Journal of Occupational & Environmental Medicine:
Occupational Risk Factors for Cancer of the Gastric Cardia: Analysis of Death Certificates From 24 US States
Cocco, Pierluigi MD; Ward, Mary H. PhD; Dosemeci, Mustafa PhD
From the Institute of Occupational Medicine, University of Cagliari, Italy (Dr Cocco); and the Occupational Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Md. (Dr Ward, Dr Dosemeci).
Address correspondence to: Mary H. Ward, PhD, Occupational Epidemiology Branch, Division of Cancer Epidemiology and Genetics National Cancer Institute, 6130 Executive Boulevard, EPN Room 418, Bethesda, MD 20892.
We evaluated the risk of gastric cardia cancer by occupation and industry in a case-control study using information from death certificates for 24 US states in 1984-1992. One thousand fifty-six cases of gastric cardia cancer were identified among men aged 20 years or more, including 1,023 whites and 33 blacks. Controls were 5,280 subjects who died of nonmalignant diseases, 5:1 matched to cases by geographic region, race, gender, and 5-year age group. Among white men, occupations with elevated risk included financial managers (odds ratio [OR] = 6.1; 95% confidence interval [CI], 1.3-28.8), janitors and cleaners (OR = 1.7; 95% CI, 1.0-2.9), production inspectors (OR = 3.2; 95% CI, 1.5-6.9), and truck drivers (OR = 1.5; 95% CI, 1.0-2.2). Industries with elevated risk included pulp and paper mills (OR = 2.0; 95% CI, 1.0-37), newspaper publishing and printing (OR = 2.6; 95% CI, 1.0-6.3), industrial and miscellaneous chemicals (OR = 2.0; 95% CI, 1.0-3.9), water supply and irrigation (OR = 5.6; 95% CI, 1.6-19.9). Among black men, risks were nonsignificantly increased for subjects employed in railroads (3 cases, 2 controls) and for carpenters (3 cases, 0 controls). We created job-exposure matrices for asbestos, inorganic dust, metal dust, lead, polycyclic aromatic hydrocarbons, nitrogen oxides, nitrosamines, sulfuric acid, fertilizers, herbicides, other pesticides, and wood dust. Among white men, a consistent pattern of risk increase by level and probability of exposure was observed only for sulfuric acid mists, with a twofold excess (95% CI, 0.6-7.3) associated with high probability of high intensity exposure. A significant 30% increase in risk was observed for those subjects with a high probability of exposure (all levels combined) to lead, and a 60% increase was observed for subjects with high-level exposure to lead (all probabilities combined). However, crosstabulation of gastric cardia cancer risk by probability and level of exposure to lead did not show consistent trends. Asbestos exposure also showed an overall 50% increase but no consistent trends among white men. None of the 12 occupational hazards showed an association with risk for black men.
While stomach cancer incidence and mortality have been declining worldwide over the past decades,1,2 there has been a recent rise in cancer of the gastric cardia and adenocarcinoma of the esophagus.3 Among whites, the male-to-female ratio for cancer of the gastric cardia is 5.6:1. Improved diagnostic capability does not seem to account for this time trend because other cancers of the upper digestive tract have not shown a similar increase.3
Few studies have specifically evaluated risk factors for cancer of the gastric cardia. Smoking and alcohol consumption were associated with risk in some4,5,5a but not in all studies.6 Dietary risk factors for gastric cancer have also been positively associated with the cardia subsite.6 A predominance of higher social class and professional occupations among incident male cases of gastric cardia cancer4,7 has been reported. However, a recent study5a found no association between education and income levels and risk of gastric Cardia Cancer. Occupational associations were recently evaluated among men and women by broad categories, using the 24 US states' death certificate database,8 and a significant increase in risk was reported for white men and white women among administrative managers, health professional, and rubber and chemical manufacturing workers. Among women, an increase in risk was associated also with work in the wholesale trade. We extended the previous investigation among men to include death certificates of an additional 3 years. Also, we evaluated more detailed occupation and industries and applied a priori job-exposure matrices for 12 occupational exposures that have previously been postulated as stomach cancer risk factors.
Since 1984, the National Cancer Institute, the National Institute for Occupational Safety and Health, and the National Center for Health Statistics have supported the coding of occupation and industry titles on death certificates from 24 US states according to the 1980 US Census occupation and industry codes.9 This database currently consists of a total of 4.5 million death certificates, covering the years 1984-1992. Details on the 24 US states' death certificates database have been reported elsewhere.10 Only usual occupation and industry is reported on the death certificate, and no duration of employment is available. Mortality records among male subjects aged 20 years or more were used to evaluate the risk of gastric cardia cancer associated with usual employment in a specific occupation and industry. Of 13,040 stomach cancer deaths among men aged 25 years or older, there were 1,056 deaths (8.1%) from cancer of the gastric cardia (ICD-9 code 151.1), including 1,023 whites and 33 blacks. Subjects of other race groups were also excluded because of small numbers. An evaluation of gastric cardia cancer risk by broad occupation and industry categories among women was included in the above-cited report.8 Five controls per case were selected from among subjects who died of non-malignant diseases, and were frequency-matched to cases by geographic region, race, gender, and 5-year age-group. Decedents in each occupation or industry were compared with all others.
