Minutes

ITEM 1: APOLOGIES FOR ABSENCE/ANNOUNCEMENTS

2. Apologies for absence were received from members Dr D Shuker, Professor G Williams, Professor P Farmer, adviser Professor H MÆ ller, and assessors Mrs K Cameron (DEFRA) and Mr A Browning (VMD).

Announcements

3. The Chairman welcomed Professor A Boobis from Department of Health Toxicology unit at Imperial College (DH Tox unit), the deputy chair of the Advisory Committee on Pesticides (ACP) (attending for item 4 & 6), Dr T England from DH Tox unit (attending for item 6) and Mr S Samuels form the Pesticide Safety Directorate (PSD) attending in the place of Mr T Holmes.

4. The Chairman thanked Professor H MÆ ller for his valuable contribution as an adviser to the COC during the sabbatical of Professor Forman.

5. Members were reminded of the need to make any relevant declarations of interest before discussion of items.

ITEM 2: MINUTES OF THE MEETING ON 7 MARCH 2002 (CC/MIN/2002/1)

6. Minutes were agreed subject to minor editorial changes.

ITEM 3: MATTERS ARISING

3.1 COC conclusions on intrahepatic cholangiocarcinoma (CC/02/21)

7. The COC had previously finalised its conclusions on the evidence for an increase in mortality from intrahepatic cholangiocarcinoma in England and Wales. Members had recommended that a number of additional studies were necessary before conclusions could be reached on whether the observed increased incidence of this tumour was real. Since the publication of the COC statement on increased intrahepatic cholangiocarcinoma, Professor Thomas (from Queen Mary's Medical School) had notified the secretariat that he was planning an internal audit of cases over the past 10 years, and that this could be used as a pilot for a national study.

8. This study proposed to use case notes to assess the accuracy of diagnosis and ICD coding. A study proposal was tabled for members to consider.

9. Members had a number of concerns over the proposals. It was agreed that the researchers had underestimated the amount of work and cost that would be involved, and suggested that it would therefore be better to examine fewer cases. The committee had concerns at the lack of clear criteria for diagnosis and the subjective nature of the decision regarding the accuracy of diagnosis would be open to bias. Members suggested that clear preset criteria should be required to categorise the cases and that an independent expert panel should be used to make the assessment.

10. The secretariat would write to Professor Thomas outlining the COC comments and would get back to the committee with the response.

11. An abstract of a paper by Professor Thomas' group, on DNA adducts in cholangiocarcinoma tissue was attached to paper CC/02/21. This paper suggested that environmental genotoxins could play a role in the development of Cholangiocarcinoma. Members requested to see the data for this study, but considered that it did not provide any significant evidence for a causative role by environmental genotoxins. The committee commented that there were few samples and a lack of control non-tumour tissue from patients.

3.2 COT/COC/COM Glossary

12. Members were informed that a draft glossary of technical terms used by the COT/COC/COM (CC/02/09) had been produced and had been aimed at the informed lay person. The Chairman noted that the level of detail was variable and suggested that the glossary should adopt a consistent level with respect to the target audience. Members were requested to send any comments to the secretariat.

3.3 Non-genotoxic carcinogens

13. The COC was reminded of the presentation by Professor Chipman and Dr Malley at the previous meeting, on biomarkers for non-genotoxic carcinogens, particularly the observation of reduced connexin plaque levels in target organs in rats following short term dosing with a number of rodent non-genotoxic carcinogens. The secretariat had received a draft report from Professor Chipman, which had also been submitted for publication and presented some additional details relating to histopathology in animals receiving non-genotoxic carcinogens. This supported the view that connexins could be used to distinguish between non-genotoxic carcinogens and cytotoxins. The secretariat agreed to forward Professor Chipman's draft report (in confidence) to members for their comments on the results and the suggestion for additional research.

3.4 Non-genotoxic carcinogens

14. Members were informed that a recent survey of the levels of acrylamide in fried and baked food in Sweden had revealed unexpectedly high levels (e.g. levels ranging from about 50-2300 ppm in chips, crisps, and rivita). The Food Standards Agency had subsequently carried out a limited survey using two different methods of analysis (one validated for determination of acylamide in foods and the other based on the methods adopted in the Swedish survey) in order to substantiate the Swedish results. These investigations had confirmed similar results to the Swedish survey in the samples of food analysed. Members recalled that the COC had concluded in 1992 that acrylamide should be regarded as a genotoxic carcinogen, and that it was also a germ cell mutagen in experimental animals. In addition it also had neurotoxic properties.

