REGULATORY

TOXICOLOGY

AND

PHARMACOLOGY

14, 78-87 (199 1)

A Critique of the Use of the Maximum Tolerated Dose in Bioassays to Assess Cancer Risks from Chemicals C. JELLEFF CARR”* AND ALBERT C. KOLBYE, JR.? *Scientific

Information

Associates, 6546 Belleview Drive, Columbia, Maryland tKolbye Associates, Bethesda, Maryland 20817

Received

February

21046;

and

2. 1991

The maximum tolerated dose (MTD) regimen for testing substances for their ability to induce cancer and other chronic diseases in laboratory rodents has been required by governmental authorities for several decades. Cancer researchers originally suggested the MTD approach and it was then adopted by the FDA and EPA. The intention was to detect the ability of any substance under any circumstances, including the most extreme, to induce cancer in laboratory rodents. We question the validity of using the MTD in animal bioassaysto evaluate risk for human cancer. The paradox is that the safer the chemical, the higher the MTD, but the higher the MTD, the more likely that biochemical distortions will result and cause cellular injury, abnormal cell rep lication, toxic hyperplasia, and toxicity-induced cancer. All chemicals are toxic at some dose, whether relevant to anticipated human exposure or vastly exceeding it. New approaches to cancertesting lifetime bioassays are needed. A minimally toxic dose is defined and suggested to avoid specific tissue toxicity detected by clinical or pathology examination in animals subchronically exposed to the test compound for 90 days. The highest subtoxic dose that can be tolerated by test animals over a long period of time is suggested as being more appropriate for carcinogenicity bioassays. 0 1991 Academic Press. Inc.

INTRODUCTION

AND

BACKGROUND

History of the MTD. For therapeutic drugs pharmacologists traditionally have used the median effective dose (MED) as a measure of a desired intensity of drug action. When drug action is compared with its toxicity, the therapeutic index (TI) is used to express the desirable therapeutic effectiveness versus toxicity. Similar notations are often used to identify the dose required to produce some measurable fraction of the maximal intensity of effect of a test substance in toxicology. The adverse side effects and toxicity of pharmaceutical agents are recognized for their importance in designing therapeutic regimens to maximize benefit and minimize risk. In the practice of toxicology there are various interpretations of how and why the maximum tolerated dose concept developed and was adopted. It grew out of a desire ’ To whom correspondence should be addressed. 78 0273-2300/91 $3.00 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved

THE MTD CANCER ASSESSMENT

79

to establish that dose which would not be lethal to animals over a lifetime of administration but would still be effective in producing some toxic manifestation of the test chemical. Greater than 10% weight loss was also to be avoided. The MTD was deemed to be especially important for lifetime bioassays of substances being tested for carcinogenicity in laboratory rodents. Justifications made for the MTD include: a To find out whether a compound could cause cancer or other chronic diseases under extremely high exposure conditions such as prolonged intensive occupational or therapeutic exposures or exposure via a substantial component of foods in the diet. l To minimize the chance that a “carcinogen” would remain undetected. l To permit the maximum level of use of a compound proposed as a food additive or pesticide in order to maximize the intended benefits in agriculture or food manufacturing. l To compensate for the relatively small numbers of animals being tested and limitations in the statistical power of the bioassay to detect significant increases in cancer incidence in test animals as compared with controls. The false negative question was addressed by the rationale that one or two missed cancers in a test group of 50 or 100 rodents would translate to a significant impact on the larger populations of humans to be protected. Accordingly, high doses were justified as a means of maximizing the detection of chemical carcinogens and minimizing their nondetection. Underlying this approach was the no-threshold theory derived from radiation biology, that any exposure to a “carcinogen,” no matter how infinitesimal, carried with it a measurable consequence of real risk for inducing cancer. This theory has never been proven for chemical carcinogenesis, and there is much experimental and epidemiological evidence to the contrary which has been disregarded and ignored. But a serious question remains. Is it sensible to attempt corrections for one source of error, that of not detecting very weak, and probably indirectly acting, “carcinogens,” by introducing another source of error-the MTD-which may distort through toxicity and perhaps tumor promotion our perception of a chemical’s propensity to induce cancer? In recent decades, criticisms of the use of the MTD in rodent bioassays for carcinogenicity testing have continued with more and more scientists calling for reexamination of the validity of this testing regimen. A pharmacokinetic or toxicokineticbased study is now believed to provide data more relevant for interspecies comparisons with man. This alternative to the MTD would be referred to as the pharmacokinetic adjusted dose (PAD) that does not exceed the primary metabolic “breakpoint.” This would avoid secondary metabolic pathways at the higher doses that are less relevant to anticipated human exposure. A symposium has addressed many aspects of the application of pharmacokinetics in toxicity risk assessments (Kavlock, 1991). While not specifically considering carcinogenicity testing in most of these papers, these reports illustrate the fundamental concepts. They also emphasize the importance of assessing cellular proliferation and/or enzymatic changes on a chemical-by-chemical basis in evaluating nongenotoxic carcinogens (Moody et al., 199 1). EVOLUTION

