Introduction

1. The COC has adopted a prudent approach to the assessment of chemical carcinogens which assumes that genotoxic carcinogens have the potential to damage DNA at any level of exposure and that such damage may lead to tumour development. Thus for genotoxic carcinogens it is assumed that there is no discernible threshold and that any level of exposure carries a risk. The general advice of the COM when considering the risk assessment of chemicals which are mutagenic in-vivo has been that it is prudent to assume a linear, non threshold dose response. Thus it is assumed that any exposure to an in-vivo mutagen is associated with some damage to DNA and consequently an increased risk of mutation leading to an increased risk of adverse health effects albeit that this may be small. In such instances the Committee has recommended that exposures be reduced to a low as is reasonably practicable. 1,2 The Expert Panel on Air Quality Standards (EPAQS) provides advice to U.K. Government Departments and Agencies on air quality issues, in particular the levels of pollution at which no or minimal health effects are likely to occur. (http://www.defra.gov.uk/environment/airquality/aqs/index.htm). EPAQS has adopted a pragmatic approach to developing air quality standards for air pollutants which are genotoxic carcinogens.3 This does not involve the use of mathematical models to estimate cancer risks because of the concerns of COC regarding such approaches.2,3

2. The Department of Health asked the COC for advice at its November 2002 meeting on whether it was feasible to identify a specific approach to modelling dose-response to carcinogenic air pollutants. The suggested approach would be restricted to modelling of data derived from acceptable epidemiology studies in exposed human populations (predominantly occupationally exposed cohorts). The derived dose-response curves could be then be used following the application of the most conservative model (considered to be the one-hit model)4 to estimate the maximum upper bound risk of cancer at environmental exposure levels. There was no intention to use data derived from long-term carcinogenicity bioassays in rodents for quantitative estimation of risks to humans. Agreement to the proposed approach would be of value to the Department in formulating advice on prioritising measures for reducing levels of air pollutants.

COC consideration of dose-response modelling for carcinogenic air pollutants

3. The Committee restated its views, published in its guidelines in 1991, that use of mathematical models to evaluate the dose-response for carcinogens, namely that extrapolation of the dose-response curve below the lowest experimental data points, taken from animal bioassay data, gave an impression of precision which cannot be justified from the approximations and assumptions used.2

4. The Committee considered that it might be acceptable to consider an approach based solely on use of epidemiological data from investigations considered to have been adequately undertaken. The COC cautioned that quantitative risk estimates based on extrapolation to dose levels of one or more orders of magnitude below the observed dose-response in epidemiology studies would involve uncertainties in that the shape of the dose-response curve could not be predicted with any degree of accuracy.

5. However the Committee noted that the one-hit model4,5 assumed that a single mutation arising from the interaction of one molecule of carcinogen with DNA lead to the development of cancer. This was likely to represent the most conservative approach to risk assessment. Thus the COC agreed that the upper bound (95% Confidence Interval) cancer risk estimate from this approach was likely to overestimate the actual risk associated with exposure to the carcinogen. The Committee agreed to this approach to carcinogenic air pollutants provided the risk estimates were used in the consideration of risk management options rather than being considered as definite values relating to cancer risk levels. The Committee considered it important to restate that any exposure to a genotoxic carcinogen carried a small, but unquantifiable, risk of cancer.

Conclusion

6. The COC agreed to a proposal from the Department of Health, that the upper bound (95% Confidence Interval) cancer risk estimate at environmental exposure levels to air pollutants, which are genotoxic carcinogens, based on data from adequately performed cancer epidemiology studies, using the one-hit model for dose-response extrapolation, could be used as an aid in deriving risk management strategies. The COC agreed it important to restate that exposure to any level of a genotoxic carcinogen carried a small, but unquantifiable, risk of cancer.

September 2003

References

1. Annual Reports of the Committees on Toxicity, Mutagenicity and Carcinogenicity of Chemicals in Food , Consume Products and the Environment. www.doh.gov.uk/com/index.htm (In particular see Annual reports for 1998 and 1999).

2. Department of Health (1991). Report on health and Social Subjects. No 42. Guidelines for the Evaluation of Chemicals for Carcinogenicity. Committee on Carcinogenicity of Chemicals in Food, Consumer Products and the Environment, London HMSO.

3. Maynard RL, Cameron KM, Fielder RJ, McDonald A and Wadge A (1995). Setting air quality standards for carcinogens; an alternative to mathematical quantitative risk assessment: Discussion paper. Human and Experimental Toxicology, 14, 175-186.

4. ECETOC* (1996). Risk assessment of carcinogens. ECETOC Monograph 24. Avenue E Van Nieuwenhuse 4, B- 1160, Brussels, Belgium. (http://www.ecetoc.org/entry.htm, *European Centre for Ecotoxicology and Toxicology of Chemicals.)

5. Cothern CR. (1986). Techniques for the assessment of carcinogenic risk due to drinking water contaminants. CRC Crit Rev in Environmental Control 16, 357-399.

Top

copyright: © | updated: 26 September 2003

	Search
	Entire site

Download Adobe Acrobat reader for PDF files

Help with Portable Document Format (PDF)