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	Mr J Battershill (HPA - Scientific)
	Mrs R Glazebrook	Dr L Hetherington (HPA - Minutes)
	Professor D Harrison	Mr S Robjohns (HPA - Minutes)
	Dr B Miller	Ms J Cleverly (HPA - Administrative)
	Professor R Roberts
	Professor D Shuker
	Professor P Vineis
	Dr N Wallis

In Attendance:
Miss S Begum	(FSA, item 6)
Mrs F Cleaver	(FSA, item 6)
Mr K Mistry	(DH)
Dr K O'Leary	(DH Tox unit, items 4, 7 & 8)
Mr K Okona-Mensah	(DH Tox unit, items & 8)

ITEM 1: APOLOGIES FOR ABSENCE/ANNOUNCEMENTS

1. Apologies for absence were received from the following members: Professor A Boobis, Ms D Howel; the following assessors: Ms A Gowers (EA), Mrs M Meldrum (HSE), Dr S Payne (National Assembly for Wales); and from Dr J Cunningham (DH Tox Unit).

Announcements

2. The Chairman welcomed Ms F Cleaver and Miss S Begum (FSA), Dr K O'Leary and Mr K Okona-Mensah (DH Tox Unit), and Mr M Hosford (EA) .

3. Members were reminded of the need to declare any relevant interests before discussion of items.

ITEM 2: MINUTES OF THE MEETING OF 13 JULY 2006 (CC/MIN/06/2)

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: REPORT ON PROSTATE CANCER AND PESTICIDE EXPOSURE (CC/06/22)

6. Dr Miller declared a personal, non-specific interest. The Chairman decided that he should continue to participate in the discussion.

7. The Committee was reminded that a statement on prostate cancer had been published in December 2004 (COC/04/S6). One of the overall conclusions was that there was some limited evidence to suggest an association between farmers/farm workers, exposure to pesticides and increased risk of prostate cancer. The possibility that such an association was causal could not be discounted and the published literature should continue to be monitored for further relevant papers.

8. The Pesticides Safety Directorate (PSD) of Defra had asked for the Committee's views on a report commissioned by Defra from the Institute of Occupational Medicine (IOM) entitled "Desk study on prostate cancer and pesticide exposure" and on whether the report impacts on the conclusion in the COC statement outlined above. The Committee was also asked to consider a recent meta-analysis by van Maele-Fabry et al (Cancer Causes Control 17: 353-373, 2006) on the risk of prostate cancer in pesticide manufacturing workers.

9. The IOM report reviewed the epidemiology on pesticide exposure and the risk of prostate cancer, and also the potential mechanisms that might underlie any link between pesticide exposure and prostate cancer. The report found that the results of a number of studies of pesticide manufacturing workers have failed to demonstrate a consistent association between pesticide exposure and prostate cancer. The results of large studies of manufacturing workers exposed to phenoxy herbicides do not suggest a link with prostate cancer.

10. In the van Maele-Fabry paper, the overall meta-analysis of 16 studies of workers ever employed in pesticide manufacturing and potential exposure to pesticides yielded a meta-rate ratio of 1.28 (95% CI 1.05-1.58). The paper combined RR estimates from both incidence and mortality studies to derive the overall RR value of 1.28. After grouping the data into specific chemical classes of pesticide, increased pooled rate ratios were observed for all groups but statistical significance was found only for phenoxy herbicides contaminated with dioxins and furans.

11. Members noted that there was no indication from regulatory animal studies that pesticides increase the risk of prostate cancer and were informed that the original hypothesis had arisen from US epidemiology studies. They questioned the wording of the conclusion of the IOM report "that there was insufficient evidence to conclude that pesticide exposure is associated with an increased prostate cancer risk" and pointed out that there was a considerable amount of relevant data but no consistent support for an association. They also commented that there was an imbalance in the discussion of mechanism of action. This proposed an endocrine disrupting mechanism based on androgen imbalance. However, there were papers which refuted this hypothesis but which had not been discussed in the report. Moreover, pesticides are not a generic group but comprise compounds of different structures and biological activities. Some may have oestrogenic properties, others do not. It is, therefore, inappropriate to propose a single mechanism for all pesticides.

