Dose and dose rate extrapolation factors for malignant and non-malignant health endpoints after exposure to gamma and neutron radiation.

Murine experiments were conducted at the JANUS reactor in Argonne National Laboratory from 1970 to 1992 to study the effect of acute and protracted radiation dose from gamma rays and fission neutron whole body exposure. The present study reports the reanalysis of the JANUS data on 36,718 mice, of which 16,973 mice were irradiated with neutrons, 13,638 were irradiated with gamma rays, and 6107 were controls. Mice were mostly Mus musculus, but one experiment used Peromyscus leucopus. For both types of radiation exposure, a Cox proportional hazards model was used, using age as timescale, and stratifying on sex and experiment. The optimal model was one with linear and quadratic terms in cumulative lagged dose, with adjustments to both linear and quadratic dose terms for low-dose rate irradiation (<5 mGy/h) and with adjustments to the dose for age at exposure and sex. After gamma ray exposure there is significant non-linearity (generally with upward curvature) for all tumours, lymphoreticular, respiratory, connective tissue and gastrointestinal tumours, also for all non-tumour, other non-tumour, non-malignant pulmonary and non-malignant renal diseases (p < 0.001). Associated with this the low-dose extrapolation factor, measuring the overestimation in low-dose risk resulting from linear extrapolation is significantly elevated for lymphoreticular tumours 1.16 (95% CI 1.06, 1.31), elevated also for a number of non-malignant endpoints, specifically all non-tumour diseases, 1.63 (95% CI 1.43, 2.00), non-malignant pulmonary disease, 1.70 (95% CI 1.17, 2.76) and other non-tumour diseases, 1.47 (95% CI 1.29, 1.82). However, for a rather larger group of malignant endpoints the low-dose extrapolation factor is significantly less than 1 (implying downward curvature), with central estimates generally ranging from 0.2 to 0.8, in particular for tumours of the respiratory system, vasculature, ovary, kidney/urinary bladder and testis. For neutron exposure most endpoints, malignant and non-malignant, show downward curvature in the dose response, and for most endpoints this is statistically significant (p < 0.05). Associated with this, the low-dose extrapolation factor associated with neutron exposure is generally statistically significantly less than 1 for most malignant and non-malignant endpoints, with central estimates mostly in the range 0.1-0.9. In contrast to the situation at higher dose rates, there are statistically non-significant decreases of risk per unit dose at gamma dose rates of less than or equal to 5 mGy/h for most malignant endpoints, and generally non-significant increases in risk per unit dose at gamma dose rates ≤5 mGy/h for most non-malignant endpoints. Associated with this, the dose-rate extrapolation factor, the ratio of high dose-rate to low dose-rate (≤5 mGy/h) gamma dose response slopes, for many tumour sites is in the range 1.2-2.3, albeit not statistically significantly elevated from 1, while for most non-malignant endpoints the gamma dose-rate extrapolation factor is less than 1, with most estimates in the range 0.2-0.8. After neutron exposure there are non-significant indications of lower risk per unit dose at dose rates ≤5 mGy/h compared to higher dose rates for most malignant endpoints, and for all tumours (p = 0.001), and respiratory tumours (p = 0.007) this reduction is conventionally statistically significant; for most non-malignant outcomes risks per unit dose non-significantly increase at lower dose rates. Associated with this, the neutron dose-rate extrapolation factor is less than 1 for most malignant and non-malignant endpoints, in many cases statistically significantly so, with central estimates mostly in the range 0.0-0.2.