Based on the 1980 US Census list of occupations and industries,9 we used job-exposure matrices to evaluate occupational exposure to 12 potential carcinogenic exposures; namely, inorganic dust, asbestos, metal dust, lead, polycyclic aromatic hydrocarbons, nitrogen oxides, nitrosamines, sulfuric acid, fertilizers, herbicides, other pesticides, and wood dust. Most of these occupational hazards were discussed as risk factors for stomach cancer in a recent review.11 An estimate of intensity level (none = 0, low = 1, medium = 2, high = 3) and probability (none = 0, low = 1, medium = 2, high = 3) of exposure to each of the 12 occupational hazards was developed by an experienced industrial hygienist (M.D.) and an occupational physician (P.C.) for each 3-digit occupation and industry code.
Intensity of exposure was estimated based upon literature information, computerized exposure databases (Occupational Safety and Health Administration files, National Institute for Occupational Safety and Health inspections database), and personal experience of the same professionals. The probability index was estimated based on the proportion of exposed workers within a given job title or industry and the number of occupations or industries with the same 1980 Census code. In addition, occupations were characterized into two groups depending upon the sources of exposure. If exposure was determined by the occupation itself regardless of industry, final intensity and probability scores were obtained by squaring the occupational scores. If exposure was determined by both occupation and industry, then the probability of exposure was dependent upon the industry, and the final score of intensity resulted from multiplying the intensity scores of both the occupation and industry. The final scores of probability and intensity of exposure were further categorized within four levels (none = 0, low = 1-2, medium = 3-4, high = 6-9).
Odds ratios (ORs) were estimated by logistic regression, and 95% confidence intervals (95% CI) by the Wald method, using the GMBO program in the Epicure software.12 Included in the logistic regression model were age (5-year age categories), marital status (never married vs ever married), rural vs urban residence, and socioeconomic status (SES) (five categories, based on Green's Standardized Scores for Specific Occupations13). Statistical significance of trends in risk by increasing intensity level and probability of exposure were tested by dividing the regression coefficients of the variables assumed as non-categorical by their standard error to generate a Z-statistic. Under the null hypothesis, this test behaves as a normal standard deviate.14 P values were two-tailed.
Average age at death of gastric cardia cancer cases did not vary by race (white men, 64.2 ± 11.1 years; black men, 64.4 ± 10.9). Among white men, never having been married was associated with a significantly reduced risk of gastric cardia cancer (OR = 0.6; 95% CI, 0.4-0.8), while the risk was positive among black men (OR = 2.4; 95% CI, 0.9-6.1). A 20% increase in risk among both study groups was associated with rural residence, defined as residence in the countryside or in a city with less than 250,000 inhabitants (white men: OR = 1.2; 95% CI, 1.0-1.4; black men: OR = 1.2; 95% CI, 0.5-2.7).
Among white men, risk increased significantly by SES category: ORs were 1.3 (95% CI, 1.0-1.6), 1.7 (95% CI, 1.4-2.2), 2.1 (95% CI, 1.6-2.6), and 2.5 (95% CI, 1.9-3.4) in consecutive, increasing SES categories, compared with the lowest category. A significant 2.5-fold increase in risk was also observed among black men in the second-lowest SES category, compared with the lowest, but numbers in the higher SES categories were not sufficient to confirm the trend observed among whites. These covariates and age in five-year groups were included in logistic regression models to estimate risk associated with each industry, occupation, or exposure.
Table 1 shows the risks for gastric cardia cancer associated with specific industries with five cases or more among white men. A significant or borderline significant increase in risk was observed among the following industries: bakery products (OR = 3.9; 95% CI, 1.3-11.2); pulp, paper, and paperboard mills (OR = 2.0; 95% CI, 1.0-3.7); newspaper publishing and printing (OR = 2.6; 95% CI, 1.0-6.3); industrial and miscellaneous chemicals manufacturing (OR = 2.0; 95% CI, 1.0-3.9); miscellaneous manufacturing industries (OR = 3.3; 95% CI, 1.3-8.4); water supply and irrigation (OR = 5.6; 95% CI, 1.6-19.9); colleges and universities (OR = 2.1; 95% CI, 1.1-4.2); and social services not elsewhere classified (OR = 7.8; 95% CI, 1.6-37.8). Notably, among industries reported to be associated with stomach cancer in other studies, risk was not increased for farming, mining, the construction industry, manufacturing of tires and other rubber products, or woodworking. Non-significant increases of twofold or more were observed for meat products and iron and steel foundries. Among black men, a non-significant positive association was observed only among persons employed in railroads (OR = 13.1, based on three cases and two controls).
