS cancer risk. Risk of meningioma was elevated among women exposed to lead, insecticides and fungicides, and who had contact with the public, but the findings were not statistically significant. There was no increase in CNS cancer risk with increas ing probability of exposure to any of the occupational hazards (Table 11). The greatest excess risk (40-50%) was associated with high probability of exposure to insecticides and fungicides (but the same excess was also observed for low probability of exposure) and to herbicides. Exploring trends in CNS cancer risk by intensity of exposure also did not show clear patterns of risk increase (Table III). The trend was inverse with intensity of exposure to methylene chlo ride. The cross-tabulation of risk by categories ofprobability and intensity of exposure did not provide further information for any of the explored risk factors, as most exposed women were in the low probability and low intensity cells. Risk in the high probability / high intensity cell was highest for exposure to insecticides and fungicides (OR = 1.5; 95% Cl 1.1, 2.1) and herbicides (OR = 1.7; 95% Cl 1.2, 2.3). After excluding subjects with any probability of exposure to herbicides, risk associated with insecticides and fungicides was still elevated (all probability categories combined: OR = 1.4; 95% Cl 1.1, 1.6; high probability / high intensity: OR = 1.4; 95% Cl 0.5, 3.7). For most exposures, cases of meningioma were very few or absent, particularly in the high exposure probability and intensity cells. Therefore, we were unable to explore trends in risk. Contact with the public showed a significant 60% excess risk associated with high probability of expo sure (test for trend: P = 0.05). DISCUSSION In this large case-control study of CNS cancer among women based on death certificates from 24 US states, we found equivocal associations with potential exposure to 11 workplace risk factors identified with the aid of jobexposure matrices. Risk showed a 20-30% increase among women exposed to EMF, methylene chloride, agricultural chemicals, and contact with the public. However, risk did not show a clear increase by probability or intensity of exposure to any of the occupational risk factors. Risk of meningioma was elevated among women exposed to lead, contact with the public, and insecticides and fungicides, but numbers of deaths were small. In our previous article [Cocco et al., 1998a], CNS cancer risk was elevated in a few industries and occupations related to EMF exposure, such as manufacture, use, mainte nance, and sale of electrical devices and telephones, as well as for air transportation. Risks were generally consistent between genders, while these occupations were seldom represented among African-American cases to evaluate race consistency in CNS cancer risk. Our new analysis of the same data with a more detailed job-exposure matrix for EMF showed a 20% increase in CNS cancer risk among women, but no increasing trend with exposure probability and intensity. The 24 states death certificate database does not provide enough information to discriminate type and CNS Cancer and Occupation in US Women 73 TABLE II. Risk of Mortality From CNS Cancer for 11 Occupational Hazards: by Probability of Exposure; Female Deaths in 24 U.S. States 1984-1992 Exposure None Cases/ controls OR Low Cases/ controls OR Probability of exposure Medium 95% Cl Cases/ controls OR 95% Cl High Cases/ controls OR 95% Cl Electromagnetic fields 10,079/41,393 1.0 2,312/8,115 1.2 1.1-1.2 255/1,045 1.2 1.0-1.4 Solvents 11,933/47,871 1.0 475/2,028 1.1 1.0-1.2 163/483 1.2 1.0-1.4 334/1,367 1.2 1.0-1.3 409/1,548 1.2 1.0-1.3 Chlorinated aliphatic hydrocarbons 11,395/45,506 1.0 791/3,258 1.1 1.0-1.2 545/1,976 1.2 1.1-1.3 249/910 1.2 1.0-1.4 Methylene chloride Benzene 12,113/48,681 12,269/49,180 1.0 1.0 756/2,839 1.2 1.1-1.3 83/311 1.2 1.0-1.6 170/711 1.0 0.9-1.2 422/1,602 1.1 1.0-1.2 28/89 119/427 1.3 0.9-2.0 1.2 1.0-1.5 Lead 12,614/50,461 1.0 214/865 1.1 0.9-1.3 94/381 1.0 0.8-1.3 58/213 1.2 0.9-1.6 Nitrosamines 12,691/50,737 1.0 174/695 1.1 1.0-1.3 88/371 1.1 0.9-1.4 27/117 1.0 0.7-1.6 Polycyclic aromatic hydrocarbons 12,616/50,463 1.0 197/772 1.1 1.0-1.3 52/224 1.0 0.7-1.4 115/461 1.1 0.9-1.4 Insecticides and fungicides 12,770/51,195 1.0 125/417 1.4 1.1-1.7 19/89 0.9 0.6-1.6 66/219 1.4 1.1-1.9 Herbicides Contact with the 12,919/51,680 1.0 7/33 1.0 0.4-2.3 7/61 0.5 0.2-1.2 47/146 1.5 1.1-2.1 public 10,021/42,196 1.0 260/1,031 1.1 1.0-1.3 99/363 1.2 0.9-1.5 2,600/8,330 1.2 1.1-1.2 TABLE III. Risk of Mortality From CNS Cancer Associated With 11 Occupational Hazards: by Intensity of Exposure; Female Deaths in 24 U.S. States 1984-1992 Exposure None Cases/ controls OR Low Cases/ controls OR Probability of exposure Medium 95% Cl Cases/ controls OR 95% Cl High Cases/ controls OR 95% Cl Electromagnetic fields 10,079/41,393 1.0 2,200/7,721 1.2 1.1-1.2 Solvents 11,933/47,871 1.0 474/1,935 1.1 1.0-1.2 616/2,507 1.1 1.0-1.3 85/299 422/1,456 1.2 1.1-1.4 151/668 1.3 1.0-1.6 1.0 0.8-1.2 Chlorinated aliphatic hydrocarbons Methylene chloride Benzene Lead Nitrosamines 11,395/45,506 12,113/48,681 12,269/49,180 12,614/50,461 12,691/50,737 1.0 1.0 1.0 1.0 1.0 