This report represents comments from the Health Protection Agency Radiation Protection Division (HPA-RPD). HPA-RPD welcomes the release of the draft document for consultation and recognises the significant work that has been undertaken. HPA-RPD is broadly supportive of the major conclusions and judgements made. This report suggests a number of minor areas for clarification or revision. CONTENTS 1 Introduction 2 General comments 3 Foundation document draft - principal conclusions and proposals of the task group 4 Foundation document draft section 1 - introduction 5 Foundation document draft section 2 - interactions of radiation with cells and tissues 6 Foundation document draft section 3 - risks of tissue injury 7 Foundation document section 4 - risks of radiation-induced cancer 8 Foundation document section 5 - non-cancer diseases after radiation exposure 9 Foundation document section 6 - risks of heritable disease 10 Foundation document section 7 - summary of principal conclusions and proposals 11 References 1 INTRODUCTION In April 2005 the International Committee on Radiological Protection (ICRP) published on their website a draft of the Committee 1 Foundation Document, Biological and epidemiological information on health risks attributable to ionising radiation: a summary of judgements for the purposes of radiological protection of humans. This document is one of several that underpin the 2005 ICRP recommendations, a draft of which has already been subject to consultation through the ICRP website. NRPB provided comment on the draft recommendations and NRPB’s successor division within the Health Protection Agency (HPA), Radiation Protection Division (RPD) welcomes the opportunity to provide comment on this foundation document. This report has been produced primarily by the authors named, however all HPA-RPD departments were invited to contribute. Several HPA-RPD/NRPB staff contributed to the development of the foundation document. This report draws as far as possible on other staff not directly involved in writing the foundation document. All sections of the foundation document are commented upon both in a general way and, where appropriate, in greater detail. 2 GENERAL COMMENTS • HPA-RPD welcome the release of the draft Committee 1 Foundation Document and other foundation documents. These provide essential underpinning for the draft 2005 Recommendations and clarify some of the issues raised in the NRPB response to the draft recommendations. • The Foundation Document draws heavily on previous reviews, for example UNSCEAR (2001), NCRP (2001) and the draft ICRP report on low-dose extrapolation of radiation-related cancer risk. Full understanding of the foundation documents requires a good knowledge of the material contained in these and associated reviews. • The principal conclusions presented at the beginning and Table 7.1 at the end of the document provide a helpful guide to the major issues, judgements and, in the case of Table 7.1 data sources and methods. • An important new policy decision regarding gender averaging of risks is presented. More explanation of the reasoning behind this decision would be welcome, particularly as there are some clear cases of gender difference. • There is an apparent desire to reduce the importance of sources of epidemiological data other than the LSS, particularly for lung cancer and liver cancer. In our opinion these additional datasets have significant merit and are generally in good agreement with LSS data. • We note that sources of uncertainty in risk estimates and weighting factors are considered but rarely quantified. This is in our opinion appropriate for the general purposes of radiological protection but there may be situations in which a detailed consideration of uncertainties would be beneficial. • HPA-RPD is broadly supportive of the major conclusions and judgements made and recognises that significant work has gone into new calculations of detriment and germline risks. 3 FOUNDATION DOCUMENT DRAFT - PRINCIPAL CONCLUSIONS AND PROPOSALS OF THE TASK GROUP This section of the draft acts as a very useful summary of the main issues covered in the document. The second bullet point (lines 192-196) concerning ‘the effect of introducing the uncertain possibility……’ is not very clear, it could perhaps be expressed more simply. In bullet point 3 (lines 198-201), it might be more accurate to say that radiation weighting factors for protons and neutrons are discussed (rather than fully developed) in the Committee 2 Foundation Document. 