15. The Food Standards Agency had held a series of stakeholder meetings. The mode of acrylamide formation in foods was not presently understood. It had been speculated that the unexpectedly high levels were possibly related to cooking time and temperature. The World Health Organisation had organised an expert consultation meeting (to be held on 27 June 2002) with the intention of evaluating new and existing data on acylamide. Professor Farmer who had investigated acrylamide toxicity and DNA adduct formation was attending as an invited expert. The Committee would be informed of the outcome of the WHO meeting.

ITEM 4: REVIEW OF THE CARCINOGEENICITY OF MALATHION (CC/02/13,14,16,17)

16. The minutes of this item are now published please see paragraphs 36 to 58. The statement is also available please see COC/03/S1.

ITEM 5: IGHRC FRAMEWORK DOCUMENT ON TOXICOLOGICAL CERTAINTY AND UNCERTAINTY FACTORS IN CHEMICAL RISK ASSESSMENT (CC/02/15)

17. The Interdepartmental Group on Health Risks from Chemicals (IGHRC) is developing cross-Government guidance on the handling of uncertainty in the toxicological hazard aspects of human health risk assessment. IGHRC intends to produce a document setting out a harmonised framework for UK Government departments, agencies and their advisory committees on how to address the uncertainties in toxicological hazard aspects of risk assessment including the derivation and application of uncertainty factors.

18. The Chairman welcomed Professor Purchase (Chair of Executive Committee of IGHRC) and asked him to make a short presentation on the work of IGHRC and the draft IGHRC Framework document on toxicological certainty and uncertainty. Professor Purchase informed members that IGHRC aimed to promote the development of improved approaches to toxicological risk assessment, to enhance coherence and consistency between Government Departments and to advocate best practice. A number of primary objectives had been set to address these aims namely, to publish a programme of work, to identify through research needs the development of innovative methods/improved approaches, to provide a forum within Government for discussing how greater coherence and consistency can be achieved, and to identify and disseminate best practice in collaboration with stakeholders. Professor Purchase overviewed the series of reports published by IGHRC, which included a weight of evidence approach to chemical carcinogens where considerable input had been derived from the COC secretariat. He noted two draft guidance documents on human exposure assessment (in preparation) and uncertainty in toxicological risk assessment (tabled for COC discussion at COC).

19. Professor Purchase briefly mentioned the ongoing literature research on variation in human toxicodynamics which was due to report in November 2002 and the future project to define a UK exposure model (contract recently awarded to Imperial College May 2002). A database (Risk Assessment Methodology Research Database RAMRED http://wads.le.ac.uk/ieh/ramred/index.htm) had been launched in 2001 in order to provide information on the key research areas. The information would be of value to researchers in developing project proposals. The Medical Research Council (MRC) and Biotechnology and Biological Sciences Research Council (BBSRC) were aware of the key areas for research and were attempting to encourage submission of research proposals. Professor Purchase also noted a number of meetings had been set up to aid in dissemination of and sharing of knowledge and experience. These included a work shop on PBPK modelling and one on probabilistic modelling (which was meeting on 27/6/2002). A course on publishing transparent risk assessments had been completed in October 2001 and a further course on this aspect might be set up in the future. One further workshop was planned on the interface between risk assessment and risk management. Professor Purchase felt that possible areas of future work could include risk assessment of chemical mixtures and terminology in risk assessment.

20. Dr Delic introduced the paper CC/02/15 and asked members for their comments. The document provided a review of the approaches used in chemical risk assessment in the UK. The draft document reflects the current position in relation to the concept of 'no-threshold' for genotoxic carcinogens as laid out in the 1991 COC Guidelines for the Evaluation of Chemicals for Carcinogenicty. It was noted in the covering paper that the COM had recently reaffirmed its position that no threshold could be assumed for in-vivo mutagens in the absence of compound specific mechanistic data to suggest otherwise. The draft IGHRC document indicated that for carcinogenicty believed to arise through a non-genotoxic mechanism of action, an approach using a conventional approach based on estimation of a NOAEL and application of uncertainty factors could be adopted. It was also recognised that in some cases an extra Uncertainty Factor may be applied to non-genotoxic carcinogens because of concern over the severity (and irreversible nature) of the effect or uncertainties in the mechanism of carcinogenicity. Members confirmed that the approach set out in the document was acceptable. Members provided a number of written comments for the IGHRC secretariat.

