Radium in drinking water and risk of bone cancer in Ontario youths: a second study and combined analysis.

OBJECTIVES

Radium induces bone sarcomas at high doses, but there is controversy about risk at low doses. A previous study in Ontario found an association between the presence of radium in birthplace water supplies and an increased risk of death from bone cancer in young people. An investigation was performed to test the findings of the previous study with an independent group of subjects for whom complete information on radium exposure would be obtained.

METHODS

A population based case-control study (238 cases; 432 controls) was conducted with incident cases of bone sarcoma identified from the Ontario cancer registry. Residential histories were collected by questionnaire and water samples were obtained and analysed for radium content.

RESULTS

There was an association between risk of osteosarcoma and birthplace exposures (odds ratios (ORs) and 90% confidence intervals (90% CIs) 1.77 (1.03-3.00) but not with lifetime measures of exposure. When lifetime exposure was dichotomised, the OR was 1.31 (0.76-2.24) for osteosarcoma. There was no trend with increasing exposure. Bootstrap resampling was used to simulate lifetime doses in a pooled analysis of 1293 subjects from the two Ontario studies. The ORs were 1.38 (1.08-1.73) for all sarcomas, and 1.44 (1.01-1.87) for osteosarcoma. Geometric mean doses in bone were about 26 mRad.

CONCLUSIONS

An association was found between the presence of radium in birthplace water supplies and increased risk of bone sarcoma in two studies. Increased risk was present for lifetime measures of exposure, but the association was not significant, and there was no dose-response trend. Our findings are compatible with the absence of risk at low doses, but they might also reflect inadequate statistical power to measure a true risk at environmental exposure levels. If the increased risk at environmental doses is causal, risk of bone sarcoma is effectively linearly related to dose over five orders of magnitude.