The linear no-threshold response: why not linearity?

Abstract

ICRP and NCRP recommend risk coefficients for use in radiation protection that are based on a linear quadratic response in the low dose region. This is a derivative of the linear no threshold (LNT) hypothesis with allowance for low dose and dose rate effects. The risk coefficients are derived from the Lifespan Study of the A-bomb survivors but are supported by many other epidemiological studies some, such as occupational, at low doses. Nevertheless, the risk coefficients are uncertain and range (90% confidence intervals) over a factor of 2-3 above and below the nominal values. Various possible dose responses in the low dose region are considered including those that may result from adaptive responses. Laboratory studies show linearity in some systems to doses as low as 2.5 mGy. Epidemiological studies include several with significant excess risks at 100 mGy or less with at least one at 10 mGy. The linear quadratic response seems, therefore, the most likely response in the very low dose region. Adopting the linear quadratic response in the low dose region does not prevent common sense judgements about dismissing small radiation risks. NCRP defined first a negligible individual risk (1987) and then an individual dose (1993) to encourage common sense judgements in the low dose region. More consideration might be given to dismissing minor risks in common sense applications in radiation protection.