Some implications for quantitative risk assessment if hormesis exists.

The existence of hormesis should impact quantitative risk assessment in at least seven fundamental ways. (1) The dose-response models for bioassay and epidemiological data should have greater flexibility to fit the observed shape of the dose-response data and no longer be forced to always be linearly increasing at low doses. (2) Experimental designs should be altered to provide greater opportunity to identify the hormetic component of a dose-response relationship. (3) Rather than a lifetime average daily dose or its analog for shorter time periods, dose scales or metrics should be used that reflect the age or time dependence of the dose level. (4) Low-dose risk characterization should include the likelihood of beneficial effects and the likelihood that a dose level has reasonable certainty of no appreciable adverse health effects. (5) Exposure assessments should make greater efforts to characterize the distribution of actual doses from exposure rather than just upper bounds. (6) Uncertainty characterizations should be expanded to include both upper and lower bounds, and there should be an increased explicit use of expert judgement and weight-of-evidence based distributional analyses reflecting more of the available relevant dose-response information and alternative risk characterizations. (7) Risk should be characterized in terms of the net effect of a dose on health rather than a dose's effect on a single factor affecting health - for example, risk would be better expressed in terms of mortality from all causes combined rather than a specific type of fatal disease.

Challenges to default assumptions stimulate comprehensive realism as a new tier in quantitative cancer risk assessment.

The current practice in carcinogen risk assessment of using a linearized multistage model and assuming low-dose linearity is based on several false premises. In many cases linearity at low doses would not be expected based on the interaction between the multiple components in the carcinogenic process. The two-stage growth models, involving multiple mutations and cell birth and death rates, provide one means of exploring these interactions. In addition, if carcinogenesis is considered to be the imbalance between invading substances and defense mechanisms, then the cancer probability depends on how much the substance increases or decrease the number of defenders or their efficiency as well as increasing or decreasing the number of invaders. Challenges to low-dose linearity and other default assumptions have stimulated the development of new risk assessment methodologies as have the need for more realistic estimates of risk, better uncertainty characterization, and greater utilization of cost-benefit analyses, and other tools for risk management decision making. "Comprehensive realism" is an emerging quantitative weight-of-evidence risk assessment methodology which is designed to reflect all of the relevant and available information and the current state of knowledge about the health risks associated with a substance.

Incorporating additional biological phenomena into two-stage cancer models.

Several difficulties arise in the implementation of the Armitage-Doll-type carcinogenesis model. Some of these difficulties have been briefly discussed herein, with particular attention to the recent finding that the magnitude of current animal-based bounds on human cancer potency is influenced by animal background transition rates. Although the ideal method of accounting for the background transition rates necessitates more direct biological information, this paper offers risk assessors some alternatives for improving current quantitative cancer potency assessment. Computer software has been developed to facilitate the use of not only the multistage model but also other dose-response models, including the following: 1) tolerance models (probit, logit, multihit, and Weibull), which assume that a distribution of tolerances exists in the population and that when a tolerance is exceeded, a carcinogenic response occurs; 2) multihit models, which assume that a carcinogenic response occurs when a tissue receives more than a specified number of hits; 3) Armitage-Doll multistage models, which explicitly model the time and/or dose dependence of each transition stage; 4) time-to-response extensions of the multistage model; and 5) extensions of the multistage model including cell proliferation (Lu and Sielken 1991, Holland and Sielken 1993). In any of the dose-response models, the dose scale is not restricted to the administered dose scale. Computer software that facilitates the use of more biologically relevant dose scales (such as delivered and biologically effective dose scales) is now available. Furthermore, the differences between species and routes of exposure in the amount of delivered or biologically effective dose corresponding to a particular administered dose can be incorporated. The dependence of an individual's dose on not only the individual's administered dose but also the individual's background dose and susceptibility can be incorporated as well. Hence, interindividual variability in dose levels and probabilities of a carcinogenic response can be considered (Holland and Sielken 1993). The possibility also exists of incorporating more cell biology into carcinogenesis theory using the MVK two-stage model. In this model, cell proliferation parameters are included, in contrast to models of the Armitage-Doll family. The MVK model offers increased potential for obtaining estimates of low-dose behavior that reflect more of the available biological data, including microdosimetry data. A Monte Carlo simulator for this model can be used in cases in which more complex biological mechanisms are included, information other than the means of population distributions is desired, and information on subpopulations is desired.(ABSTRACT TRUNCATED AT 400 WORDS)

Policy principles for utilizing science in decision-making on chronic health issues.

Scientific advances will continue to contribute to our understanding of latent chronic diseases related to chemical exposure. Regulatory agencies must deal with a complex matrix of emerging scientific information, a diversity of potential risk situations, and a variety of statutory prescriptions for protecting public health. Seven policy principles are proposed for facilitating integration of the latest scientific thought into the administrative decision-making process. The principles relate to distinguishing between risk assessment and risk management, analysis of all relevant information in developing a risk assessment, consideration of weight-of-the-evidence and more probable than not criterion on key assumptions, scientific peer review of assessments, scoping scientific input appropriately with the nature of a specific regulatory activity, emphasizing research which enhances the basis of risk assessment, and education and communication on risk matters. The policy principles are interdependent; collectively they need endorsement and promotion by the scientific and regulatory communities and by policy leaders in federal and state governments in the interest of establishing a framework for further improving the basis of critical decisions for protecting public health.