OF SCIENTIFIC

OPINION

In 1984 following the report of the National Toxicology Program Review Panel for Chemical Carcinogenesis Testing and Evaluation and other reports, Burchfield et al.

80

CARR

AND

KOLBYE

(1974), Hottendorfand Pachter (1982), Kolbye (1982, 1983), the Nutrition Foundation (1983), and Moolenaar (1983), this paper was drafted but not published. Recent developments critical of the use of the MTD warrant publication because dose criteria based on pharmacokinetic data within the toxicokinetic range now appear to be more appropriate. In recent years numerous publications have emphasized the value of pharmacokinetic methodology. There are many examples, e.g., Somers ( 1987), Ames et al. (1987), Lave et al. ( 1988), Watanabe et al. (1988), Vocci and Farber ( 1988), Sheehan et al. (1989) EPA (1989), OMB (1990), Hart and Jensen (1990), Gregory (1990), Eason et al. (1990), Ames and Gold (1990), Cohen and Ellwein (1990), and Apostolou ( 1990). A MINI-TD

APPROACH

IS RECOMMENDED

We believe that the MTD should be converted to the minimally toxic dose (MINITD/minTD/PAD) or the highest subtoxic dose (HSTD) that can be tolerated by the test animals over a long period of time without demonstrable toxicity or histopathology. The high dosage selected should not produce a significant shortening of the lifespan or weight loss or demonstrable organ or tissue toxicity. Organ toxicity usually indicates disruption of cell and tissue function and structure. Potential effects of carcinogenicity and chronic toxicity over a two-year exposure span can be expected if preliminary subchronic exposure studies cause organ toxicity observable by clinical testing or histopathology. Some studies have ended disastrously because the MTD proved to be too high (Burchfield et al., 1974). Resources were not used to best advantage. When carcinogenicity studies are discovered to involve excessive doses, based on histopathology observations of either subchronic (90-day) studies or interim “take-off’ animals in lifetime studies, either the experiments should be modified to include a wider range of lower doses or the dosage regimens within the toxic range should be terminated entirely. Nondosage “recovery” studies may be appropriate to rescue potentially useful animal organ/tissue repair and response data after animals have been exposed to toxic doses. Tissue and organ toxicity data should be observed and recorded carefully in order to select dosage regimens in future studies to avoid toxicity and to test hypotheses concerning “safe doses” or “negligible risk human exposure ranges.” Otherwise the final test data will remain controversial, subject to criticism from reputable toxicologists, and will impair regulatory judgments and be the subject of legal challenges. Such inaccuracies and irrelevancies also encourage private litigation involving unfounded allegations of cancer causation and risk. “WORST

CASE” IS NOT

EFFECTIVE

PREVENTION

Any chemical administered in large enough doses to animals will cause toxic effects. It is a different matter, however, to firmly establish the safety of a substance in the more reasonable concentrations consumed by man. Highly toxic doses in animals, far above exposure limits or the amounts consumed by man, may not be relevant and may bear little relationship to the real life situation. This was pointed out 16 years ago by Burchfield et al. (1974).