12. It was noted that the IOM paper mentioned that there were reports linking the incidence of prostate cancer to a western diet including high levels of red meat and that cooked meat contains PhIP, which induces prostate cancer in rats.

13. Members considered the van Maele-Fabry paper and commented that, whereas the IOM paper was simply a narrative review, the van Maele-Fabry paper was a meta-analysis which could possibly result in a conclusion which differs from the individual papers. They noted that it seemed to be a careful review in terms of methods and conclusions but no test had been carried out for heterogeneity. Moreover, adjustment for confounding might have been unsatisfactory, as little is known about the risk factors for prostate cancer. Members also noted the lack of good exposure data in the epidemiology studies on prostate cancer and pesticides and commented that it was unclear whether manufacturing workers or farmers had the greater exposure.

14. In response to the questions asked of the Committee, members confirmed that the studies considered in the narrative review by IOM did not provide consistent support for an association between exposure to pesticides and risk of prostate cancer in farmers/farmworkers or pesticide manufacturing workers. The meta-analysis by van Maele-Fabry et al (2006) was considered to be of good quality although there were some unresolved methodological issues. It did support a weak association. Causality could not be inferred from the available data.

15. Members were asked to comment on the difference found in the results for the phenoxy herbicides in the IOM and the van Maele-Fabry et al (2006) papers. They replied that possible reasons included: different levels and durations of exposure, differences in the range of compounds to which subjects were exposed, differences in the levels of PCDD/F contamination, random variation in background incidence on top of a weak casual association with herbicide exposure, and random variation in background incidence.

16. The secretariat was asked to draft a short statement for the Committee's consideration which updated the 2004 statement.

ITEM 5: COMPARATIVE RISK ASSESSMENT: APPLICATION OF THE MOE APPROACH FOR COMMUNICATING THE RISKS OF EXPOSURE TO GENOTOXIC CARCINOGENS (CC/06/20)

17. Members were reminded that the comparative risk assessment (CRA) project was last discussed at the COC meeting in July 2006. Following the Committee's endorsement of the COM's suggestion that a risk professional should be invited to advise on how the COC could develop the CRA project further, Professor Lynn Frewer (Wageningen University, The Netherlands) had been approached and had agreed to participate in the discussions. Unfortunately, she was unable to attend this meeting but had undertaken to provide comments for tabling at a future meeting.

18. In addition, members were reminded of the FSA's request for COC advice on the potential usefulness of the Margin of Exposure (MOE) approach, which is being developed by the European Food Safety Authority (EFSA), WHO and the International Life Sciences Institute (ILSI) as a way of prioritising the risks associated with unavoidable exposure to genotoxic carcinogens in food. Paper CC/06/20 addressed this and provided a brief overview of a paper by Barlow et al (2006) that reports on an international conference organised by these organisations. Members were also provided with a number of worked examples of MOEs calculated by the WHO/FAO Joint Expert Committee on Food Additives (JECFA) and in a paper by O'Brien et al (2006).

19. It was noted that the MOE concept had been incorporated in the guidance document for risk assessment of carcinogens under the EU Risk Evaluation, Assessment and Authorisation of Chemicals (REACH) initiative, and that it was used for risk ranking, not for risk assessment.

20. Members considered that the use of the MOE to prioritise the risk of exposure to genotoxic carcinogens was consistent with the most recent COC guidelines and would also contribute to international harmonisation in this area. The Chairman considered the usefulness of the MOE in aiding the general public to rank risks would be a critical test of the concept and commented that the comments from Professor Frewer would be valuable.

21. The COC discussed the relative merits of using BMDL₁₀ or T25 and noted that using the BMDL₁₀ would result in a different MOE from the T25. Thus it was not possible to compare MOEs derived using the BMDL₁₀ from those derived using the T25. It was noted that the BMDL₁₀ takes the quality of the data into consideration, due to the use of the use of confidence intervals, whilst the T25 does not. The T25 was usually derived from poorer data sets. Overall, it would be necessary to decide which to use on a case-by-case basis depending on the quality of the available data.

