The regulation of carcinogenic hazards

An Analysis of the Carcinogenesis Testing Experience of the National Cancer Institute

Tabulations of the results of the initial approximately 200 NCI carcinogenesis bioassays conducted in mice and rats were published recently. However, no analysis of the results were provided. Such an analysis is the subject of this report.

Ninety eight NCI bioassays were positive for carcinogenicity. Only three bioassays (<2%) were negative in both species and the ratio of positive to negative bioassays was 33:1. The remaining 47% of the bioassays were considered to be suggestive or inconclusive and it is inappropriate to refer to the total positive rate as 50%. The liver was the predominant tumor site and the only tumors produced in either rodent species in 19% of the positive bioassays were in the livers of mice. The high ratio of positive to negative results and the use of the mouse liver as a discriminating target organ raise serious concerns about the carcinogenesis protocol and in the methods of evaluating the results. Several recommendations are offered which may be useful in the current reexamination of bioassay procedures.

Estimating the magnitude of carcinogenic effects in long-term animal studies

Carcinogenicity studies seek to compare the incidence of tumours in animals exposed to the substance under investigation and animals used as controls. The conventional method of analysis is the Peto test, which assumes that tumours are either instantly fatal or have no effect on mortality and requires a judgement to be made regarding the lethality of each tumour. Such an assumption seems unrealistic and the judgement is often difficult to make and unreliable. The need for such a judgement and the assumption of extreme lethality can be removed by using parametric multi-state models. In this modelling approach the transition of animals between the states 'alive without a tumour', 'alive with a tumour' and 'dead' is modelled mathematically. This paper compares the Peto test with tests based on two parametric multi-state models in terms of the sensitivity of the tests to detect carcinogenicity. The sensitivity, or power, is shown to be low for commonly used numbers of animals, depending chiefly on the expected total number of animals with tumours. The Omar and Whitehead multi-state model is found to be slightly more powerful than the Dewanji et al. model and at least as powerful as the Peto test. Provided the parametric assumptions are appropriate, this method thus gives a test that is more sensitive than the Peto test and enables estimation of tumour onset and mortality rates without the requirement of tumour lethality judgements.

Pharmacokinetic data in the evaluation of the safety of food and color additives

Safety evaluation of food and color additives intended for human use is usually based on toxicity data obtained from animal studies; human data are rarely available. The extrapolation of animal data to humans is often controversial. The important role that pharmacokinetic data could play in the safety evaluation of food and color additives is now widely recognized. This paper reviews the current scientific knowledge concerning the application of properly designed pharmacokinetic studies to the evaluation of the safety of food and color additives. In principle, pharmacokinetic data can be useful not only in designing, interpreting, and extrapolating animal toxicity studies to humans, but also in providing insight into the mechanisms of toxicity. Examples of such applications are provided.

Evaluating the safety of food and color additives with pharmacokinetic data

Pharmacokinetic studies are designed to quantify, as a function of time, the processes associated with absorption, distribution, metabolism, and excretion of a chemical in experimental animals or in humans. Such studies have played an important role in drug safety evaluation and could be very useful in the safety evaluation of food and color additives. This presentation provides an overview of the potential use of metabolic and pharmacokinetic data in the design and evaluation of toxicological studies and in the assessment of the potential hazard to humans from exposure to food or color additives.

Adjudicating cancer causation: scientific, political, and legal conflicts

Lawsuits concerning cancer causation resort to scientific argumentation. Yet, the apparent ambiguities of science confuse the courts, the juries, and the public. This is especially so with regard to official regulatory definitions of cancer causation that carry the weight of law. At the heart of this problem is a prevailing misunderstanding of science and the scientific method, and of the limits of current scientific knowledge about cancer. Moreover, current regulatory policies encourage the public to perceive official cancer risk assessments as if they were scientifically derived and accepted, even though official fine print readily admits they are not. Some recent court decisions have begun to recognize these difficulties with a body of precedent, and this may result in future rulings influenced more by objective appraisals than by reliance upon official but contingent assumptions.

The risks of handling cytotoxic drugs. I. Methods of testing exposure

Results of various biological and physical/chemical tests of the urines or blood of health-care personnel working with cytotoxic drugs are discussed. The outcomes of these tests are conflicting and inconclusive. The physical/chemical tests seem to be an alternative method. However, until now it has not been possible to establish the threshold concentration in urine or blood beneath which no effect has to be expected. Therefore, the interpretation of the concentration of cytotoxic drugs and/or the metabolites in the urine or blood is difficult. As long as one will not be able to provide conclusive data on the health hazards when working with cytotoxic drugs, protective measures have to be taken in order to lower the risk as much as possible.

