Comments on the Draft 2005 Recommendations of the International Commission on Radiological Protection Prepared by the US Environmental Protection Agency/Office of Radiation and Indoor Air (with input from staff in EPA’s regional radiation program offices) The U.S. Environmental Protection Agency (EPA)/Office of Radiation and Indoor Air (ORIA) is pleased to offer comments and suggestions to the International Commission on Radiological Protection (ICRP) concerning the Draft 2005 Recommendations. ORIA commends the ICRP and its chairman, Prof. Roger Clarke, for the public consultation process that is being followed in the development of these new recommendations. Since we now understand that additional foundation documents are to be made available in 2005 that will lay the foundation for the proposed new recommendations, we look forward to also reviewing them and following the continued evolution of the current document beyond the now-postponed 2005 publication date. ORIA is pleased that the new recommendations stress the important role of stakeholders in implementing a national system of radiation protection. While strongly endorsing this approach, we also would emphasize that the principle of justification can still play an important role in the decision-making process for radiation protection professionals. It is true that justification of new practices involves many more stakeholders than just the radiation protection community, but on a more routine level, justification can be a useful concept in operational health physics distinct from optimization (though the two processes may overlap). Regarding the structure of the proposed recommendations, it seems that its purpose changes at various points. In parts, it reads like a high level set of recommendations; in others, it is a detailed review of current scientific information (e.g., Section 3); and in Section 11, it reads more like an action plan for future work. A suggestion would be to transfer some of the more detailed information to the foundation documents and to greatly reduce the section on environmental protection pending future work in this area. A second observation is perhaps more of a question, and that is whether the new set of recommendations are meant to replace or to add to the recommendations found in Publication 60. It could be read either way. Addressing this issue in the next revision of the recommendations will contribute to the improved coherence that is an aim of the new proposals. Specific comments and suggestions follow: Quantities: ORIA supports ICRP’s efforts to distinguish between the radiation weighted dose and effective dose. For continuity and clarity, we would support retaining the use of “deterministic effect” to describe tissue reactions. The proposed tissue weighting factors are a logical simplification from the ones given in Publication 60, but for those authorities who have recently invested in the switchover from Publication 26 to Publication 60, additional modification of the tissue weighting factors is unlikely to be cost-effective at this time. Limits and constraints: As a regulatory agency, EPA is accustomed to the use of a limit as an enforceable quantity, whether for describing a source or a class of sources. The ICRP, on the other hand, defines a limit as something which is, in practice, unenforceable – viz. the individual dose received from the sum of all controllable sources under routine conditions. It would be to the regulatory community’s advantage if ICRP could restructure its terms to better coincide with their standard usage. By setting the maximum individual constraint for the public at the same level as the public dose limit, i.e. 1 millisievert (mSv), the confusion surrounding the question of what regulators should refer to as a legally enforceable quantity is compounded (limit, constraint, optimized constraint?). In addition, setting the individual limit and annual constraint at the same level appears to reinforce an assumption that no individual in society is likely to be exposed to more than one significant source of ionizing radiation in a year. If this is the basis for the two values being the same, then this supposition should be tested or at least better defended. We believe that individual sources of radioactivity outside of licensed facilities are more often than not easily controllable at levels well below 1 mSv, which suggests that constraints for individual sources would be better set at some fraction of the public dose limit. It may then lead one to conclude that the four dose values presented in the draft recommendations are insufficient for fully characterizing the range of constraints needed in the proposed system of radiation protection. On the other end of the recommended dose range, we have some concerns that the maximum constraint for emergency responders is too low at 100 mSv. For example, in the United States, an emergency worker may voluntarily elect to receive 250 mSv (or more) to save a life. When viewed as a likely one-time exposure, this protective action guideline falls well within the cumulative (lifetime) dose limit allowed for workers and is also below the level of concern for any serious deterministic effects. Exclusion of radiation sources: The Table of Recommended Exclusion