COC Meeting 1st March 2007
Minutes
Present:
| Chairman: | Members: | Secretariat: |
| Professor D Phillips |
Dr C Allen | Ms F Pollitt (Scientific - HPA) |
| Dr P Carthew | Dr D Benford (Scientific - FSA) | |
| Professor P Farmer | Dr L Hetherington (HPA - Minutes) | |
| Mrs R Glazebrook | Mr S Robjohns (HPA - Minutes) | |
| Professor D Harrison | Ms J Cleverly (HPA - Administrative) | |
| Ms D Howel | ||
| Dr B Miller | ||
| Professor R Roberts | ||
| Professor D Shuker | ||
| Professor P Vineis | ||
| Dr N Wallis |
| In Attendance: | |
| Dr P Edwards |
(HPA) |
| Dr D McElvenny | (HSE, item 7) |
| Mrs M Meldrum | (HSE, item 5) |
| Mr K Mistry | (DH) |
| Mr K Okona-Mensah | (DH Tox unit, item 9) |
| Dr L Rushton | (Imperial College, item 7) |
| Assessors: | Observers: |
| Dr D Andrew (PSD) | Hong JeongMi (Korean FDA) |
| Mr A Browning (VMD) | Lee Seon-Hwa (Korean FDA) |
| Mr M Hosford (EA) | Lee Jeongho (Korean FDA) |
| Dr H Stemplewski (MHRA) | Mr A Ward (NICE) |
| Contents | Paragraph(s) | |
| Item 1: | Apologies/announcements | 1-3 |
| Item 2: | Minutes of meeting held on 16 November 2006 (CC/MIN/2006/3) | 4 |
| Item 3: | Matters arising not covered by later agenda items |
5 |
| Item 4: | Draft Working Paper on prostate cancer and pesticide exposure (CC/07/3) |
6 - 8 |
| Item 5: | Technical guidance for derivation of DNELs and risk Characterisation of non-threshold effects in the context of REACH (CC/07/4) |
9 - 17 |
| Item 6: | Assessing the risks of the acute or short-term exposure |
18 - 26 |
| Item 7: | HSE programme of work on occupational carcinogens: Project to update epidemiological intelligence: progress report (CC/07/5) |
27 |
| Item 8: | Non-Hodgkin's lymphoma - overview paper (CC/07/7) | 28 |
| Item 9: | Risk communication: Comments from Dr Fischer (CC/07/2) | 29 - 31 |
| Item 10: | COC Annual Report for 2006 |
32 - 33 |
| Item 11: | Any other business | 34 |
| Item 12: | Date of next meeting |
35 |
ITEM 1: APOLOGIES FOR ABSENCE/ANNOUNCEMENTS
1. Apologies for absence were received from Professor A Boobis, Mr J Battershill (HPA secretariat), and the assessors Mr S Samuels (PSD), Dr B Viegas (Defra), and Dr S Dyer (DH). It was noted that written comments on the papers had been received from Professor Boobis.
Announcements
2. The Chairman welcomed Mrs M Meldrum (HSE), Dr D McElvenny (HSE), Mr K Okona-Mensah (DH Tox Unit), and the following observers: Hong JeongMi, Lee Seon-Hwa and Lee Jeongho (all Korean FDA) and Mr A Ward (NICE).
3. Members were reminded of the need to declare any relevant interests before discussion of items.
ITEM 2: MINUTES OF THE MEETING OF 16 NOVEMBER 2006 (CC/MIN/2006/3)
4. The minutes were agreed subject to minor editorial changes.
ITEM 3: MATTERS ARISING NOT COVERED BY LATER AGENDA ITEMS
5. There were no matters arising.
ITEM 4: DRAFT WORKING PAPER ON PROSTATE CANCER AND PESTICIDE EXPOSURE (CC/07/3)
6. The secretariat reminded members that, at the last meeting, they had considered a Defra-funded narrative review by the Institute of Occupational Medicine of the available studies on prostate cancer and pesticide exposure, and a recent meta-analysis of the risk of prostate cancer in pesticide manufacturing workers. The committee had decided to summarise its conclusions on these papers in a short supplementary statement to its 2004 statement on prostate cancer. CC/07/3 contained a draft for comment.
