Scientific analysis of the proposed uses of the T25 dose descriptor in chemical carcinogen regulation - An ECE TOC workshop overview

The uncertainties that surround the methods used for risk assessment of exp osure to carcinogens have been highlighted by a recent document advocating an approach based on the T25 dose (the dose giving a 25% incidence of cance r in an appropriately designed animal experiment). This method relies on de rivation of the T25 dose then assesses risk at the exposure dose using prop ortionality provided by a linear extrapolation (T25/ linear). To promote di scussion of the scientific issues underlying methods for the risk assessmen t of chemical carcinogens, the European Centre for Ecotoxicology and Toxico logy of Chemicals (ECETOC) hosted a one-day workshop in Brussels on 10 Nove mber 2000. Several invited presentations were made to participants, includi ng scientists from regulatory authorities, industry and academia. In genera l, it was felt that there was sufficient basis for using the T25 dose as an index of carcino-genic potency and hence as part of the hazard assessment process. However, the use of the T25 in risk assessment has not been valida ted. The T25/linear and other extrapolation methods based on metrics such a s LED 10 assume linearity which may be invalid. Any risk calculated using t he T25/linear method would provide a precise risk figure similar to the out put obtained from the Linearised Multistage (LMS) method formerly used by t he Environmental Protection Agency (EPA) in the United States of America. S imilarity of output does not provide validation but rather reflects their r eliance on similar mathematical approaches. In addition to the T25 issue, e vidence was provided that using two separate methods (linearised nonthresho ld model for genotoxic carcinogens; no-observable-effect level with a safet y factor (NOEL/SF) method for all other toxicity including non-genotoxic ca rcinogens) is not justified. Since the ultimate purpose of risk assessment is to provide reliable information to risk managers and the public, there w as strong support at the workshop for harmonisation of approaches to risk a ssessment for all genotoxic and nongenotoxic carcinogens. In summary, the T 25 method has utility for ranking potency to focus efforts in risk reductio n. However, uncertainties such as the false assumption of precision and non -linearity in the dose-response curve for tumour induction raise serious co ncerns that caution against the use of T25/linear method for predicting hum an cancer risk.