THE

MTD

CANCER

ASSESSMENT

81

Because the primary purpose of toxicologists, public health authorities, and regulatory agencies should be to prevent human disease when such is preventable, appropriate dosage regimens are essential preconditions to obtaining toxicologic data usable for risk evaluation. It does no good to scare the public about chemical exposures that exceed by far the actual exposures they encounter in their daily lives at home and at work. Admittedly, this implies a threshold for regulation. The pharmacokinetic characteristics of the test compound should be determined in the most relevant species/strain of test animals by measuring absorption, metabolism, and excretion at reasonable doses that provide relevant data. A dose range should be selected with the highest dose that approximates the biological detoxification and excretory capabilities of the test animals. Any dosage exceeding these capabilities is dose-induced nonlinear kinetics. Such doses may induce biochemical and physiological changes and marked toxicity and carcinogenicity not encountered at lower doses. SUBCHRONIC

STUDIES

AS A BASIS FOR THE MTD

The subchronic study has been described in detail in the report of the Scientific Committee of the Food Safety Council (1980). The discussion includes test species, route of administration, dosage levels, diet, animal husbandry, clinical observations, and pathology considerations. This critique of the MTD is restricted to the determination of the carcinogenicity and the effects of oral doses in rodents in long-term studies such as those conducted by the National Toxicology Program for chemical carcinogenesis testing and evaluation. The prediction of the MTD has been described by Sontag et al. (1976) as that dose that will not produce “unwanted side effects” based on the prechronic preliminary studies. Thus, the MTD would be the highest dose given during the chronic study that can be predicted not to alter the animal’s normal longevity from effects other than carcinogenicity. It has been generally accepted that this highest dose will cause no more than a 10% body weight loss over time as compared with control rodents. The tacit assumption is made that this degree of weight loss may or may not be a sign of toxicity. A frequent problem of dietary administration of test substances related to body weight is the lack of accurate records of food intake. A similar problem is encountered when test substances are administered in drinking water. In diet preparation percentage weight is the unit of concentration; the toxicologist is interested in actual compound intake, generally expressed on a mg/kg body wt/day basis. Thus, dietary concentration, in the absence of food consumption data, may be misleading, if not actually meaningless, since it is affected by various factors, including caloric density and palatability, and especially by the diminishing rate of food intake per unit of body weight as the animal grows to maturity. One aspect of toxicity may be appetite depression leading to reduced dosage of test material as well as of nutritionally essential dietary components thus confounding nutritional insufficiency with toxic effects. Of particular significance in the subchronic studies using weanling rodents is the weanlings’ initial high metabolic activity and growth rate which results in the consumption of several times as much food as consumed by the fully grown animal, per unit of body weight. Unless the dietary concentration of the test material is adjusted weekly, or at least biweekly, to compensate for the changing rate of food intake, the

