THE RELATIONSHIP OF CANCER INCIDENCE TO TERRESTRIAL RADIATION AND POPULATION DENSITY IN CONNECTICUT, 1935–1974

The relationship of cancer incidence to terrestrial radiation and population density was investigated. Cancer incidence was obtained using 40 years of age-standardized data from the Connecticut Tumor Registry, and environmental radiation was estimated using data from an airborne gamma radiation survey of the entire state. These variables were examined ecologically, using the 169 towns of the state as the analytic units in a weighted regression analysis. The study design involves a large population base in a state having relatively high terrestrial radiation exposure levels overall and reasonable variation in exposure between towns. For all cancer combined, only one of the eight sex-specific analyses by decade yielded a significant radiation regression coefficient, and this was negative. In the sex- and site-specific analyses, almost all the coefficients for radiation were not significantly different from zero. In contrast, significant positive relationships of cancer incidence with population density were found for all cancer, for cancer of the lung for both sexes, for stomach, colonic, and prostatic cancer for males, and for lymphomas, thyroid, breast, and ovarian cancer for females. Both the radiation and population density relationships were adjusted for socioeconomic status. Socioeconomic status was significantly negatively associated with stomach and lung cancer in males and with cervical cancer in females; it was also positively associated with lymphomas and breast cancer in females. A power calculation revealed that, despite the relatively large size of this study, there was only a small probability of detecting a radiation effect of the strength anticipated from previous estimates.