Table 2 presents the ORs for occupations with five exposed deaths or more among white men. A significant or borderline significant increase in risk was observed for financial managers (OR = 6.1; 95% CI, 1.3-28.8); janitors and cleaners (OR = 1.7; 95% CI, 1.0-2.9); production inspectors, checkers, and examiners (OR = 3.2; 95% CI, 1.5-6.9); and drivers of heavy trucks (OR = 1.5; 95% CI, 1.0-2.2).
Some non-significant excess risks corresponded to results among industries: printing machine operators (OR = 1.4; 95% CI, 0.5-3.7), bakers (OR = 2.0; 95% CI, 0.5-7.8), and water and sewage treatment plant operators (OR = 5.2; 95% CI, 1.1-25.0). The last two occupations, both based on three cases, are not reported in the tables. Other non-significant increases in risk among occupations previously associated with stomach cancer11 were observed for firefighters, brickmasons and stonemasons, and construction laborers, but not for farmers, farmworkers, and butchers and meatcutters. Among blacks, only carpenters showed a non-significant excess risk (three cases and no controls).
The application of the 12 job-exposure matrices for potential occupational risk factors for gastric cardia cancer showed an increase in risk associated with the probability of exposure to asbestos and sulfuric acid among white men (Table 3), with a significant positive trend for sulfuric acid. The increase in risk associated with high probability of exposure to asbestos was based on only one death. High probability of exposure to lead was associated with a significant 30% increase in risk among white men, and the risk was 1.2 for all probability categories combined among black men (data not shown). Risk patterns by intensity of exposure (Table 4) showed an inverse trend for asbestos, while sulfuric acid again showed a significantly positive trend (P < 0.01). The highest intensity level of lead exposure was associated with a significant 60% increase in risk.
More detailed analyses were possible only for white men. Among whites, risk patterns associated with exposure to asbestos, lead, and sulfuric acid were also explored within categories of probability and intensity of exposure (Table 5). A consistent pattern of increasing risk by probability within intensity level categories and by intensity level within probability categories was observed for sulfuric acid, with a few significantly positive trends. Lead showed some increase in the highest level of exposure, but only one case was represented in the high-level/high probability of exposure cell. Risk patterns associated with probability and intensity of exposure to asbestos were inconsistent. However, only one subject had a high probability of exposure to asbestos.
The results of the present case-control analysis suggest that occupational exposure to sulfuric acid mists may be a risk factor for cancer of the gastric cardia. Applying a job-exposure matrix to the occupation/industry combinations in the death certificates of study subjects resulted in significantly increasing risks by probability and intensity level of exposure. Exposure to sulfuric acid mists occurs in the pulp and paper mills industry, where we observed a significant increase in risk. Exposure to acid mists also occurs among water and sewage treatment plant operators, who showed a fivefold increase in risk in this study (OR = 5.2; 95% CI, 1.1-25.0, based on three cases; not shown in the tables). Other exposures in these jobs include chlorine and its acids. Gastroesophageal reflux has been suggested as an important risk factor for adenocarcinomas of the esophagus and gastric cardia,15 possibly through the production of premalignant epithelial changes after contact with the acidic content of the stomach. Whether the inhalation and subsequent ingestion of acid mists in occupational settings might also cause irritation and subsequent premalignant changes in the gastric cardia mucosa is unknown. Significant increases in gastric cardia cancer risk were also associated with high probability (all levels combined) of and high-level (all probabilities combined) exposure to lead, by crosstabulation of risk by the two exposure metrics did not show consistent trends. Only one death was represented in the high probability/high level of exposure cell. Results for asbestos were also inconsistent. Only one subject had high probability of exposure to asbestos, and risk decreased by increasing level of exposure in the low- and medium-probability categories. None of the other occupational hazards explored using job-exposure matrices showed any pattern of increasing risk either by probability or intensity level of exposure.
The present analysis of risk of gastric cardia cancer by occupation and industry showed some similarities with reports in the literature for all stomach cancer combined.11 However, occupations-such as farmers, and butchers and meatcutters-or industries-such as mining, construction, manufacturing of tires and other rubber products, and wood working, previously associated with stomach cancer,11-were not associated with gastric cardia cancer risk in this study among white men. A non-significant excess risk for carpenters was limited to black men. Also, exposure to inorganic dust, metal dust, and nitrosamines, possible candidates as occupational carcinogens for gastric cancer,11 did not show an association in this study. These disparities may be due to the limitations of this analysis in evaluating occupational exposures, or they may reflect a different etiology for gastric cardia cancer, compared with other cancers of the stomach.