22. A system for banding MOE values proposed by the FSA was also discussed. This expanded on proposals for the interpretation of the magnitude of the MOE that had been made by JECFA and EFSA , where there was a consensus that an MOE greater than 10,000 indicated low concern. Members considered that a banding system would be feasible but were unsure where the banding should fall.

23. Committee Members considered that such a banding system might be used to improve the communication of advice on genotoxic carcinogens to wider audiences but it would be necessary to get the description of the bands right for a lay audience. It was suggested that the MOE banding system could be used to put into perspective known risks such as tobacco and alcohol, but there was no consensus on this. The Committee considered that the interpretation of the banding and terminology used for wider audiences would need to be considered more closely and requested a future discussion on this area once the comments from Professor Frewer and, possibly, other risk communication experts have been obtained.

24. Members supported an extension of the use of the MOE and a banding approach to genotoxic carcinogens that are environmental contaminants, and suggested that polycyclic aromatic hydrocarbons and benzene would be suitable candidates.

ITEM 6: BETEL QUID, PAN MASALA AND ARECA NUT CHEWING (CC/06/18)

25. Members were informed that areca nut was an ingredient of betel quid or pan masala, which is chewed as an aid to digestion and as a stimulant. Areca nut may have limited use as a food ingredient. The COC last considered the carcinogenicity of areca nut in 1994. Since then, a number of relevant new papers had been published. The Food Standards Agency was seeking the view of the COC on whether the new information was sufficient to warrant updating its previous advice, and whether any conclusions could be drawn about the use of areca nut as a food ingredient.

26. In the previous COC assessment, the Committee had considered a range of human epidemiology studies, animal carcinogenicity studies and in vivo and in vitro mutagenicity studies. On the basis of the previously available evidence, the COC concluded the following:

• There was evidence of mutagenic and carcinogenic activity of areca nut extracts and derived compounds in experimental systems. In particular, the potent carcinogenic activity of the areca-derived nitrosamine, 3-(methylnitrosoamino)-propionitrile (MNPN), had been confirmed and methyl and cyanoethyl adducts had been detected in the DNA of the target tissues in which the tumours developed. There was evidence that endogenous nitrosation of areca nut alkaloids can occur in animals and humans; and areca nut derived nitrosamines, including MNPN, have been detected in the saliva of betel quid chewers.
• There were very limited data from epidemiological studies on the effect of betel or areca nut products without tobacco, which did not allow any conclusion to be drawn. There was, however, sufficient epidemiological evidence of a link between the chewing of betel quid containing tobacco and cancer in humans.
• The Committee concluded that the use of these products without tobacco was possibly carcinogenic in humans.

27. In 2003, the International Agency for Research on Cancer (IARC) assessed the use of betel quid, and concluded that both the chewing of betel quid and areca nut should be categorised as Group 1 carcinogens (carcinogenic to humans). In its conclusions, IARC stated that there is sufficient evidence in humans to conclude that betel quid chewed without tobacco causes oral cancer. IARC (2003) also concluded that there was sufficient evidence in animals to confirm the carcinogenicity of betel quid and areca nut without tobacco. This followed a previous IARC review in 1985, which concluded that there was inadequate evidence for the carcinogenicity of betel quid chewed without tobacco.

28. Annexes 3 and 4 of the paper CC/06/18 summarised a number of epidemiology studies published since 1993. Annex 3 contained human epidemiology studies that considered the effects of betel quid usage and oral cancer or pre-cancerous lesions. Further human studies which described possible relationships between areca nut use and the incidence of other forms of cancer, such as liver cancer, as well as some animal and in vitro studies, were provided at Annex 4. Annex 5 contained tables of epidemiology data assessed by IARC in 2003.

29. Members considered the new epidemiology studies that looked at the relationship between oral cancers and areca nut use, new epidemiological studies that looked at areca nut use and liver cancer, and new animal studies which attempted to identify mechanisms for the carcinogenic potential of areca nut extracts.