21. Members welcomed the document. As a general comment it was thought that more tabulation of information would make the document more accessible. Members commented on the interaction between risk assessment and risk management, which was mentioned in section 2.1 (page 5 of the draft) and commented that these should be undertaken as separate procedures. It was pointed out that in some instances there would be a limited range of risk management options and this could influence the approach used to risk assessment, but not the outcome of any risk assessment. Members commented that the Bench Mark Dose (BMD) approach made use of the data from all dose levels and avoided uncertainties in trying to set a No Adverse Effect Level (NOAEL). However the BMD required more doses and might result in greater animal usage. In respect of risk assessment of chemical carcinogens, members cautioned against a "numerical" approach to the use of data derived from long term bioassays in animals in risk assessment. It was noted that the interpretation of long-term cancer bioassays was influenced by, often subtle, interpretation of histology.

ITEM 6: REVIEW OF COC GUIDELINES ON CARCINOGEN RISK ASSESSMENT: DISCUSSION PAPER (CC/02/18)

22. The COC had given some initial consideration to this issue at its previous meeting (March 2002). Paper CC/02/18 reviewed the application of mathematical models to carcinogenicity data derived from long-term animal bioassays, the proposed use of potency estimates derived from animal bioassays, the derivation of a threshold of regulation for carcinogens, and the derivation of minimum risk levels for genotoxic carcinogens. The objective of the paper was to initiate discussion of the scientific basis for various approaches to carcinogen risk assessment for genotoxic carcinogens. It was hoped that the discussion could be used to produce a first draft of a revised version of the COC guidelines on chemical carcinogenicity risk assessment.

23. Members had concerns over the use of all animal mathematical models for carcinogen risk assessment when extrapolating from the relatively high dose levels used in animal bioassays to below the observed region of a dose-response relationship i.e. to levels of human environmental exposure. Some of the approaches used complex mathematics and others were fairly simplistic, but they involved a large degree of uncertainty and lacked scientific understanding of the underlying biological processes and mechanisms leading to carcinogenicity. The different types of models developed to derive risk estimates from animal bioassays were considered to vary greatly in their predictions.

24. It was agreed that the Bench Mark Dose (BMD) point of departure analysis such as the use of a dose associated with a 10% excess risk of cancer (LED10) was theoretically preferable to the use of NOAEL/LOAELs from the cancer bioassay and the application of uncertainty factors. These points of departure approaches mad better use of all the data to estimate the dose-response curve. The use of NOAEL/LOAELs was considered to be limited due to being very dependent on the chosen dose levels and dose spacing. However, potential problems with a BMD approach to carcinogenicity risk assessment involved the requirement for the use of large numbers of animals (in excess of current guidelines) and the associated need to produce a positive response in more than one test group. Additionally, there were considerable problems associated with scaling animal BMDs to humans, which also applied to other approaches.

25. Regarding the use of a linear-dose response, there were concerns over the lack of scientific justification to support the extrapolation to orders of magnitude below the observed dose range. It was accepted that there was a linear-dose response for the formation of DNA adducts, an early stage of the carcinogenic process. However, there were no biomarkers for the following steps leading to cancer and insufficient understanding of the mechanism. Members considered there would probably be a number of individual steps contributing to the total process, which might involve a number of sequential individual thresholds.

26. Members suggested that it might be informative to consider the analysis of examples of sets of carcinogenicity bioassay data using different mathematical models and applying sensitivity analysis. Members agreed it would be useful to review some worked examples, particularly with regard to quantitative risk assessment models not considered in 1991 COC guidelines. Two issues were considered important, fitting a particular model to available data points and then the method of extrapolation from the point of departure. Members also noted that the selection of the top dose could profoundly influence the outcome of subsequent modelling.