4 FOUNDATION DOCUMENT DRAFT SECTION 1 - INTRODUCTION The introduction provides a useful context for the body of the document that follows. Some of the document summarises previous judgements while a number of new or revised judgements are also presented. Specific comments are given in Table 1. TABLE 1 Specific comments on draft section 1 Paragraph/line(s): Comment 1/276: Care needs to be taken in the use of ‘effective dose’ to ensure consistent use in all ICRP documents. 5 FOUNDATION DOCUMENT DRAFT SECTION 2 - INTERACTIONS OF RADIATION WITH CELLS AND TISSUES This section gives an overview of current knowledge on radiation biophysics and radiation biology relevant to the three major classes of health effects - cancer, hereditary diseases and tissue reactions. A wide range of topics is considered and consequently, there is heavy but appropriate dependence upon previous reviews rather than primary literature. However, additional references would be useful for the statements on lines 435-438, 486-493, 495-499 and 557. Specific points are given in Table 2. TABLE 2 Specific comments on draft section 2 Paragraph/line(s): Comment 2.1/319, 359: Two references are inconsistent between text and reference list. 2.3.1/409-412: The role of homologous recombinational repair (HRR) in fixing mutations is given little weight. However, there is good evidence for HRR factors in affecting cancer risk (at least for spontaneous cancer); for example, BRCA1 and BRCA2. 2.3.1/425-429: The role of cell cycle checkpoints in providing ‘opportunities for repair’ is not particularly well supported by experimental evidence, it is however often assumed. 2.3.2/461-464: Points a and b appear to contradict each other. The statement of ‘no evidence that adaptive responses… in a) appears contrary to the original observations of adaptive responses in human lymphocytes. 2.6/650-653: Sentences require re-drafting for clarity. 2.7/664-699: Many readers will be familiar with the terms ‘initiation’, ‘promotion’ and ‘progression’ in relation to carcinogenesis. It may be useful to mention these in relation to multistage carcinogenesis. 2.7/665: Suggest ‘…. on our understanding of the complex process….’ 6 FOUNDATION DOCUMENT DRAFT SECTION 3 - RISKS OF TISSUE INJURY Judgements relating to risks of tissue injury are covered in section 3. We note and welcome the change in terminology from ‘deterministic effects’ to ‘tissue injury/tissue reactions.’ The revisions of threshold doses for 1% incidence of tissue reactions are appropriate and draw on relevant literature. Presentation of the basis for selecting 15 mSv as the public dose limit for the lens of the eye would be useful. There is no real justification given (see specific comment below). TABLE 3 Specific comments on draft section 3 Paragraph/line(s): Comment 3.1.3/879: Suggest ‘….desquamatory reactions in epithelial tissues…..’ 3.1.3/880-882: Lymphocytes are depleted from circulation primarily by migration into tissues following radiation exposures rather than through apoptosis. 3.1.5/1061-1062: The true benefit of growth factor therapy can be questioned. Bone marrow recovery was noted but patients died from later arising organ reactions such as pneumonitis. If sentence is retained, suggest ‘However, the growth factors were considered to be of some benefit.’ 3.1.5/1102-1105: Suggest rewording of the definition of RBE: The RBE of a high versus low LET radiation is defined as the ratio of absorbed doses from a reference low LET radiation and the high LET radiation that result in the same level of biological effect. 3.1.7/1165-1168: With a threshold of about 1.5 Gy for induction of cataracts, a dose limit of 150 mSv (occupational) seems sufficiently conservative for general application and it is not clear why another factor of ten reduction should be applied to specify a dose limit of 15 mSv for the public. Can some additional justification be provided? Table 3.2/1228-1237: The use of >1.0 as DMF values can be confusing. Should this be taken to indicate a qualitative effect that has not been quantitated? If so, the use of >1.0 gives a false sense of a defined quantified effect. 