Click for next page ( 472


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 471
Perspectives The objective of risk assessments is to provide society with estimates of the likelihood of toxic injury in human populations as a consequence of exposure to naturally occurring or man-made chemicals. Fundamen- tally, the assessment procedure involves the development of a hypothetical dose-response curve for an untested species (humans) on the basis of exposure studies in laboratory animals followed by projection of the "curve" to estimate the levels of exposure that may be considered safe. In practice, the risk assessor is faced with the problem of predicting the likelihood of injury to humans exposed to low concentrations of ma- terials from data gathered in studies of animals receiving high doses, often by a different route of administration. The complexity of this three-way extrapolationdose, species, and route presents a formidable chal- lenge. Not the least of the risk assessor's concerns is the problem of relating an applied "dose" to the "dose" of the active material in target tissues. Traditionally, "dose" has been understood to mean the amount of test material administered to a test animal per unit time. As discussed extensively by various speakers in this workshop, the toxic response of an animal is related to exposure through two sets of factors: pharmaco- k~neticfactors, which govern the concentration in target organs and the interaction of the material with the site of action; and pharmacodynamic factors, which govern the sequence of events that result from the inter- action and lead to the manifestation of the toxic response. Clearly, if it were accepted that the severity of a toxic response depends on the con- centration or the area under the curve (AUC) of the toxicant at its site of 471

OCR for page 471
472 PERSPECTIVES action, "dose" could be defined in terms of these pharmacokinetic pa- rameters. Extrapolation across dose, species, and route could then be performed based on the actual concentration or AUC of the toxicant at the site of action, rather than on the applied dose. The task of the assessor would then be greatly eased, and the prediction would be much improved. The rapid development of computer hardware and software during the last decade has markedly increased the diversity of the mathematical tools applicable to the pharmacokinetics of environmental chemicals. In par- ticular, computers have permitted the application of physiologically based pharmacokinetic (PB-PK) models to define "dose" in pharmacokinetic terms. As was amply demonstrated in the workshop, relevant PB-PK data can be used to reduce uncertainty in extrapolation and risk assessment. It is now theoretically possible to simulate concentrations of virtually any xenobiotic substance or any of its metabolites in any organ, regardless of the complexity of the systems. PB-PK analysis has particular value when absorption of inhaled or ingested material might be incomplete and when capacity-limited enzymatic processes can cause nonlinearity between in- gested dose and AUC of the parent material in blood and various organs. An improved definition of dose is especially useful when toxicity results directly from the action of a parent material. Moreover, allometric methods for extrapolation between species are most likely to be valid when such toxic materials are eliminated predominantly by exhalation or urinary excretion. It was also made clear in the workshop that the potential power of the approach is matched by potential pitfalls and hence that simplification of extrapolation of dose, species, and route must be based on an understand- ing of both general principles of toxicology and specific information on the mechanism of the toxicity under consideration. The pharmacokinetics of toxic metabolites can be much more complex than those of the parent substance. Various speakers made it clear that in such cases, supplemen- tary information on the mechanisms of action must be provided before pharmacokinetics can be usefully applied to risk assessment. Unfortu- nately, in current practice such critical information is often missing from the results of pharmacokinetic studies presented to risk assessors, and the usefulness of pharmacokinetic studies is correspondingly limited to esti- mation of the bioavailability of the parent material. Problems in applying pharmacokinetics to risk assessment tend to fall into three broad categories: inadequacies of the present data base, incomplete definition or understand- ing of underlying processes, and inherent uncertainties in the modeling process. Inadequacies of the data base are largely correctable with present tech- nology. As emerged in many of the discussion sessions of the workshop, refinement and adaptation of the protocols for both pharmacokinetic and