Levels is largely based on values given in the IAEA report, DS 161 (now RS-G-1.7). For radionuclides of natural origin, RS-G-1.7 states that the value of 1 Bq/g was “selected on the basis of consideration of the upper end of the worldwide distribution of activity concentrations in soil provided by UNSCEAR.” It then states that “[d]oses to individuals as a consequence of these activity concentrations would be unlikely to exceed about 1 mSv in a year, excluding the contribution from the emanation of radon, which is dealt with separately in the BSS.” Given that the exclusion level for these radionuclides of natural origin (other than K-40) is based on a value corresponding to the proposed individual dose limit and maximum individual constraint, ORIA encourages the ICRP to revisit the rationale for recommending these values for exclusion. As an example of the problems this value could present to regulators, the 1 Bq/g value for Ra-226 is over five times higher than EPA’s cleanup level for radium in surface soil at uranium mill tailing sites. On the other hand, 1 Bq/g may be too low a value for some other naturally occurring radionuclides. In addition, there needs to be a clear distinction made between levels appropriate for the release of small quantities of solid material from regulatory control and levels at which residual radioactivity can remain at uniform concentration in the environment (i.e., cleanup levels). In this respect, it must be noted that the exemption/exclusion limits described by IAEA’s RS-G-1.7 are for “..‘moderate’ amounts of material and that for larger amounts additional consideration is necessary.” Even more importantly, the IAEA states on page 8 of the Safety Guide that: The examples of excluded types of exposure given in the BSS include exposure from “unmodified concentrations of radionuclides in most raw materials”. The reference to unmodified concentrations points to the fact that the processing of some raw materials, which may have typical concentrations of radionuclides of natural origin, may generate products or wastes that have higher concentrations of radionuclides or give rise to exposures that should not be excluded from regulatory control. The reference to exposure from most raw materials suggests that exposure from some raw materials should not be subject to exclusion. Thus, whichever the cause of the exposure — whether it results from the modification of the chemical or physical form of the material, thus enhancing its radionuclide content in processing, or simply because the material inherently has a relatively high radionuclide content — the regulatory body should recognize that there are some exposure situations that warrant consideration and control (e.g. exposure situations in industries in which material containing radionuclides of natural origin is handled or used and where exposure is attributable to its processing)... The misuse of materials containing naturally occurring radionuclides particularly from mine site wastes excluded and ultimately recycled for home construction, or construction of houses on recycled mine sites, may commonly give rise to individual exposures far exceeding public limits, even though the radium content of the soil or wastes might not exceed 1 Bq/g. Further clarification by ICRP on the consequences of such exposures to naturally occurring radionuclides at these lower concentration limits is merited, and exclusion from the recommendations is not advisable in our view. Radon: Table 8 state that the Commission regards the radon levels of 600 Bq m-3 (16.2 pCi/L) for homes and 1500 Bq m-3 (40.5 pCi/L) for the workplace as Maximum Constraints. Other statements affirm that national agencies will apply optimization to arrive at the most applicable level at which to act in their countries. However the last sentence in paragraph 180 is somewhat confusing in that it recommends that there should be no attempt to reduce public exposure further. This sentence needs to be clearer in specifying that it is the optimized level set by national authorities that is being referred to, not the value in Table 8. Environmental Protection: We concur that the concept of developing assessment endpoints is of value. It is a communication tool to both risk managers and the general public. In following this approach, it will be particularly useful to clearly state why a selected group of organisms is valuable. This then leads to the selection of measurement endpoints (or measures of effects) which in this case would be the reference animals and plants. As noted in the document, there are a number of receptor categories which are not presented. Specifically, we recommend that much of the details of the discussion on the approach to evaluating non-human receptors be deleted and deferred to future work of the Commission. We suggest that only a general acknowledgment of the issues remain. We do not believe that the approach presented will be functional and could easily result in conclusions which will later need to be reversed. The proposed approach to protecting the environment should be harmonized with the approaches taken for other environmental stressors. Effects on individuals may or may not affect the overall ecosystem. The current approach may not be helpful in assessing