7. The following comments were made:
- The statement should be redrafted to indicate that it was the committee, and not current members, who published the statement in 2004.
- In paragraph 3, it should be made clear that there is a large number of studies of agricultural workers exposed to pesticides and that the results have been inconsistent ie these are two separate points.
- The section referring to the androgen imbalance hypothesis should be redrafted. A member offered to provide a suitable reference.
- The section in brackets at the end of paragraph 3 should be deleted.
8. A number of other minor drafting comments were made. The final statement would be agreed by Chairman's action.
ITEM 5: TECHNICAL GUIDANCE FOR DERIVATION OF DNELs AND RISK CHARACTERISATION OF NON-THRESHOLD EFFECTS IN THE CONTEXT OF REACH
9. Dr Carthew declared a personal, non-specific interest.
10. The Health and Safety Executive (HSE) made a short presentation to introduce this item. Members were informed that the item was on the agenda to inform members about progress with the development of technical guidance for the risk assessment of substances under REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals - a new European Regulation). The COT had discussed this issue at its recent meeting and some COC members were present at that discussion.
11. In contrast to the broad overview presented to the COT, the paper for the COC focused on the proposed guidance for dealing with "non-threshold" genotoxic carcinogens. Two approaches are proposed: one is based on linear extrapolation from animal bioassay data to a predetermined low level of risk and the other is based on application of a large uncertainty factor (UF) to a suitable reference point on the dose-response for carcinogenicity. The latter approach was included because the UK, unlike most other EU member states, does not use the linear extrapolation approach. Both approaches generate a Derived Minimum Effect Level (DMEL) which is defined as the level of exposure above which humans should not be exposed. Members were informed that both the predetermined low level of risk and the uncertainty factors were decided at policy level and were non-negotiable.
12. Members considered that the UF and DMEL methods are essentially equivalent but differ with respect to risk communication. Members noted that the UF approach differed from the MOE approach in that, in the REACH methodology, a predetermined UF of 10,000 was to be used. Members were concerned about the derivation of the UF of 10 for 'the nature of the carcinogenic process' given the limitations of current knowledge about DNA repair.
13. There was discussion of the units chosen for the DMEL values with some members favouring the use of moles/micromoles rather than dose/weight/day, because the biological response is based on the interaction of molecules of a chemical, not the weight administered. However, it was pointed out that this could be difficult since the purity of the active ingredient is not always known and it would be difficult to assess substances, as opposed to single chemicals, on a molar basis.
14. It was noted that, in the draft Reference Preliminary Technical Guidance Document (pp76), it is proposed that "biological criteria" should be used to assess the relevance of a carcinogenic response, rather than statistics alone. Concern was expressed as to how an assessor can judge that a non-significant change in tumour incidence is biologically significant. All tumours are biologically significant. It is important to know whether they are compound related and this is, in part, based on statistical analysis. It was considered that one of the advantages of the BMD approach is that it does not require a categorical conclusion as to whether the response in a particular dose group is "real" or not.
15. It was noted that EFSA is undertaking further assessment of the MOE approach and that ILSI is developing MOEs for 12 compounds, with results due to be published in autumn 2008. The committee was also informed that an EFSA opinion on aflatoxins to be published shortly used three risk characterisation approaches. This would be circulated to committee members.
16. A member commented on the Threshold of Toxicological Concern (TTC) approach to setting DMELs for somatic cell mutagens with no cancer data. He considered that the TTC approach, with the exclusion of those groups of compounds suggested by JECFA and others (dioxins, nitrosamines, etc), was a pragmatic approach for prioritisation of chemicals for which there are only limited toxicological data and that it could be applied to both genotoxic carcinogens and to mutagens. In his opinion, the published analysis, using, for example, the Gold database, has shown that a TTC of 0.15 micrograms per day is adequately protective (at an assumed minimum risk level of 10-6) even for genotoxic compounds for all but the exceptions noted in the papers. However, it was noted that the COM considered it premature at present to set levels such as TTCs for mutagens with no cancer data, as no agreement had been reached about how to rank such mutagens.