27. Overall, the committee had not changed its previous views on mathematical modelling using animal data and did not think that any further research to define or develop mathematical models would be worthwhile. It was accepted that attempts had been made to produce more biologically based mathematical models, but these still contained too many uncertainties regarding the mechanism of carcinogenesis and thus assumptions in producing risk estimates. Key uncertainties in the carcinogenicity process included lack of information on the frequency of mutations caused by DNA adducts, the efficiency of DNA repair, and the process by which mutations lead to transformed cells and eventually neoplasia. It was agreed that further chemical specific research into the mechanism of carcinogenicity would be helpful for risk assessment purposes and might help to define thresholds. Thus in addition to the proposal from COM for work to establish thresholds for mutations, in-vitro studies could be used to examine the dose-response for relevant chemical specific metabolic activation/detoxication or aspects of mechanism (i.e. whether O⁶ methylation was saturable), which might help in assessing the potential for carcinogenesis. Studies involving DNA repair deficient animals could also be useful in this respect.

28. Members considered potency estimates (such as T25) could have some pragmatic use in carcinogen risk assessment, such as helping regulators to prioritise, but not for quantifying cancer risks. The proposals to use T25 in cancer risk assessment made crude assumptions such as relative ranking in the observed dose range was maintained at low doses, and that relative animal potency could be used to scale down to human exposure. Members heard that the additional relative ranking work suggested by COC for comparing animal with human potency estimates had not been taken forward since the quality of epidemiology on many chemicals was too limited, and due to the need for considerable expert epidemiological input.

29. A 'threshold of regulation' has been proposed as a method for setting regulatory minimum risk level for carcinogenic contaminants. It is derived from an evaluation (involving linear extrapolation down to 10^-6 excess cancer risk) of substances in the Gold Carcinogen Potency Database, which uses TD50 as a potency estimate. The FDA originally introduced this approach as a method for reducing toxicological data requirements for low levels of chemicals leaching from food contact materials. Recent workshops had attempted to develop the concept that a de minimis value (Threshold of toxicological concern (TTC)) could be identified for any chemical, including those of unknown toxicity, taking chemical structure into consideration (Barlow et al (2001) Thresholds of toxicological concern for chemical substances present in the diet. Report of a workshop, 5-6 October 1999, Paris, France, Food Chem Toxicol 39;893-905). It had been proposed to extend TTC to carcinogenicity, but with criteria to exclude chemicals with certain structural alerts for mutagenicity and TCDD. It was noted that JECFA had considered using a 'threshold of regulation' of 1.5 micrograms/d, which excluded some structural alerts (including dioxins and nitrosamines) due to their potency. The committee felt that this 'one dose fits all approach' would have limited use, as the proposed exclusions covered the major classes of genotoxic carcinogens and TCDD. Although members had reservations about such an approach, the committee agreed to consider the TTC again when a draft COC guideline document became available.

30. Members considered that under certain specific circumstances, where there was an unavoidable but low exposure to genotoxic carcinogen contaminants, a pragmatic approach to estimating an exposure level that is associated with a minimum risk could be used, although it should always be recognised that there was some risk and ALARP should apply. This process might involve dividing a minimum tumourigenic dose by a factor (such as 10,000) derived by expert judgement to produce a margin of exposure considered to present a minimum risk to health. Any exposure, however, should be kept to as low as reasonably practical.

31. The committee was made aware of a paper (van den Brant et al. 2001. The contribution of epidemiology. Food Chemical Toxicology 40, 387-424) that could be helpful in assessing the value of human epidemiological data in carcinogenicity risk assessment. The paper suggested criteria for categorising the data. The categories included 1, sufficiently good to derive a dose-response relationship, 2 adequate to check the plausibility of an effect, and 3 useless. Members agreed to review this paper and others from this edition of Food and Chemical Toxicology at the next meeting.

32. The Chair thanked members for their comments. The secretariat and the DH Toxicology unit would use comments raised in the discussion to help produce a draft revised version of the COC guidelines on the evaluation of chemicals for carcinogenicity.

