In Finland many groundwater supplies have naturally very high levels of radionuclides. A case-cohort study was undertaken to examine the effect of internal exposure to radon and other radionuclides (radium and uranium) through drinking water on the risk of stomach cancer.

A cohort of people using drilled wells as their drinking water source was identified. The source population was defined as persons born between 1900-1930 residing at the same address from January 1 1967 to December 31 1980, and living within small area units identified as having at least nine-tenths of the population outside the municipal drinking water system (n= 144,627). The Finnish Cancer Registry was used to identify all stomach cancer cases diagnosed from 1981-1995 within the base cohort (n= 1,492). A reference group (control subcohort) of 4,590 people was randomly selected within the base cohort, stratified by age and sex to match the distribution of cases.

Information was gathered on the characteristics of wells and their use as a drinking water source. Most of the subjects were not long term users of drinking water from wells. The final study population consisted of 371 subcohort members and 107 stomach cancer cases that had used drinking water from drilled wells prior to 1981. Individual water consumption was not assessed. Water samples were taken from drilled wells for 274 subcohort members and 88 stomach cancer cases between July and November 1996. The drinking water was analysed for activity concentrations of radon, radium-226 and natural uranium using radiochemical and alpha spectrometric methods. Samples were analysed without knowledge of whether they came from the water supply of a case or a control.

Mean radon concentrations in the well water were 310 Bq/I for the stomach cancer cases and 320 Bq/I for the subcohort. The median radon concentration for these two groups was the same, 130 Bq/I. The mean radium concentrations were 0.024 Bq/I for the stomach cancer cases and 0.030 Bq/I for the subcohort, and the median radium concentrations were 0.007 and 0.010 Bq/L respectively. The median natural uranium concentration was 0.07 Bq/L for both cases and subcohort and the mean concentrations were 0.21 and 0.45 Bq/I respectively. The total alpha activity in the drinking water for both groups was similar. In about 15% of wells the concentration of radionuclides (expressed as Bq/L) exceeded action levels set by many regulatory authorities.

Overall, no association was found between the risk of stomach cancer and exposure to radon, radium or uranium. In fact, the hazard ratios for stomach cancer decreased for water with higher levels of all three radionuclides, and this finding was not affected by adjustment for age and sex.

The results of this study do not indicate an increased risk of stomach cancer associated with radon, radium or natural uranium in drinking water in this population with very high exposure levels. The study did not assess other risk factors for stomach cancer (Helicobacter pylori infection, salted and smoked foods) or protective factors (fresh fruit and vegetables).