7 FOUNDATION DOCUMENT SECTION 4 - RISKS OF RADIATION INDUCED CANCER This section provides some very important background to the draft 2005 recommendations and covers the use of a new tissue weighting scheme and methods for calculating radiation detriment. The move to calculating cancer risks on the basis of incidence is appropriate and welcomed. Sufficient detail is given on the methods for calculating detriment and the highlighting of uncertainties provides readers with a good feel for the robustness of the calculations. Section 4.4.3 considering prenatal (in utero) irradiation is based on the Executive Summary of Publication 90. Reading this section of the draft foundation document in isolation does not appear to provide a balanced summary of the data and their implications. There are three additional areas where more clarity would be welcome. Firstly, there appears to be a desire to dismiss sources of epidemiological data other than the LSS, particularly radon-induced lung cancer and Thorotrast-induced liver cancer, rather than to use these data to draw comparisons and support the LSS data. More detail on this point is given in Table 4. Secondly, it is considered that more justification of the Commission’s policy decision on gender averaging could be provided. Again more detail on this point can be found in Table 4. Finally, in considering risks to bone, ICRP 60 estimates are used. However, there is a problem with the bone estimate as it is based on Ra-244 data, calculated on the basis of average bone dose. There is a mis-match between this and biokinetic/dosimetric models which calculate radionuclide doses to the bone surface (currently a 10m endosteal layer). As pointed out by Puskin et al (1992), the risk estimate would be a factor of nine lower if calculated on the basis of bone surface dose. The situation is further complicated by current moves to widen the bone target region thus reducing doses. Some discussion of these additional factors would be beneficial. Further specific comments are given in Table 4. TABLE 4 Specific comments on draft section 4 Paragraph/line(s): Comment 4/1266-1272: Suggest restructure sentence to: ‘The conclusions reached by the task group on the implications of fundamental and animal data are used: i) to guide the projection of higher dose epidemiological data for the purposes of estimating cancer risk in the low dose region of interest and ii) to consider the application of dose and dose-rate effectiveness factor (DDREF) …and low dose rates.’ 4.1.2/1354-1355: Suggest delete ‘to be’ from sentence. 4.1.2/1361: Suggest delete ‘…therefore…’ 4.1.3/1409: Suggest ‘…published report of the Committee Examining Radiation Risks of Internal Emitters (CERRIE 2004).’ 4.2/1430-1431: Suggest ‘…a DDREF value of about 2 is implied’ rather then current wording ‘…value of 2 or less.’ 4.4/1468-1469: Suggest ‘…recommendations on the transport of risk between populations, the estimation of radiation detriment and the derivation of tissue weighting factors.’ 4.4.1/1475-1476: ‘… averaging gender and age at exposure-specific lifetime risk estimates…’ Can this be redrafted for greater clarity? 4.4.1/1478-1481: It would be useful to mention that the alternative approaches referred to in the sentence ‘Because of the uncertainty… from alternative models.’ are discussed in 4.4.1.1. 4.4.1/1490: Suggest delete ‘generally’. 4.4.1/1492-1498: This is a rather confusing description of how effective dose is obtained as the sum of weighted equivalent doses to organs/tissues. ‘… in some useful partition of the human body..’ appears superfluous. ‘…. dose equivalents…’ = equivalent doses. 4.4.1/1498-1500: The last sentence says: ‘The components of detriment are essentially the same for cancer and hereditary disease and, if desired, these detriments may be combined’. However, page 99: ‘In a strict sense, genetic risk coefficients cannot be compared or combined with those for cancers’. This inconsistency should be addressed. 4.4.1/1502: Suggest replace ‘For generality…’ with ‘In general…’ 4.4.1.1/1543-1545: The sentence ‘These effects were …goodness of fit’ could be clarified by adding ‘... when modelling cause-specific cancer types.’ 4.4.1.1/1597-1598: Suggest ‘…in the course…’ 4.4.1.1/1619-1638: Relative weight given to non-LSS data. a) Are there really significant differences between lung cancer risk estimates based on the LSS and radon-exposed miners? In our opinion, not when uncertainties in dose estimates, RBE, etc are taken into account. Suggest that ‘reassuringly close’ rather than ‘significantly different’ is more appropriate. b) The Publication 60 liver cancer risk was based on Thorotrast data with no caution regarding their application to low LET radiations. Why are generalisations to low-LET radiations now regarded as problematic? Applying an RBE of 20, risk estimates based on the LSS and on Thorotrast exposed groups are remarkably similar (see, for example Harrison and Muirhead 2003), although there are, of course, substantial uncertainties in dose estimates for Thorotrast. Nevertheless, the similarity of these (published) risk estimates might be seen to weaken the case for dividing the LSS based estimate by two on the basis of a reported interaction between hepatitis and radiation in the LSS. 4.4.1.2/1758, 1765 and 1766: Is the descriptor ‘dose-specific’ necessary? 