ITEM 7: PAPERS FOR INFORMATION

33. The following papers were tabled for information:

7.1 Shen L et al (2002). DNA methylation and environmental exposures to chemicals in human hepatocellular carcinoma. (CC/02/19)

72. FSA Report on Review of Scientific Committees. (CC/02/22)

ITEM 8: ANY OTHER BUSINESS

8.1 Geographical epidemiology of prostate cancer in great Britain Jarup et al In J of Cancer vol 97, 695-699, 2002.

34. This paper was provided for comment. Members asked whether there was any evidence for a significant variation of chemical exposure throughout the UK. The committee agreed that the study had been adequately performed and that no evidence for significant geographical variation in prostate cancer incidence had been produced. However, members were aware of evidence for an increasing trend. This in part, might be due to improved diagnoses and registration. It was felt that it was appropriate and timely to consider the evidence with regard to potential chemical induction of prostate cancer.

ITEM 9: DATE OF NEXT MEETING

35. 21 November 2002.

PARAGRAHS 36 TO 58 ADDED on 28 March 2003

ITEM 4: REVIEW OF THE CARCINOGEENICITY OF MALATHION (CC/02/13,14,16,17)

36. Members were asked to consider declarations of interest at the previous meeting. The manufacturer of malathion is Cheminova Agro A/S (referred to as the pesticide data holder). Pesticide approval holders are Petlife International Ltd, and United Phosporous. Human licences are held by EC De Witt and Co Ltd, Seton Products Ltd and Ultra Chemicals Ltd.

37. Malathion is an organophosphorous insecticide. It has been marketed in the UK for use in agriculture and horticulture since 1956. There are currently three products with approvals for use in agriculture, home garden and use in pigeon lofts. A number of products containing malathion are also licensed as human medicines for use in the control of head lice.

38. The Advisory Committee on Pesticides (ACP) is reviewing the available toxicological information on malathion as part of its ongoing review of organophosphorous compounds. The ACP has asked for advice from COM on mutagenicity, and COC on carcinogenicity of malathion. It is proposed that a joint COC/COM statement be produced and that this be finalised shortly after the COM meeting on 10th October 2002.

39. The COC was informed that malathion and some of its metabolites/impurities had the potential to methylate DNA. The COM had concerns regarding the potential importance of impurities present in technical grade malathion in the observed mutagenicity in a number of studies it had reviewed. The COM agreed that conclusions should be drawn on technical grade malathion and had requested information from the data holder on the influence of manufacturing process on the formation and levels of impurities. The COM had agreed that technical grade malathion was not mutagenic in bacteria or in an in-vitro UDS assay using hepatocytes. However the data holder had recently undertaken an in-vitro cytogenetics assay in human lymphocytes and a mouse lymphoma assay and had documented positive results with technical grade malathion. The COM considered technical grade malathion to be mutagenic in-vitro. There were a considerable number of in-vivo mutagenicity tests available. A negative result had been reported in an adequate oral bone marrow mouse micronucleus assay. However positive results had been produced when technical grade or malathion of higher purity had been tested in the bone-marrow of mice using the intraperitoneal route. There was insufficient data for the COM to draw conclusions regarding the mutagenicity of technical grade malathion in other tissues. The COM would review its position after consideration of additional studies (both published and from the data holder) and further information on impurities in technical grade malathion at its next meeting.

4.1 Carcinogenicity of malathion

40. The Committee was asked to review the conduct, adequacy and results from the available carcinogenicity bioassays in rats and mice where malathion or malaoxon (a metabolite of malathion) had been administered in the diet. The relevant data were summarised in the overview paper CC/02/13. Results from the test reports and details of the peer reviews of the pathology were appended as Annexes. Some correspondence between the data holder and the EPA was also included in these Annexes. Members were made aware of an overview from the data holder which provided additional data on the conduct of the studies undertaken at Huntingdon Life Sciences U.S.A. and on historical control incidence of nasal tumours in F344 rats from the US National Toxicology Program. This document had been provided to members by e-mail by the secretariat and had now also been tabled for discussion. Members attention was also drawn to the evaluation of the carcinogenicity reported by the EPA (CC/02/14) and a Scientific Advisory Panel review of malathion carcinogenicity (CC/02/16) which were provided as background information.

Malathion administration in the diet

41. In reviewing these studies, members agreed that particular reference should be made to the evaluation of nasal tissue. Members noted that nasal tissue was generally not examined in carcinogenicity bioassays until around 1990.