30. The Committee agreed with the above IARC (2003) conclusions. Members considered that most of the evidence suggested that areca nut is a site of contact carcinogen acting by a genotoxic mechanism, and noted that there was some evidence that a non-genotoxic oxidation mechanism could also be involved. Members considered that the carcinogenic mechanism was likely to involve the nitrosation of precursor chemicals present in areca nut. It was not clear why there appeared to be an increased site-of-contact cancer risk for those who swallowed the juice. The COC also considered that there was no clear evidence that the use of areca nut without tobacco caused primary liver cancer or other cancers at distant sites.

31. The COC concluded that there was now sufficient evidence for a causal link between the chewing of areca nut without tobacco and oral cancer and that, overall, areca nut should be regarded as carcinogenic to humans. Members also agreed that the use of areca nut as a food ingredient was potentially carcinogenic.

ITEM 7: ACUTE T25 - POSSIBLE APPROACH TO POTENCY RANKING OF SINGLE EXPOSURE GENOTOXIC CARCINOGENS (CC/06/19)

32. The Committee was reminded that the Health Protection Agency and other government departments and agencies sometimes have to provide advice on the carcinogenic risk following a single exposure to a genotoxic carcinogen, for example, following a chemical accident. There is evidence from animal studies that a single exposure to potent genotoxic carcinogens may be associated with higher cancer risk during later life stages. Because the COC recommends the prudent assumption that there is no threshold for genotoxic carcinogens, government agencies can only give generic advice indicating that any carcinogenic risk from a single exposure, such as following a chemical accident, is likely to be very small. It would be helpful if more information could be given on carcinogenic risks following such exposures. One way may be to grade the potency of genotoxic carcinogens following a single exposure. It may be that any risk from a single exposure to low potency genotoxic compounds could be regarded as negligible.

33. The DH Toxicology Unit had therefore investigated the available animal data on single exposure to genotoxic carcinogens and the subsequent induction of tumours. A major problem was the limited amount of data in this area from which to assess potency and grade compounds. It had previously been proposed that the nine most commonly studied compounds in the Calabrese database could be the subject of a pilot study. It was considered that it would be useful to compare the rank order of potency of these compounds following single exposure to that obtained following conventional long-term carcinogenicity bioassays. If the rankings were similar, it could be possible to make assumptions about acute potency in those many cases where data are only available from chronic studies.

34. The data available for this exercise had proved to be limited and it was possible to calculate acute T25 values for only 5 of the 9 chemicals (dibenz(a,h)anthracene, N-nitrosodiethylamine, N-nitrosodimethylamine, methylnitrosourea and benzo(a)pyrene). Chronic T25 values were calculated for all 9 chemicals. These were compared with the TD₅₀ values from the Gold and Zeiger database and a good correlation found. However, there was no apparent correlation in the ranking of calculated acute T25 values and chronic T25 values for the 5 chemicals for which acute T25 values could be calculated.

35. Members agreed that the above approach was not useful, due to the limited data, and that there would not be any value in following this further. Members were not surprised that there was no correlation between the rank order of acute T25 values and the rank order of chronic T25 values as most of the chronic values had been obtained after oral dosing, whereas a range of routes had been used for the single dose exposures. Another potential problem was the difference in age of the animals at dosing ie the comparison had been made between lifetime exposures in the chronic studies and mainly newborn animals in the acute studies. It was noted that it would also be difficult to obtain comparative data between chronic and acute studies for the same species, strain and sex. The Committee also noted that the T25 value was a concept intended to relate to long-term exposure, which could produce chronic cytotoxicity and repeated DNA repair, which contribute to the carcinogenic risk. Lifetime cancer incidence depends on the accumulation of risk and one would not necessarily expect a correlation between the risk from acute exposure and that from chronic exposure.

36. Members stated that clarification was needed on whether the concern was about the consequences of single exposures or of short-term exposures. If the latter, it might be more useful to compare short-term with long-term exposures, rather than using single dose studies. However, such data were rarely published and some of the available data had been used in the above exercise. One member pointed out that the NTP had conducted some animal studies, which suggested that short-term exposure in early life produced a greater lifetime cancer risk than would have been expected from a linear extrapolation. It was suggested that, if the only data came from long-term studies, it might be possible to apply some adjustment factor when providing advice on acute exposure.