4.4.1.2/1779-1786: ‘In accordance with current ICRP procedures, intermediate and final numerical risk estimates presented here are gender-averaged’. This is presented here and elsewhere (4.4.1.4) as a policy decision but no explanation is given on this important issue. Doses will soon be calculated separately for males and females using voxel phantoms, before averaging. Given differences in doses and risks, it might be argued that these should be recognised in the system of radiation protection. Some explanation of why gender specific data are not considered appropriate for radiation protection would be useful. In addition, it seems clear there are specific circumstances when the use of gender-specific values would be useful. Mention could be made of the potential uses of gender-specific data. 4.4.1.2/1797-1799: ‘There is rough equivalence between dose-specific excess absolute risk (EARLSS) and the product of excess relative risk (ERRLSS) and the base-line rates for the population of Japan…’ Can the reasons for the equivalence being ‘rough’ be made clear? 4.4.1.2/1813-1817: This reference to the relative merits of multiplicative and additive models on the basis of the role of radiation and other initiators/promoters is very interesting but difficult to follow. Perhaps this is not the best place to expand on these ideas fully but they are not easily understandable from the single sentence given. Perhaps a modest amount of additional supporting evidence or explanation can be added. 4.4.1.2/1825 and 1834: Should ERRUS(1/2) be ERRUS(p=0.5)? 4.4.1.2/1852: Suggest replace ‘radiation-associated excess rate’ with EAR for consistency. 4.4.1.3/1939-1942: ‘… the lethality fraction for breast cancer has decreased in the past 15 years…but this appears to have little impact on the relative detriment estimates.’ Why not? A more specific statement would be welcome. 4.4.1.3/1994: In tables 4.1(a) and (b) detriment is the product of lethality adjusted nominal risk and relative cancer free life lost. However, the values in the table do not equate exactly. This is due to rounding of values - it might be useful to mention this in a footnote. 4.4.1.3/2000-2001: Lethality is shown as ‘q’ here but is represented by ‘k’ in the main body of the text. Suggest replace ‘k’ for ‘q’ here. 4.4.1.4/2044-2047: The scheme referred to here is that in Publication 26 rather than Publication 60. In Publication 60 the remainder was specified as the mass weighted mean of a list of ten tissues. A ‘splitting rule’ is applied so that if any of the remainder tissues receive a committed equivalent dose that exceeds corresponding doses to all named tissues, it is singly given a weighting of 0.025, ie half of the total of 0.05 for Remainder tissues/ organs. Doses are then non-additive. 4.4.1.4/2047: Suggest ‘… which receive the highest doses ie, a non-additive system.’ 4.4.1.4/2049-2051: Suggest clarify sentence: ‘The principal reason for not mass weighting Remainder is that it is then dominated by doses to organs/tissues with the greatest masses, particularly muscle.’ 4.4.1.4/2062-2065: Footnotes to Table 4.3. The suggestion is that ‘mouth’ should be removed from the ET region of the HRTM (it was listed as part of this region but not included as a source or target) and include ‘Oral mucosa’ as a Remainder tissue as part of the HATM. The approach is then consistent. 4.4.2/2101: Is the reference to Table 4.2a correct? Should it be 4.1a? 4.4.3/2122 onwards: This paragraph questions the findings of the OSCC study but does not discuss the main reservation - that the risks of solid cancer after irradiation in late fetal life from this study are very different from those for irradiation in early infancy in the LSS. 4.4.3/2124-2127: This sentence refers to the second largest study of prenatal x-ray exposure which was much smaller and, given uncertainties, was consistent with the results of the OSCC. It also refers to cohort studies, of which only one has sufficient power to conclude anything - Court-Brown et al - and the reliability of its results are questioned by one of the authors (Doll). 4.4.3/2127-2129: This sentence, referring to similar life-time risks of cancer after irradiation either in utero or in early childhood does not make clear whether the comparison includes the risk of childhood cancer after in utero irradiation. 