17. HSE stated that the comments from the committee would be reported to the EU working group drawing up this guidance. The Chairman asked that the COC be kept informed of developments.
ITEM 6: ASSESSING THE RISKS OF ACUTE OR SHORT-TERM EXPOSURE TO CARCINOGENS (CC/07/1)
18. The Department of Health and the Health Protection Agency are occasionally asked to provide advice on the health risks of acute or short-term exposure to genotoxic carcinogens and have sought the advice of the COC on how this might be done. At the last meeting members decided that it would not be appropriate to use an acute T25 approach. One member had also informed the COC that there were some papers in the literature which might indicate a way forward. Three suggested papers were made available to the committee.
Halmes et al 2000
19. Halmes et al (2000) stated that conventional risk assessments are generally predicated on the assumption that cancer risk increases as a function of the cumulative carcinogen dose. For exposure to a carcinogen at a given rate, this would mean that the excess cancer risk is a function of the duration of exposure. The authors tested this assumption by comparing the tumour response in NTP stop-exposure studies with that in a standard 2 year study on the same chemical. In many cases, the response in the stop exposure study was greater than the response that would be predicted from the 2 year study. The authors concluded that, in those cases where it was possible to calculate equivalent averaging times for tumour/sites, the results suggested that short-term exposures (ie 13 to 66 weeks) were generally more effective in producing tumours than continuous long-term studies would predict.
Bos et al 2004
20. This was mainly a theoretical paper, which considered whether short-term exposure (1 - 10 days) to genotoxic carcinogens may contribute to tumour development and, if so, whether this contribution to cancer risk could be quantified. A pragmatic approach was proposed, which used the premise that tumour incidence is linearly related to the cumulative dose of a chemical and incorporated the principle of the Virtually Safe Dose (VSD) associated with an "acceptable" risk level. The approach then applied factors to scale up from low level exposure daily over 70 years to the dose which might be acceptable if exposure was only for 1 day, or 2-10 days.
Murdoch et al 1992
21. This paper considered approaches to estimating the lifetime risk associated with intermittent or time-dependent exposure to carcinogenic substances. The lifetime equivalent daily dose (LEDD, the received dose divided equally over a lifetime) has been used in the USA to estimate the risk associated with short-term exposure. Murdoch et al (1992) used a mathematical approach to generate a lifetime equivalent constant dose (LECD), which gave the same lifetime risk as the actual time dependent exposure pattern. They state that a comparison between the LECD and the LEDD gives a measure of the accuracy of risk estimates based on the LEDD and a measure of correcting such estimates. In some circumstances, use of a lifetime average daily dose would underestimate cancer risk by 2-5 fold. However, the authors concluded that it is possible to place plausible upper bounds on the error in estimates of risk based on the LEDD.
22. The COC considered the above papers. Members were unhappy with the concept there was a simple linear relationship between duration of exposure and cancer risk from genotoxic carcinogens for the following reasons: DNA repair processes could be significant at low doses, a non-linear response could occur due to the complexity of carcinogenic process, and genotoxic carcinogens may have different effects eg at high doses some genotoxic carcinogens could also promote cancer via a cytotoxic mechanism. The relationship could also be affected by latency. It was assumed that this concept had derived from observations about the relationship between radiation dose and cancer risk. It was noted that there were few relevant epidemiological data on chemicals but it appeared that, for smoking, the risk was sublinear with time, at least for lung and bladder cancer.
23. A member commented that the papers had been selected with a view to assessing the risks of short-term exposure (eg for several months) to a low dose of a carcinogen. However, the data in Halmes et al (2000) related to high dose studies, so it was unclear whether the results could be extrapolated to low doses, when a different mechanism of action may apply. The overall conclusion from the papers appeared to be that, at high doses, the relationship of risk with time was supralinear, ie a short-term exposure gave a higher risk than would be anticipated from long-term dosing, but it was not clear if this relationship held at low doses. Some members proposed that it could be assumed that this was the case. However, another member considered that the available data at low doses strongly indicated linearity.