802/3,549 1.0 1.0-1.1 370/1,316 1.3 1.1-1.5 266/976 1.1 1.0-1.3 187/676 1.2 1.0-1.4 19/79 1.0 0.6-1.7 643/2,272 1.3 1.1-1.4 140/593 345/1,255 1.2 1.1-1.4 152/668 245/959 1.1 1.0-1.3 200/805 138/623 1.0 0.8-1.2 41/160 212/878 1.1 0.9-1.3 58/226 1.0 0.9-1.2 1.0 0.8-1.2 1.1 0.9-1.3 1.1 0.8-1.6 1.1 0.9-1.5 Polycyclic aromatic hydrocarbons 12,616/50,463 1.0 157/608 1.2 1.0-1.4 141/501 1.2 1.0-1.5 66/348 0.9 0.7-1.1 Insecticides and fungicides Herbicides Contact with the 12,770/51,195 12,919/51,680 1.0 1.0 125/380 3/15 1.5 1.2-1.8 0.8 0.2-2.8 27/123 7/38 1.0 0.7-1.5 0.9 0.4-1.9 58/222 51/187 1.2 0.9-1.7 1.3 0.9-1.8 public 10,021/42,196 1.0 980/3,309 1.2 1.1-1.3 1,177/3,385 1.2 1.1-1.3 802/3,030 0.9 0.8-1.0 74 Cocco et al. characteristics of EMF exposure in our job-exposure matrix. Therefore, we conclude that such a study design is not powerful enough to exclude an association between CNS cancer and occupational exposure to EMF. Poor occupational information in the death certificates was a major limitation in this study. Occupation and industry listed on the death certificate represent only a fraction of the work history for each subject, either the "usual" or the last occupation. The 3-digit US Census code may have not been specific enough to accurately identify exposures. Also, women might have different job duties or exposure levels than men in the same occupation. The resulting misclassification of exposure may have impaired the specificity of our job-exposure matrix and weakened positive associations. These disadvantages are particularly important in studies involving women, as the reliability of the occupational information in the death certificate is poorer for women than men [Schade and Swanson, 1988]. Diagnostic bias is also likely to occur in death certificatebased case-control studies, as mortality from all causes combined is generally greater and reliability of death certificate diagnosis is poorer among low SES groups. Therefore, low SES occupations could be underrepresented among cases (subjects with a well-defined diagnosis) and overrepresented among controls (including other and more generic causes of death), which could obscure true associa tions with occupational risk factors more frequent in low SES groups. To address the possibility of diagnostic bias, we controlled for SES in our analysis and evaluated exposures resulting from a variety of occupations and industries, instead of individual occupations and industries. Finally, the use of death certificates provides only limited possibility to control for confounding or effect modification by lifestyle factors or other occupational expo sures. We adjusted for marital status and residence in the analysis to reduce the effect of lifestyle factors on our results. In conclusion, we did not find evidence of a strong increase in CNS cancer risk among women potentially exposed to several occupational hazards suspected to in crease risk. However, numbers were not sufficient to rale out a more modest association. Interpreting inconsistent findings by gender may be difficult as women and men, although sharing the same occupational title, may perform different duties and have different exposures. Specially designed job-exposure matrices, based on the exposure experience of the female workforce, could be an asset. Further research is warranted to understand whether different duties at work or different gender-linked susceptibility to workplace carcino gens may play a role in gender-related variations of CNS cancer risk. REFERENCES Breslow NE, Day NE, 1980. Statistical methods in cancer research. Volume I. The analysis of case-control studies. Lyon, Trance: International Agency for Research on Cancer, IARC Sci Pub No 32. Cocco P, Dosemeci M. 1999. Peritoneal cancer and occupational exposure to asbestos: results from the application of a job-exposure matrix. Am J Ind Med 35:9-14. Cocco P, Dosemeci M, Heineman EF. 1998a. Occupational risk factors for cancer of the central nervous system: a case-control study on death certificates from 24 US states. Am J Ind Med 33:247-255. Cocco P, Figgs L, Dosemeci M, Hayes R, Linet MS, Hsing AW. 1998b. Case-control study of occupational exposures and male breast cancer. Occup Environ Med 55:599-604. Cocco P, Ward MH, Dosemeci M. 1999. Occupational risk factors for cancer of the gastric cardia: analysis of death certificates from 24 U.S. states. J Occup Environ Med (in press). Green LW. 1970. Manual for scoring socio-economic status for research on health behavior. Public Health Rep 85:815-827. Inskip PD, Linet MS, Heineman EF. 1996. Etiology of brain tumors in adults. Epidemiol Rev 17:382-414. Panneggiani L, editor. 1983.1LO encyclopaedia of occupational health and safety. Geneva: International Labour Office. Preston DL, Lubin JH, Pierce DA. 1990. Epicure. Seattle: HiroSofi International. Schade WJ, Swanson GM. 1988. Comparison of death certificate occupa tional and industry data with lifetime occupational histories obtained by interview: variations in the accuracy of death certificate entries. Am J Ind Med 14:121-136. Zenz C, Dickerson OB, Harveth EP, editors. 1994. Occupational medicine, third ed. Chicago: Mosby Yearbook.