OCR for page 471
Perspectives 473 toxicologic procedures will be needed to allow fruitful comparisons, es- pecially when chronic administration of the test compound might induce or impair the enzymes responsible for its elimination. Because such changes are not always predictable, it would be advantageous to incorporate phar- macokinetic studies into the toxicologic protocols and, in the case of chronic exposure, to repeat the pharmacokinetic studies at intervals throughout the exposure period. If it is not feasible to integrate the phar- macokinetic and toxicologic studies directly, it will be important to ensure that the pharmacokinetic studies are done under toxicologically relevant conditions. It will also be important to develop protocols for toxicity studies that would show whether toxicity is caused solely by the parent substance or by its metabolites. Much work remains to be done to collect and improve estimates of kinetic parameters and to determine how they vary within and among human and animal populations. Answers to many of the questions in this category can be expected to accumulate as ex- perience with PB-PK risk assessment grows and as the integration between pharmacokinetic and toxicologic studies improves. Appreciation of problems in understanding the underlying processes can also be expected to grow as experience in the use of PB-PK modeling in risk assessment accumulates. Some of the problems can be predicted from present knowledge; others will appear as our sophistication increases. Knowledge of the kinetics of the interaction of biologically active sub- stances with putative sites of action and of how physiologic processes modify the interactions is sparse. As it increases, we will probably be able to improve pharrnacokinetic and risk-assessment equations. In ad- dition, the role of pharmacodynamic factors needs to be defined. Examples of problems in this category include differences in longevity between test animals and human species, differences in DNA repair rates, differences in transformation frequencies, and differences in mechanisms of "pro- motion." Those problems cannot be addressed by the use of present procedures for risk assessment and have necessarily been largely ignored. Research stimulated by the opportunities and challenges posed by PB-PK assessment, however, could reduce uncertainties associated with the con- cept of "dose" and thereby allow us to focus on the roles of pharma- codynamic factors, which then may be incorporated into more detailed models. A lack of correlation between the concentration in target tissues and response would be of special interest to toxicologists, because it would disclose major interspecies or intraspecies differences in the pharmaco- dynamics of the toxicity, which may profoundly affect the validity of risk assessment. As to inherent uncertainties in the modeling process, some factors are unlikely to be completely validated soon. Biologic systems are highly complex, and numerous factors influence the effective rate constants that

OCR for page 471
474 PERSPECTIVES describe the metabolic processes. Modeling consists of simplifying the system by selecting steps whose estimated rate constants will allow the best fit of a mathematical model to the biologic reality. Clearly, the more detailed the model, the greater will be the theoretical utility, but also the more numerous will be the sources of uncertainty and the data necessary to estimate values of the parameters of the models. In practice, a com- promise between complexity and simplicity must be reached. It is crucial to develop procedures for identifying the critical parameters in these com- plex models (i.e., by performing sensitivity analysis) and to develop ap- propriate statistical tools for estimating the parameters and the confidence limits of these estimates in complex models. The latter task is complicated. Estimates of some of the parameters can be obtained from specific ex- perimental data (whose experimental error can be directly estimated), but others are based on collective experience from a variety of sources (such as Bayesian input as to the prior "believability" of the values estimated by the computer during optimization). Several participants dealt with these topics and identified them as important for the development of pharrna- cokinetic modeling in risk assessment. Perhaps the most important idea that emerged in the workshop was the need to integrate mechanistic information with pharmacokinetic studies and with the risk assessment process. For example, it is essential to know whether toxicity results from the action of a parent material or from its metabolites. If the ultimate toxicant is a metabolite, it is important to obtain information on the biologic disposition and stability of the metab- olite and on the pharmacokinetic interrelationships among organs of for- mation, organs of elimination, and target sites. Availability of appropriate mechanistic information markedly increases the power of PB-PK analysis and the scientific credibility of extrapolation across doses, species, and routes of administration. As illustrated by many participants, the incor- poration of toxic active/reactive metabolites into PB-PK models compli- cates the analysis. But the insight that can be gained from the integration of knowledge of the pharmacokinetics and the mode of action of toxic metabolites will focus attention on the dominant factors that govern the manifestation of different toxicities, and that attention will in turn lead to the identification of other important risk factors. Appreciation of the roles of such risk factors is likely to lead to recognition of highly sensitive subpopulations that are not discernible by present testing processes. The workshop brought together a large group of people from diverse backgrounds who share an interest in the application of PB-PK in risk assessment. The size of the audience and the extent of discussions, both formal and informal, indicate a high degree of interest in the critical evaluation of present and future roles of PB-PK analysis in risk assessment. Clearly, pharmacokinetic studies will have a large role in the elucidation

OCR for page 471
Perspectives 475 of mechanisms of toxicity and of the dominant factors that account for intraspecies and interspecies differences in the incidence of toxic effects. Even in the absence of knowledge of the mechanisms of toxicity, phar- macokinetic studies will be essential for assessing the bioavailability of materials subjected to toxicity tests. As was well illustrated by this workshop, PB-PK analysis has genuine promise for improving risk assessment. The basic tools are available. It remains for us to use and improve them. In doing so, we shall explore the strengths and weaknesses of PB-PK analysis, evaluate how such anal- ysis may improve the process of risk assessment, and identify the areas of research that will lead to further improvement of our procedures and concepts.

OCR for page 471