24. As regards the individual papers, Members commented that, in the study by Halmes et al (2000), it was unlikely that the data from stop-exposure studies of at least 13 weeks duration could be extrapolated to the exposure durations of concern (<10 days). Members also noted that there were some problems with the analysis conducted by Halmes et al, such as the use of tumour responses from some stop exposure studies that were not considered significant in the long-term NTP studies. Overall, members agreed that this study would not aid the risk assessment of short-term exposures of less than 10 days to genotoxic carcinogens.
25. Members considered that the paper by Murdoch et al (1992) was a theoretical approach that involved a number of assumptions which it had not been possible to test. Overall, it was not considered that this provided a useful approach.
26. The COC noted that the Bos et al (2004) paper was mainly theoretical and used mathematical modelling to develop risk estimates from animal carcinogenicity data (an approach that is not recommended by the COC). Although there were areas of uncertainty and a number of assumptions in the proposed approach, members considered that extrapolation from lifetime exposures to short-term (up to 10 days) exposures for low doses would probably overestimate the cancer risk. Members considered that although there was an assumption that certain subpopulations would be more susceptible than others, there was no suitable data to allow any quantitative estimation and associated adjustment to cancer risk estimates. However, it was suggested that it may be possible to adapt the method by using the MOE approach and that this might provide a pragmatic approach to the risk assessment of short-term exposures to genotoxic carcinogens, although there would be some associated degree of uncertainty.
ITEM 7: HSE PROGRAMME OF WORK ON OCCUPATIONAL CARCINOGENS:
PROJECT TO UPDATE EPIDEMIOLOGICAL INTELLIGENCE: PROGRESS REPORT (CC/07/5)
27. The minutes of this item will be published when the work is completed and formally published.
ITEM 8: NON-HODGKIN'S LYMPHOMA - OVERVIEW PAPER (CC/07/7)
28. This item was postponed to the next meeting.
ITEM 9: RISK COMMUNICATION: COMMENTS FROM DR FISHER (CC/07/6)
29. At the request of the committee, Dr Lyn Frewer of the Social Sciences Department of the Marketing and Consumer Behaviour Group at Wageningen University, Netherlands, had been invited to attend the March meeting of the COC to discuss the committee's project on risk assessment and the proposal for the application of the Margin of Exposure (MOE) approach for communicating the risks of exposure to genotoxic carcinogens (discussed at the last two meetings). Unfortunately, Dr Frewer was unable to attend the meeting, but a colleague, Dr Arnout Fisher, had sent comments and papers to the COC.
30. Members thanked Dr Fisher for his assistance. The COC agreed that it was important to consider the communication of uncertainty associated with carcinogenic risk assessment eg it would be useful to provide some indication of whether there was likely to be an overestimate or underestimate of the carcinogenic risks or whether this was unknown.
31. It was agreed that lay statements were helpful and should be used when possible. Members considered that it may be possible to provide some perspective on the cancer risks and confirmed their support of a 'Margin of Exposure' approach. They also agreed that it may be useful to contact organisations experienced in communicating health risks to the public, to seek their advice on cancer risk communication, such as Cancer Research UK. However, overall, members did not consider that they could take this topic much further.
ITEM 10: COC ANNUAL REPORT 2006 (CC/07/2)
32. Members asked for the following changes to the draft report:
- The preface should include the actual number of years that Professor Blain served as chairman.
- The section in paragraph 3.3 re newborn animals should be made clearer.
- The end of paragraph 3.9 to be altered to: "Most of the agents classified by IARC as Group 1 and 2A human carcinogens were positive in animal studies".
- The last sentence of paragraph 3.10 to be deleted.
- The end of paragraph 3.27 to be altered to: "The committee also noted the increasing literature on the importance of epithelial-mesenchymal interactions in causing and sustaining tumours and of epigenetic changes that may precede or accompany mutation".
- The section on horizon scanning to be moved to the end.
33. Subject to these and a few minor editorial changes the report was agreed.
ITEM 11: ANY OTHER BUSINESS
34. There was no other business.
ITEM 12: DATE OF NEXT MEETING
35. 12 July 2007