Hazard evaluation for complex mixtures: relative comparisons to improve regulatory consistency

The traditional "absolute decision-making" process used by federal regulatory agencies to derive permissible exposure concentrations for hazardous substances is initiated by an evaluation of the "weight-of-evidence" that a substance is a potential human carcinogen. Subsequent conservative procedures applied variably to noncarcinogens and carcinogens yield exposure limits for individual substances based on "data-sparse, model-intensive" techniques which may lack consistency and have difficulty directly addressing the hazards from complex mixtures. This paper describes how a "relative decision-making" technique applicable to complex mixtures can supplement the "absolute" approach currently used. Estimates obtained through this "data-intensive, model-sparse" technique may be evaluated by comparisons to estimates representing a range of hazards "generally regarded as safe" derived through analyses of chlorinated drinking water, cigarette smoke condensate, and other common human exposures. Comparisons are also used to evaluate the relative degree of consistency in risk estimates between 58 suspect human carcinogens analyzed by the U.S. Environmental Protection Agency Carcinogen Assessment Group and by the authors.

Hormesis, Gompertz functions, and risk assessment

A historical survey of the literature indicates that benefits derived from low doses of toxic substances have been reported over many centuries. Hippocrates, Paracelsus, Arndt, Schulz, and Hahnemann (founder of homeopathy) have all reported that low doses of toxic substances may be "stimulatory" or otherwise beneficial. Assessment of mortality data from modern-day bioassay studies indicates that low-dose animal exposure to a variety of toxic agents can, through an unknown mechanism, also induce beneficial changes which promote health and prolong life (longevity hormesis). This nonspecific and apparently reversible phenomenon has been modeled kinetically through use of age-specific mortality rate and a generalized Gompertz function; the basic assumption is that mortality in an interval is a function of the weighted sum of intensities of physiologic injury during that interval. It was assumed that longevity-enhancing hormetic reduction in population injury may be decremented from life-shortening injury produced through the aging process and concomitant toxicity. At low exposure levels, a net reduction in age-specific mortality rate can sometimes be observed. The implications for risk assessment are significant. It is tacitly assumed in generating virtually all estimates of risk that toxic manifestations observed at higher doses are the sole effects elicited at lower doses. This appears to be qualitatively incorrect.

The calculation and use of carcinogenic potency: a review

Six methods of estimating carcinogenic potency now in use to some extent by regulatory agencies are examined. It is concluded that none of these methods is adequate for regulatory decision making when used alone, as is usually the case, but that all are useful as a contribution to the whole risk assessment process. The needs for further development and improvement of these methods, where appropriate to human risk, are identified and discussed in view of the growing use of potency estimates in regulatory decision making.

Permissible concentrations of chemicals in air and water derived from RTECS entries: a "rash" chemical scoring system

Many chemicals are of concern to human health, but only a few have epidemiologically derived risk estimates. About 45,000 chemicals are listed in RTECS, most of which have had some testing in subhuman models. RTECS entries range from cellular effects through organoleptic damage to lethality, with many pathological endpoints listed, including mutagenic changes, irritation, teratogenesis, cancer, mortality, etc. However, it is difficult to extend any biological test results to human risk assessments. If the results are extended, the degree of validity is highly uncertain. This paper describes a logical basis for using the entire complex spectrum of test results to evaluate the overall toxicological potency of a chemical to be assayed (i.e., an interviewing chemical) and describes how to derive tentative, permissible concentrations in air and water for any particular chemical for which no regulatory guidance exists. This approach has been tested for 16 reference chemicals discussed in NIOSH Criteria Documents, EPA-CAG reports, etc. The evaluations are uncomplicated, but occasionally it is difficult to match RTECS entries for two different chemicals. Difficult comparisons may require some familiarity with experimental design and the toxicological literature. One important product of this novel approach is that a distribution or array of potency values is obtained for any chemical evaluated. This distribution reflects many uncertainties stemming from low statistical power, experimental design, pharmacological processes, interspecies variability, dose rate, biological effect monitored, route of treatment, etc. The array of relative values for a particular chemical reflects many different biological and physical conditions. The distribution of the array helps to index a composite toxicological profile for many different biological effects resulting from numerous treatment protocols. To minimize the effect of extreme sensitivity of certain (perhaps novel) biological test models, possible errors in the RTECS data-base, and possible human pharmacological insensitivity to a particular chemical and/or a particular route of administration, we consider the interquartile range (i.e., the central 50%) of the array of relative potency values between two chemicals being compared as a practical measure of uncertainty. Thus, the range in response derived from variability in relative potency should be useful in addressing the range of response in man as estimated from extrapolations of test data.

