CANCER INCIDENCE AND RISKS IN SELECTED AGRICULTURAL SETTLEMENTS IN THE NEGEV OF ISRAEL

Medical staff of two Negev kibbutzim invited epidemiologists to help t hem investigate cancer rates among their members. Our objectives were (a) to determine whether the cancer rate in the kibbutzim was elevated or abnormal and (b) to determine the role of agricultural and other r elevant exposures if cancer incidence was elevated. We validated cases of cancer by kibbutz records and by surveying other information; we c omputed expected values on the basis of the age-sex-calendar period an d site-specific cancer incidence rates reported by the Israel Cancer R egistry for the entire population; and we compared the data for the 2 kibbutzim with data derived for similar age and sex groups in 2 other kibbutzim, which were assumed not to have increased cancer rates, in a ddition, we planned and conducted a case-referent study, including the design, pretest, and use of questionnaires, including data about life time exposures (i.e., type of work and its duration, agricultural and industrial chemicals, smoking and alcohol use, demographic variables, health experiences, and family history). In only one of the kibbutzim, for which high cancer rates were suspected, was there significant exc ess for all sites in persons who were less than 40 y of age, in one of the ''comparison'' kibbutzim, we found increased cancer rates overall . Much of the excess in the high cancer kibbutzim was in hematological cancer (i.e., leukemia and lymphoma). Multiple years of work in field s, orchards, and landscape, as well as orchard work that commenced bef ore 1960, were associated with increased risk of cancer (p < .08). We also found an association between cancer rate and numbers of industria l chemicals used (p < .08). Pipe and cigarette smoking were also assoc iated with increased cancer incidence. In the multivariate analysis, t he association with calendar year in which orchard work was started an d multiple exposures to industrial chemicals was stronger than associa tions noted in the univariate analyses. Although duration of agricultu ral work or multiple industrial exposures were clearly associated with increase in cancer risk, we were unable to identify the causal role o f specific agent(s). Nonetheless, educational programs for cancer prev ention can be based, in part, on the results of such a study.