37. Members agreed that the margin of exposure (MOE) approach (discussed in item 5) might assist in providing advice following chronic exposure but deferred the issue pending consideration of the NTP data described above.

ITEM 8: HORIZON SCANNING 2006 (CC/06/21)

38. The annual horizon scanning paper was presented for members' comments. A number of items had been selected following literature searches and these were presented in CC/06/21, together with a review of the outcome of the 2005 horizon scanning exercise. Members and assessors were also invited to suggest topics for further work.

39. CC/06/21 sought clarification on two topics identified for further work in the 2005 exercise. In response, members asked to see the data on trends in breast cancer before deciding whether further work was necessary on this endpoint. On computational systems biology, it was suggested that, as a first step, it might be useful to consider reviews in this area or the work carried out by the Institute of Environmental Health Sciences (IEHS), with a view to a possible presentation to the Committee on this topic.

2006 exercise (priorities are marked high (H) or medium (M))

40. Dibenzo(a,l)pyrene: It was suggested that this could be considered as part of the work on the carcinogenicity of mixtures (see below).

41. Formaldehyde: It was suggested that formaldehyde might form Schiff bases which can reform formaldehyde at sites remote from the site of contact. The Committee considered that a further review of formaldehyde might be useful (M).

42. Hexachlorobenzene: Members did not support further work on this chemical.

43. Studies with transgenic animals: Members did not support further work on this topic.

44. Toxicogenomics: Members considered that it would be useful to follow up on the role of toxicogenomics in predicting carcinogenicity and in confirming or proposing a mechanism of action (H). This might be a joint COC/M/T project. A member noted that he had attended a meeting which indicated that ochratoxin A might affect DNA by oxidative damage and this was in line with the results of one of the studies in CC/06/21. Another member noted that such a mechanism had also been proposed for aristolochic acid and suggested that this should be considered as a general topic.

45. Thresholds for genotoxic carcinogens: Members supported further work on the characterisation of the dose-response at low exposures (H).

46. Mechanisms of action: Members were interested to see the paper by Sielken et al (2005) (M).

47. Potential biomarker: Nitrative DNA damage: Members did not support further work on this topic.

48. A possible framework to derive quantitative risk estimates based on cancer epidemiological data: Members did not support further work on this topic.

49. Methodological flaws in epidemiological studies linking carcinogen-DNA adduct levels with cancer risk: Members were interested to see the paper by Rundle (2006) (M).

50. Mode of Action (MOA): the Committee was informed by a member that the IPCS had updated its MOA framework to include Human Relevance Framework (HRF) considerations and a publication was currently in press, together with a few illustrative case studies. The Committee decided that, when these papers were published, it would review them, along with the paper by Holsapple et al (2006), in the context of taking a new look at the HRF (H).

51. Carcinogenicity of mixtures and mutational fingerprints: Members agreed that both these topics should be taken forward jointly with the COM (H).

52. Other topics: It was proposed that the COC should consider the carcinogenicity of depleted uranium, including uranium pyrolysis. It was also noted that the Committee needed to keep up-to-date on developments in the assessment of carcinogenicity of nanoparticles. The Committee considered a publication on cooking fumes exposure and lung cancer among Chinese nonsmoking women and asked the secretariat to discuss with the HPA Air Pollution Unit the extent of existing UK work on indoor air pollution due to cooking fumes.

ITEM 9: ANY OTHER BUSINESS

53. There was no other business.

ITEM 10: DATE OF NEXT MEETING

54. 1 March 2007

ACTIONS

ITEM	ACTION	RESPONSIBILITY
4. Report on prostate cancer and pesticide exposure	Update COC statement/working paper	Secretariat (HPA)
5. Comparative risk assessment	Obtain expert advice Rethink banding proposal	Secretariat (HPA and DH TU) Secretariat (HPA and FSA)
6. Betel quid, Pan Masala and areca nut chewing	Prepare statement/working paper	Secretariat (FSA)
8. Horizon scanning	Take priority items forward	Secretariat

Top

copyright: © | updated: 6 March 2007