Health risks attributable to radiation

Draft document: Health risks attributable to radiation
Submitted by Yoshiya Shimada, National Institute of Radiological Sciences
Commenting on behalf of the organisation

2005 July 21 Comments on FD-C-1 NAKAJIMA T. SHIMADA Y. YOSHINAGA S. KANDA R. OHMACHI Y. FURUKAWA A. FD-C-1 is well written. The following are our comments. 1) The newly coined term of Òtissue reactionÓ should not be used and the term for effect has to be kept as the traditional Òdeterministic effectÓ (2.6). Tissue reaction and deterministic effect were used in the past as two separate terms to describe radiation effects in addition to genetic and carcinogenic effects. Although tissue reaction and deterministic effect overlap each other for some cases, the former is more general term than the latter which is used more strictly for risk assessment purpose. Thus, the use of the new term is likely to create confusions which should be avoided. 2) It may be true to judge that there are threshold doses for induction of malformation and severe mental retardation for fetuses, likely around 100 mSv and 300 mSv, respectively (3.2. p.28). Although a non-threshold dose response cannot be excluded for IQ losses, FD-C1 concludes that the effects of a few tens of mGy would be undetectable and therefore of no practical significance. To avoid the misunderstanding that these is also a dose-threshold for induction of cancer for in utero exposures, additional sentence is needed stating that LNT should not be applied for cancer induction in embryo and fetus (See 4.4.3). 3) The discrepancy in the nominal risks, detriment and tissue weighting factors between FD-C-1 and the ICRP draft for 2005 Recommendations should be dissolved. For example, nominal risk coefficient for breast cancer for whole population is 69 in Table 4.1 in FD-1 whereas it is 121 in Table A.1 of the draft. Tissue weighting factor for breast is 0.08 in FD-1 while it is 0.12 in the ICRP draft for 2005 Recommendations. 4) Tissue weighting factors for remainder tissues should be shown in a separate category in Table 4.3 so that readers can see them easily. 5) The use of the term Òdetriment adjusted probability coefficientsÓ (lines 212, 2087, and elsewhere) has to be avoided and changed to much simpler term of ÒdetrimentÓ. The detriment is a complicated mixture of a probability coefficient of cancer/genetic effect, fatality, loss of life expectancy, etc. Thus, Òdetriment adjusted probability coefficientÓ is repetitious. 6) Sentences for or against LNT/existence of threshold are intermixed in Chapter 4.4.5 which is likely to create confusion. The paragraph would better start by mentioning the possibility of threshold based on the biological mechanism and epidemiological data and then move onto the LNT, which the ICRP judges as a better model at this moment. This issue is not resolved and remains to be discussed. Therefore, the title may be ÒThe possibility (or arguments) of a low dose threshold for cancer riskÓ.