Despite the spontaneously decreasing incidence rates, cancer of the stomach is still one of the most frequent forms of cancer on a global scale. Our long-term objective is to understand the etiology and provide the scientific basis for prevention of stomach cancer with different locations (cardia or other) and histologic types (intestinal or diffuse). A population-based case-control study has been initiated in a large area of central and northern Sweden. The study population has the following important features: (1) the study-base is exactly defined; (2) all individuals are obliged to resort to hospitals within a defined area; (3) close collaboration with all departments included in the diagnosis and treatment of stomach cancer insures early identification of all incident cases and permits interviews to take place before surgical treatment; (4) virtually all cases are confirmed histologically; (5) the incidence rates vary almost two-fold between the five counties of the study, indicating a considerable heterogeneity in exposure to causative agents; (6) the intake of cured (notably salted) and smoked food has been extensive and the use of snuff widespread in the study population.
Specific aims. To test the hypotheses that causal relationships exist between gastric cancer and 'dietary patterns' during adolescence and past dietary exposure to nitrous compounds, salt and PAHs (before the age of 20, and 20 years prior to interview), and to evaluate the role of tobacco smoking and snuff dipping (notably in cancer of the cardia) and associations between antibodies to Helicobacter pylori in the serum and stomach cancer. The proposed protective effects of fruits, vegetables, vitamins C and E, retinol, beta-carotene and allium vegetables will also be examined. The role of prolonged use of H-2 receptor blockers, SES, occupational exposures, peptic ulcer, pernicious anemia and familial occurrence of cancer as determinants of the risk will also be assessed. Methodology. In an extended pilot study, about 80% of all eligible patients younger than 80 years and diagnosed consecutively as having castric cancer, and population controls, stratified by age and gender, have undergone professional face-to-face interviews. A preliminary evaluation of interview data shows that the validity (concordance in dietary reports between interview subjects and their siblings) and the reliability (estimated from repeated interviews after one year) are both on an acceptable level. Univariate, stratified and multivariate analyses will be carried out using standard epidemiologic methods with the relative risk as the central measure of association. Encouraged by the well- functioning routines, we propose an extension of the study to include 600 cases and 1,200 controls in order to increase the power to study risk by tumor location and histologic type, as well as infrequent exposures and interactions between nutritional factors and between diet and smoking.
Showing the most recent 10 out of 13 publications