42. NCI 1978, Osborne Mendel rat. The study authors had reported that there was an increase in thyroid follicular cell carcinomas/adenomas in females in the high dose group. A peer review of pathology had concluded that there was no significant increase in thyroid tumours. Members noted there was no examination of nasal tissue, non-neoplastic pathology was reported in females but not males, there were no full necropsies of decedents and the control groups were too small. Overall it was agreed that this study was inadequate by contemporary standards and no conclusions could be drawn.

43. FDRC 1980, Sprague Dawley Rat. The study authors had not reported any evidence for a treatment related carcinogenic effect. The committee noted that there was no examination of nasal tissue. With this limitation noted, the committee felt the study had in general been adequately undertaken.

44. NCI 1979, Rat F344. There was an increase incidence in phaeochromocytomas in low dose but not high dose males, and the absence of adrenal medullary hyperplasia lessened the significance. There was limited evidence for increased leukaemia in low dose males. It was agreed that this study provided no consistent evidence for a carcinogenic effect. Members noted that examination of nasal tissue had not been undertaken in this study, but agreed that overall the study had in general been adequately conducted.

45. HLS 1993-1996, Rat F344. The COC agreed that this bioassay was conducted to modern standards and was the critical carcinogenicity study in the rat. Dietary levels used were 100 ppm (reduced to 50 ppm at month 3), 500 ppm, 6000 ppm and 12000 ppm. Members agreed that the top two dose levels in males exceeded the Maximum Tolerated Dose (MTD) level. This was based on reduced mean body weight (-16.8% at month 18 in the 12000 ppm group and -11.1% at 6000 ppm) and statistically increased mortality compared to control animals. In females the MTD was exceeded at 12000 ppm (based on reduced mean body weight at 24 months; -16.8% and significant mortality compared to controls) but was not exceeded at 6000 ppm (mean reduction in body weight <5.1% compared to controls, 62% survival to termination). The study investigators had documented two nasal tumours in males (an olfactory epithelial adenoma in one animal at 6000 ppm and an olfactory epithelial carcinoma in one animal at 12000 ppm). The COC noted that an independent peer review of the pathology had been undertaken. Members felt that the report of the peer review was satisfactory and agreed that the classsification of tumours used in this review should be used in the evaluation of this cancer bioassay. Following a peer review of the pathology, a consensus agreement was reached that there were three nasal respiratory adenomas (two (one male/one female) at 12000 ppm and one female at 6000 ppm) and one olfactory epithelial adenoma in one male at 6000 ppm. These tumours were associated with severe non-neoplastic pathology in the nasal tissue at 6000 and 12000 ppm (including epithelial degeneration, hyperplasia and inflammation). The peer review also documented four tumours of the oral palate. The data holder had provided a concise summary of the neoplastic pathology in animals with nasal tumours. Members noted that based on the relatively small laboratory historical control data, the incidence of nasal tumours was highly statistically significant. However it was evident from the EPA review and information supplied by the data holder that the background incidence of nasal tumours was unclear and depended upon the number of nasal sections used in bioassays. In modern bioassays this would comprise 5 sections whilst only two sections were used in the bioassays in the laboratory historical control database. The historical control range from 20 bioassays undertaken as part of the NTP program was 0-2% in both males and females. Members considered that overall there was evidence of a tumourigenic response to malathion in the nasal tissue of the rat but the observed tumours were induced in tissue which was subject to severe ongoing nasal inflammation. Members agreed there was an increase in liver tumours at the two highest dose levels and that this was likely to be due to an effect of treatment on the background pathology of ageing animals.

46. NCI 1978, Mouse B6C3F1. Members noted there was no examination of nasal tissue, there were no full necropsies of decedents and the control groups were too small. A statistically significant increase in combined hepatocellular carcinoma and neoplastic nodule were reported in males animals by trend analysis. No increase was reported when analysed by pair wise comparison or by life-time analysis. Members noted that the terminology used to describe histological lesions of the liver was now considered outdated and that there was a high incidence of hepatocellular carcinoma in historical control groups. The committee agreed that by contemporary standards this bioassay was inadequate. Members noted the evidence of coughing and wheezing in most high dose animals from week 72 which provided some evidence for an effect on the respiratory tract.