Carcinogenicity testing of antitumor agents

Carcinogenicity testing of antitumor agents in animal bioassays has been proposed because of the potential for carcinogenicity of this class of agents and the expectation that such testing may indicate prospectively the target organs of any related human oncogenesis. The literature reveals the anticipated confirmations in animals of the carcinogenicity of many antitumor agents. Furthermore, these agents have been associated with human tumors in numerous case reports. Review of the literature also indicates the inability of animal studies to predict the sites of carcinogen-induced tumors in man. The carcinogenic risk assessment of antitumor agents should begin with the determination of the ability of the agent to interact with DNA. Those agents which are capable of alkylating or binding DNA should be tested for mutagenic and teratogenic potential. The presumption of carcinogenicity should be made for DNA-reactive, mutagenic/teratogenic antitumor agents without requiring confirmation in long-term carcinogenicity bioassays in large numbers of animals. The inability of carcinogenicity studies in animals to accurately predict potential human tumor sites must also be recognized.

The power and interpretation of the carcinogenicity bioassay

Carcinogenicity is a major consideration in the assessment of risks due to environmental chemicals. The carcinogen bioassay therefore is a very important component of the battery of toxicological tests used in hazard evaluation. The strengths and limitations of this bioassay are discussed with emphasis upon the unresolved practical considerations, the interpretation of negative results, the significance of tumors induced in the presence of a high background incidence of naturally occurring tumors, and the difficulties in transspecies extrapolation. These factors, in combination with consideration of the biological mechanisms of chemical cancer induction, will be valuable in assessing the potential risk to man posed by individual chemicals.

Comparative risk analysis of technological hazards (a review)

Hazards are threats to people and what they value and risks are measures of hazards. Comparative analyses of the risks and hazards of technology can be dated to Starr's 1969 paper [Starr, C. (1969) Science 165, 1232-1238] but are rooted in recent trends in the evolution of technology, the identification of hazard, the perception of risk, and the activities of society. These trends have spawned an interdisciplinary quasi profession with new terminology, methodology, and literature. A review of 54 English-language monographs and book-length collections, published between 1970 and 1983, identified seven recurring themes: (i) overviews of the field of risk assessment, (ii) efforts to estimate and quantify risk, (iii) discussions of risk acceptability, (iv) perception, (v) analyses of regulation, (vi) case studies of specific technological hazards, and (vii) agenda for research. Within this field, science occupies a unique niche, for many technological hazards transcend the realm of ordinary experience and require expert study. Scientists can make unique contributions to each area of hazard management but their primary contribution is the practice of basic science. Beyond that, science needs to further risk assessment by understanding the more subtle processes of hazard creation and by establishing conventions for estimating risk and for presenting and handling uncertainty. Scientists can enlighten the discussion of tolerable risk by setting risks into comparative contexts, by studying the process of evaluation, and by participating as knowledgeable individuals, but they cannot decide the issue. Science can inform the hazard management process by broadening the range of alternative control actions and modes of implementation and by devising methods to evaluate their effectiveness.

Toxicological evaluation of malonaldehyde: a 12-month study of mice

The chronic toxicity of malonaldehyde (MA) was evaluated using Swiss female mice. Beginning at 10 wk of age, MA was administered in the drinking water for 12 mo to groups of 50 animals at levels of 0.1, 1, and 10 micrograms/g body wt.d with 100 controls. The highest dose was associated with increased mortality (28% versus 12-14%). MA had no effect on body weight, organ weight, hematological indices, or the incidence of lesions in 27 tissues examined. More liver lesions were observed in the three treatment groups than in the controls (p less than 0.05), and the histopathologic scores for severity of lesions were significantly increased in the groups that received the two higher levels of MA. The liver lesions included anisokaryosis, changes in cytoplasmic volume with architectural derangements, necrosis and neoplastic changes (nodular hyperplasia, hepatoma, and hemangioma). There was no significant increase in specific neoplasms in the treated groups, but the incidence of total neoplasms and neoplastic lesions was dose-dependent (4%, 8%, and 12%, respectively) (p less than 0.01). There was only one neoplasm (a hemangioma) among the controls (1%). Three animals (6%) given the highest dose of MA developed stomach neoplasms. In terms of human dietary exposure to MA, the lowest level of MA used in this study is about 10 times the estimated average daily intake of MA by the Canadian population on a body weight basis.