4.4.3/2131-2132: The OSCC data suggest that there is some risk associated with 1st trimester irradiation but, given the likelihood that doses during this period were greater, due to the use of different procedures, no firm conclusion can be reached regarding differences in sensitivity between trimesters. 4.4.3/2132-2134: While it is probably not possible or desirable to produce wTs specifically for in utero irradiation, it should be possible to discuss implications of current data for the ICRP scheme. This applies also to overall risks of in utero irradiation (final paragraph of section), ie that risks to the foetus are a few times greater than to the population as a whole. The same, most likely, applies to young children as well. 4.4.3/2134-2137: The last sentence raises issues of DDREF and RBE that could be presented in a more positive way. 4.4.5/2209-2218: The message on potential involvement of adaptive responses, genomic instability and bystander signalling is more positive than given in section 2 and the principal conclusions. 4.4.5/2222-2223: Some of sentence can be deleted, it is repetitious. 4.4.5/2268: Substitute ‘prove’ for ‘provide’. 8 FOUNDATION DOCUMENT SECTION 5 - NON-CANCER DISEASES AFTER RADIATION EXPOSURE The judgement to exclude emerging data on potential risks of non-cancer disease following radiation exposure is sound. It will nonetheless be important to keep new data under review as effects at a population level could be significant. There are no specific comments on this section. However the references used do not appear in the reference list. 9 FOUNDATION DOCUMENT SECTION 6 - RISKS OF HERITABLE DISEASE The section on risks of heritable disease is the longest and most detailed. It is a technically complex area and many assumptions have to be made and justified. We note two major changes: i) the projection of risks of heritable disease over two generations, ii) revision of background incidences of heritable disease. While there remains some uncertainty on the first change, on the basis of current evidence, it is not unreasonable. The use of revised baseline frequencies of heritable disease makes good use of new data/estimates. The level of detail provided in this section is significantly greater than in others. However, sufficient collation and distillation of key points is given. Specific comments on this section are compiled in Table 5. TABLE 5 Specific comments on draft section 6 Paragraph/line(s): Comment 6.3.2/2566: Suggest ‘…entirely upon mouse data…’ 6.3.2/2582: Suggest ‘These factors were considered…’ 6.3.4/2924: Suggest ‘(or are induced)’ 6.3.4/2933: Suggest ‘mutation rates’ 6.3.4/2956-2957: Suggest ‘taking into account gene…’ 6.4.1: It should be made clear that when quoted risks are in terms of per million live births. This is not apparent until Table 6.3 is read in detail. 6.5.1/3281: Suggest ‘…and also incorporated…’ ie substitute ‘also’ for ‘besides’. 10 FOUNDATION DOCUMENT SECTION 7 - SUMMARY OF PRINCIPAL CONCLUSIONS AND PROPOSALS This section provides a very useful summary with reference to major sources of information drawn upon and methods used. Two specific comments are relevant in this section. TABLE 7 Specific comments on draft section 7 Paragraph/line(s): Comment 7/3470-3472: ‘The conclusions and proposals…..are for broad purposes of prospective planning in radiological protection.’ There is scope for greater clarity on this issue in FD-C-2 (referred to in this paragraph). Most of the conclusions and proposals also apply to retrospective consideration of doses and risks. Table 7.1: Point 7, conclusions, suggest ‘First and second generation…’ 11 REFERENCES Harrison JD and Muirhead CR (2003). Quantitative comparisons of cancer induction in humans by internally deposited radionuclides and external radiation. Int J Radiat Biol 79: 1-13. ICRP Publication LDR-C-1. Low dose extrapolation of radiation-related cancer risk. Awaiting publication in Annals of the ICRP following website consultation. NCRP (2001). National Council on Radiation Protection and Measurements. Evaluation of the Linear-Non-threshold Dose-Response Model for Ionizing Radiation. NCRP Report No. 36. National Council on Radiation Protection and Measurements, Bethesda MD. Puskin JS, Nelson NS and Nelson CB (1992). Bone cancer risk estimation. Health Physics 63: 579-580. UNSCEAR (2001). United Nations Scientific Committee on the Effects of Atomic Radiation. Hereditary Effects of Radiation. UNSCEAR Report to the General Assembly with Scientific Annex, United Nations, New York.