47. HLS 1992-1994, Mouse B6C3F1. The COC agreed that this bioassay had been conducted to modern standards. Dietary levels of 100 ppm, 800 ppm, 8,000 ppm, and 16000 ppm were used. Members agreed that the Maximum Tolerated Dose level was exceeded at 16 000 ppm in both males (-20 % reduction in mean body weight compared to controls at week 78) and in females (-16% reduction in mean body weight compared to controls at week 78). The MTD was exceeded in males at 8000 ppm (-14% reduction in mean body weight compared to controls at week 78). Members noted that the 8000 ppm dose level in females equated to the MTD (reduction in mean body weight for females was -9.7%). The study investigators reported the finding of hepatocellular adenoma, hepatocellular hypertrophy and increase liver weight in most mice (male and female) at 8000 ppm and 16000 ppm. A Peer review of histology in male mice essentially confirmed the liver findings reported by the study investigators. The Peer review group also identified significant and excessive non neoplastic pathology of the nasal tissue similar to that reported in F344 rats. The Committee agreed that the liver tumours were treatment related.

Malaoxon administration in the diet

48. NCI 1979, Rat F344. A statistically significant increase in thyroid C-cell adenomas and carcinomas were reported in females at the top dose. A Peer review had been undertaken. The Committee agreed with the conclusion reached by the peer review group that the evidence for a carcinogenic effect in the thyroid was equivocal. Members noted that examination of nasal tissue had not been undertaken in this study. The Committee noted the high dietary level was 1000 ppm, but agreed that, based on the preliminary 13 week study (no effects on body weight or histology at 2000 ppm), a higher dose level might have been used.

49. HLS 1993-1996, Rat F344. The COC agreed that this bioassay had been conducted to modern standards. Dietary levels of 20 ppm, 1000 ppm, and 2,000 ppm were used. Members agreed that the Maximum Tolerated Dose level had not been exceeded in this study (-1.4% to -7.1% reduction in body weight compared to controls in males and -4.0% to -8.8% in females at termination). However a dose related increase in mortality was reported in males. A significant dose-related trend in mononuclear cell leukaemia was reported in males but no statistically significant increase was evident in pair wise comparisons with controls. The incidence of this tumour was within the historical control for the laboratory. A significant increase in testicular interstitial cell tumours was reported at 2000 ppm when the data were corrected for time to tumour identification. However the observed increase was within the historical control incidence. There was thus no evidence for any treatment related increase in neoplasia. The study investigators found significant non-neoplastic pathology of the nasal tissue and respiratory tract. This included significant purulent inflammation of the lungs. There was evidence of food particle deposition in many animals in this study.

50. NCI 1979, Mouse B6C3F1. Members noted that examination of nasal tissue had not been undertaken in this study. The Committee noted the high dietary level was 1000 ppm, but agreed that, based on the preliminary 13 week study (no effects on body weight at 2000 ppm), a higher dose level might have been used. No treatment related neoplasia was reported.

Overall malathion carcinogenicty evaluation

51. Members agreed that the assessment of the significance of the observed nasal tumours in F344 rats in the modern study undertaken by HLS U.S.A. was the critical neoplasm for discussion. The Committee agreed that the finding of liver tumours in F344 rats and B6C3F1 mice in the modern studies should also be evaluated in detail. The findings reported in earlier studies undertaken in the 1970s' and 80s' contributed little weight of evidence to the assessment of malathion.