the total impact on the environment. We suggest not using multiples of background as the metric for protection of the environment. The use of background as a metric for protection of the environment can be a very complex undertaking. The small spatial range of some species would prohibit the use of a generic background. Each species would require a specific background determination due to proximity to the source, diet, and other exposure factors. Thus, the determination of background for a reference animal would be required at every location where protection of the environment is considered. The approach presented for protection of non-human receptors could lead one to interpret that a set of default or “reference” species models can be developed and from these a “safe” set of ambient activity levels can be established. Application of the standard ecological risk assessment process is also introduced as a possible tool within the development of these reference species. There are several challenges in the proposed method which must be overcome to produce a useable endpoint. In human health risk assessments, the assessment endpoint is defined -- it is humans. Additionally, the level of protection is established within a range for protection of individuals. Risk characterization leads to the establishment of acceptable exposures and by using conservative inputs, an acceptable concentration in the environment can be determined. This process is not this straightforward for non-human receptors. In particular, there could be confusion on the intended use of reference plants and animals. For example, selection of reference plant and animal models, by default, selects the assessment endpoints, but this somewhat arbitrary approach could lead to incorrect conclusions. We can say we are to be protective of populations and not individuals, but, when we do this, the individuals become the population. This is population ecology (deme) theory. With respect to the effects, again we do not have a measure. Relatively minor effects (repairable tissue damage) are likely to result in death or effect reproductive failure (nesting) through a variety of means. These could include reduced feeding, inability to avoid predators, and alteration in behaviors. This leads to the issue of what effect you put into the effects assessment for the model. You cannot know this unless you know the environmental setting and the effects possible for the species and system present and the level of protection which is desired. In the end the situation leads to only a couple of outcomes. If the approach taken is to use very conservative assumptions for the risk assessment (so that we are sure that significant adverse effects are unlikely under any environmental setting), the result will be extremely conservative exposure estimates. On the other hand, institutionally defining the assessment endpoints and by default the level of protection can be criticized as lacking in scientific rigor. (We note that paragraph 249 refers to heightened concern at exposures that are orders of magnitude greater than background whereas Figure B1 in the annex refers to concern at several times background. This discrepancy should be addressed.) The only other alternative is to define the process, which is the approach taken to date in EPA’s ecological risk assessment guidance. This is the approach that we recommend. One of the major challenges in the area of environmental risk assessment has been the development and use of the ecological exposure models. In the area of evaluating the effect of ionizing radiation, there is a greater possibility of developing standard reference models, as the mechanism of tissue effects is not radically different between organisms, as it is with different chemicals. However, the discussion presented leads one to believe that the ramifications of adverse effects observed in laboratory organisms will mirror those which would occur in the “wild”. While that may be true for large doses of radiation which would result in mortality, it may not be true for low dose effects. In the wild, organisms have set breeding periods, and do not have food and cover provided; any effect on their normal behavior is likely to result in the death of that organism due to directed predation, reduced feeding success or other causes not directly a result of the radiation exposure. This could be particularly true in eco-regions with ‘extreme” or harsh conditions, such as arctic conditions. Reduced feeding for even a short period of time during the summer can easily result in increased winter mortality. The result is that there appears to be a disconnect in the “effects assessment” since the adverse effects may be secondary to the direct effect of the ionizing radiation. In addition to external stressors, another factor which should be considered is whether experimental data collected from a species of pure bred laboratory animals (e.g., rats) can be extrapolated to the more genetically diverse populations of that species present in any ecosystem of concern. Pure bred strains could show more or less resistance to radiation than would be found in nature. In closing, ORIA again thanks the ICRP for the opportunity to participate in the development of its new recommendations.