Risk of malignancies in relation to terrestrial gamma radiation in a Swedish population cohort.

Results of epidemiological studies on terrestrial gamma radiation (TGR) and related malignancies have not been consistent. This study is a thorough examination of this relationship. Records on all individuals living in two Swedish counties in 1973, along with their annual dwelling coordinates during the 28-year follow-up period, were retrieved from the National Archives and Statistics Sweden. We used Geographical Information System (GIS) to match the individuals' dwelling coordinates annually to the TGR given in 200×200 m grids produced by the Geological Survey of Sweden. Cases of malignancies and deaths were retrieved from the Swedish Cancer Register. During the follow-up period 61,503 incident cases were included in the analyses and in total 11 million person-years were recorded. Cox regression was used both in a linear continuous model and analyses of six exposure categories. Adjustments were made for sex, age, and population density. The hazard ratio (HR) per 100 nanoGray/hour (nGy/h) was significantly increased for total malignancies and for several sites; however, contrary to expectations, an obvious and anticipated linear exposure-response relationship could not be identified. With the lowest exposure category (0-60 nGy/h) as reference, a statistically significantly increased HR for total malignancies was seen in all exposure categories, except in the highest category 96-366 nGy/h. For breast cancer, thyroid cancer and leukaemia an obvious exposure-response could not be seen.