52. Members agreed that tumours of the nasal tissue in F344 rats were rare and that there were difficulties in assessment particularly with regard to the adequacy of historical control data from the laboratory. Members noted the additional sectioning and microscopic evaluation of additional slides from the HLS U.S.A. suggested that the U.S. NTP historical control data for F344 rats might be more appropriate. Overall it was agreed that following a Peer review, that a tumourigenic response had been documented in this study. Members noted that the observed tumours were all benign and included (an olfactory epithelial adenoma arising from the Bowmans' gland in a male fed 6000 ppm and a respiratory epithelial adenoma in a male fed 12000 ppm. In females one animal at 6000 ppm and one at 12000 ppm had respiratory epithelial adenomas. All tumours were well defined and there was no evidence of pleomorphism or atypia. The Committee considered that the significance of these tumours had to be assessed against a background of severe inflammation, which exceeded that found in inhalation carcinogenicity studies with chemicals that were directly irritant to the nasal tissue. It was unclear from the HLS malathion study in F344 rats to what extent inhalation of food particles contributed to the induction of this inflammatory response in this study. Members agreed that direct irritant effects on the nasal passages was possible when animals were fed powedered diets. It was noted that the food pots used in the malathion study had some degree of covering which might have limited inhalation of food particles. It was noted that there was clear evidence to show the effect of food particles in the induction of localised inflammation of the nasal passages and the lungs in the malaoxon bioassay in F344 rats undertaken at the same laboratory as the critical malathion F344 rat bioassay. Members felt that the proposal from the data holder that de-esterification of malathion in nasal tissue to form acids was also potentially plausible. It was noted that nasal tissue would have appropriate metabolic capacity to metabolise malathion to its corresponding diacid metabolites. The Committee discussed the possibility of a genotoxic mechanism in the induction of nasal tumours. However the weight of evidence including the formation of tumours only at excessive doses, the evidence for severe prolonged localised inflammation in the target tissue, the lack of multi- organ response suggested that a non-genotoxic mechanism was probable in this instance.

53. In further discussion, members were asked by MCA to clarify the conclusion with regard to mechanism of tumours in the nasal tissue. Members heard that COM had provisionally considered that malathion had potential to induce mutations in bone marrow of mice following intraperitoneal administration but would consider further published data and information submitted by the data holder at the next COM meeting of 10 October 2002. The Chair asked if representative members of COC could also attend the COM meeting in October to report on the COC conclusions with regard to the mechanism of nasal tumours seen in F344 rats.

54. The Peer review pathology report of the Huntingdon Life Sciences bioassay in F344 rats conducted between 1993 and 1996 reported a consensus incidence of 5/70 (7.1%) hepatocellular adenomas in females at a dietary level of 12000 ppm (ca 868 mg/kg bw/day). The historical control incidence for the laboratory was 0-5.4% ( n= 254). The Peer review report noted that there were a number of non-neoplastic microscopic findings in the five animals with tumours, which included hypertrophy, congestion and vacuoloation. The COC agreed that the time of death and liver pathology seen in these animals would appear to be generally consistent with ageing pathology in this strain of rat. There was a significant increase in mortality in the female high dose group from month 18 to 24 (above that seen in controls) resulting in limited numbers of animals available at terminal sacrifice and thus the possibility existed that there might be some animals which could have developed liver tumours being missed. However there was no evidence of a dose response in females and only one hepatocellular adenoma was found in treated males (in a high dose male). All the available pathology evidence suggested that these tumours are most likely to be an effect of malathion on the background pathology of ageing F344 rats and are unlikely to be of significance for human health.

55. Members noted that the increased incidence of hepatocellular adenomas in male and female B6C3F1 mice was associated with clearly overt increases in liver weight and hepatocellular hypertrophy and are unlikely to be significant to humans. It was suggested that, given the evidence for hepatocellular hypertrophy induced at doses exceeding the MTD, that it was highly likely that these tumours were induced through a non-genotoxic mechanism of technical grade malathion.

56. Members agreed that there was no evidence for a carcinogenic effect of malaoxon, the main metabolite of malathion and that this conclusion should also be reported back to COM.

57. The COC drew the following overall conclusion to be forwarded to the COM at its next meeting in October. Members agreed that technical grade malathion had been tested in four long-term dietary bioassays in rats and two long-term dietary bioassays in mice. The most recent studies undertaken in F344 rats (1993-96) and in B6C3F1 mice (1992-94) were adequate for the evaluation of carcinogenicity. There is evidence for tumourigenicity in the nasal tissue and liver (females only) of F344 to rats fed high doses of malathion. The nasal tumours were associated with severe ongoing inflammation, which is most likely involved in the mechanism of tumourigenesis. There was evidence for liver tumours in female F344 rats and male and female B6C3F1 mice. The weight of evidence suggested that these liver tumours are not relevant to human health assessment.

58. Members agreed to consider a draft statement to be circulated by post before the COM meeting and to attend and discuss all of the available data with the COM.

ACTIONS

ITEM	ACTION	WHO
4. Review of malathion	Draft statement/report conclusions to COM	Secretariat
6. Revision of COC Guidelines	Draft 1st draft guidelinesr	DH Toxicology Unit

Top

copyright: © | updated: 28 March 2003