Brenner D J, Miller R C, Huang Y, Hall E J
Center for Radiological Research, Columbia University, New York, New York 10032.
Radiat Res. 1995 Apr;142(1):61-9.
Domestic radon risk estimates are typically based either on data for uranium miners or on data derived from A-bomb survivors; comparison of domestic radon risk estimates derived from these two disparate sources represents an important test of their reliability. There is currently a significant discrepancy of about a factor of three between domestic radon risk estimates generated with these two independent methods. To base such risk estimates on the data for A-bomb survivors, who were exposed mainly to low-LET radiation, requires a quality factor for alpha particles from random progeny; the final risk estimate is then directly proportional to this quality factor. We have used the most extensive quantitative in vitro data set currently available at high LET for an oncogenic end point, to make the best estimate we can that could be used as a basis for a quality factor. Our best estimates of values appropriate for the quality factor for radom progeny are significantly lower than those currently used (20-25) in estimating lung cancer mortality due to randon. Specifically, our best estimate for home dwellers is around 10. In addition, because of the different geometry in the bronchial epithelia of nonsmokers compared to smokers, our best estimate of an appropriate quality factor for home dwellers is about 18% greater than that for miners; thus our best estimate of the "effective K factor" to convert to effective dose/WLM in home dwellers from effective dose/WLM in miners would be increased by this factor. Based on a quality factor of approximately 10, the dosimetrically based estimate of radon-induced mortality would be approximately 35,000 per year in the U.S. rather than the value of approximately 70,000 obtained using a quality factor of 20. The value of 35,000, while larger than the values based on data for miners (approximately 20,000), is much smaller than previous estimates of approximately 70,000 based on dosimetric methods; thus risk estimates based on the two approaches, dosimetric and epidemiological, may be partially reconciled. Finally, a quality factor of 10 would reduce the proportion of the collective effective dose caused by radon progeny from the currently accepted value of 55% down to about 38%.