Cancer of the gastric cardia is largely underreported on the death certificates. SEER incidence data indicate that it accounts for about 25% of all gastric cancer among men,8 while it accounts only for 8.1% of stomach cancer deaths in the 24 US states' death certificate database. We observed an association of gastric cardia cancer with higher SES, which may be due to differential underreporting by the SES of the decedent. We adjusted risk estimates by SES to prevent underestimation of the gastric cardia cancer risk associated with some occupations and industries belonging to low SES.
Interpretation of results from hypothesis-generating studies is difficult, as chance is a likely explanation when multiple comparisons are involved. Suggestions have been made to correct the level of statistical significance by the number of comparisons.16 Instead, we presented only risk estimates with at least five exposed deaths and explored the consistency of trends by probability and intensity of exposure.
Another limitation in the present study is that occupational information in the death certificates is limited to only one occupation and industry title, with no duration of employment. Any resulting exposure misclassification would probably be non-differential. As a consequence, underestimation is most likely for risk associated with high-exposure categories, and positive trends such as those that we observed for exposure to sulfuric acid would be possible only under very extreme conditions.17 On the other hand, using the 24 US states' database provided the largest number of gastric cardia cancer deaths studied to date, which allowed the first detailed exploration of occupational associations. Further studies with more detailed exposure assessments, based on complete work histories and incident cases, are needed to further evaluate hypotheses generated by the present study.
1. Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide incidence of eighteen major cancers in 1985. Int J Cancer. 1993;54:594-606.
2. Howson CP, Hiyama T, Wynder EL. The decline in gastric cancer: epidemiology of an unplanned triumph. Epidemiol Rev. 1986;8:1-27.
3. Blot WJ, Devesa SS, Kneller RW, Fraumeni JF Jr. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA. 1991;265:1287-1289.
4. Kabat GC, Ng SKC, Wynder EL. Tobacco, alcohol intake, and diet in relation to adenocarcinoma of the esophagus and gastric cardia. Cancer Causes Control. 1993;4:123-132.
5. Vaughan TL, Davis S, Kristal A, Thomas DB. Obesity, alcohol, and tobacco as risk factors for cancers of the esophagus and gastric cardia: adenocarcinoma versus squamous cell carcinoma. Cancer Epidemiol Biomark Prevent. 1995;4:85-92.
5a. Gammon MD, Schoenberg JB, Ahsan H, et al. Tobacco, alcohol, and socioeconomic status and adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst. 1997;89:1277-1284.
6. Palli D, Bianchi S, DeCarli A, et al. A case-control study of cancer of the gastric cardia in Italy. Br J Cancer. 1992;65:263-266.
7. Powell J, McConkey CC. Increasing incidence of adenocarcinoma of the gastric cardia and adjacent sites. Cancer. 1990;62:440-443.
8. Ward MH, Dosemeci M, Cocco P. Mortality from gastric cardia and lower esophagus cancer and occupation. J Occup Med. 1994;36:1222-1227.
9. Bureau of the Census. Alphabetical Index of Industries and Occupations. [PHC80-R3]. Washington, DC: US Department of Commerce; 1982.
10. Burnett CA, Dosemeci M. Using occupational mortality data for surveillance of work-related diseases of women. J Occup Med. 1994;36:1199-1203.
11. Cocco P, Ward MH, Buiatti E. Occupational risk factors for gastric cancer: an overview. Epidemiol Rev. 1996;18:218-234.
12. Preston DL, Lubin JH, Pierce DA. Epicure. Seattle, WA: Hirosoft International Corp; 1990.
13. Green L. Manual for scoring socioeconomic status for research on health behavior. Public Health Rep. 1970;85:815-827.
14. Breslow NE, Day NE. Statistical Methods in Cancer Research: Volume I-The Analysis of Case-Control Studies. [IARC Scientific Publications No. 32.] Lyon, France: International Agency for Research on Cancer; 1980.
15. Chow WH, Finkle WD, McLaughlin JK, et al. The relation of gastroesophageal reflux disease and its treatment to adenocarcinomas of the esophagus and gastric cardia. JAMA. 1995;274:474-477.
16. Rothman KJ. Modern Epidemiology. Boston, MA: Little, Brown and Co; 1986.
17. Dosemeci M, Wacholder S, Lubin J. Does nondifferential misclassification always bias a true effect toward the null value? Am J Epidemiol 1990;132:746-7
THE HUMAN TOUCH
If you call Allegheny Petroleum Products Company, don't expect a nasal-voiced recording to ask you to hold, push a button, or listen to a menu. In fact, the company, in the Pittsburgh suburb of Wilmerding, is so proud of its ban on automated telephone systems that its brochures tout its use of human operators.
-From Schogol M. Personal Briefing. Philadelphia Inquirer, April 14, 1998, p D3.
© Williams & Wilkins 1998. All Rights Reserved.