82

CARR

AND

KOLBYE

rapidly growing postweaning animal will be markedly overdosed until the anticipated adult intake is attained. Postweaning rats on normal diets consume, based on body weight, about 2.5 times the adult level, which they attain between 15-20 weeks of age; however, on test diets this multiple may vary depending on many factors. Substances added to the diet in excess of 5% of the total diet weight usually will exceed the level that animals will eat and may lead to nutritional deficits. A number of observations and guidelines deal with using subchronic tests to select doses for chronic toxicity testing. The key endpoints are organ specific and/or systemic pathology, body weight and organ weight alterations, mortality, clinical signs, and laboratory measurements of hematology, urinalysis, and clinical chemistry. The slope of the prechronic dose-response curve may influence the range and number of doses selected for the chronic toxicity test. The high dose for chronic toxicity testing generally lies between the dose producing a positive endpoint and the dose producing a minimal effect in the three-month subchronic test. This is the dose issue that is under criticism for reasons that will be presented. The first need is to select appropriate doses in order to obtain reliable data from chronic studies. To assessthe nonneoplastic toxic potential of a substance the highest test dose should be minimally toxic, but the carcinogenic potential of the substance may be grossly exaggerated by such doses. Lower doses should be used to span the toxic to the no-observed-adverse-effect range and the resulting data on cancer incidence examined in relation to dose. Overtly toxic test doses are of little or no value for quantitative human risk assessment in relation to low-level exposure if there are differences in metabolism, pharmacokinetics, or detoxification at lower exposure levels. Although the ceiling for minimal toxicity should be induced in subchronic studies, doses that cause overt toxicity or organ dysfunction should be avoided in lifetime studies. Improving the selection of appropriate doses will come from understanding the metabolism, pharmacokinetics, transformations, and excretion patterns of the test substance. We should relate the form and amount of the administered compound to the metabolites or transformation products which actually elicit adverse biological effects (Conti and Bickel, 1977). The need for metabolic studies became more widespread as methodology improved and regulatory requirements for classes of chemical products such as food additives or agricultural chemicals made metabolic information mandatory. The suggested approach avoids many scientific controversies now confronting us. Excessive doses will commingle and confuse the effects of primary carcinogens with other compounds, not really “carcinogens.” These induce cancer only by creating secondary toxicity which leads to a promoter-like effect. Excessively toxic doses may also effect toxicityinduced initiation by aberrant methylation of DNA (Barrows and Shank, 198 1). Apostolou (1990) reviewed a large number of carcinogenicity studies based on the use of the MTD and pointed out that a more systematic use of pharmacokinetics coupled with a better understanding of the cellular mechanisms of carcinogenicity would have provided a better data base. Because the latter mechanisms are largely unknown at present we should attempt to avoid rodent studies that provide results that are irrelevant in estimating human risk. VALUE

OF PHARMACOKINETIC

STUDIES

As knowledge of metabolic and pharmacokinetic changes in toxicology has increased, it is now recognized that such data characterizes the test substance and helps in planning

THE MTD CANCER ASSESSMENT

83

the strategy of approaches to safety evaluation. This knowledge is central for species selection, definition of dose regimens, and the conduct of in viva tests in subchronic studies to provide the following information: (a) Whether steady-state tissue concentrations are achieved; (b) Time required to reach a steady state with a particular dosage regimen of the test substance; (c) Total metabolic load; (d) How the metabolic and pharmacokinetic characteristics change with time, dose, and dose schedule. Determination of the circumstances when metabolic and/or excretory systems become overloaded as a result of high doses is most important. This implies that pharmacokinetic studies should bracket the doses selected and should include normal biochemical changes as well as those of the animal under the extreme stress levels of toxic doses. Gregory (1990) called this the “delivered dose” to be used in making quantitative risk assessments and it is of value in determining toxicity resulting from different routes of administration. It is often assumed that the principle justification for this information was to make interspecies comparisons so that the selection of test doses of one species could be translated to that of man most effectively. There are fallacies in this concept. It is assumed that human metabolism will he studied as part of the interspecies comparison, but it may be difficult. In some instances it may be sufficient to establish the metabolic pathways in an experimental animal such as the primate Mucacu. It is impossible to find a nonhuman species completely comparable to man in regard to metabolic pathways. CHEMICALLY

REACTIVE

METABOLITES

Knowledge of the formation in the test animal of chemically reactive metabolic products can provide insight about irreversible effects and neoplastic potential. Depending on their activity patterns, chemically reactive metabolites may bind covalently with a variety of tissue macromolecules such as proteins, RNA, DNA, and glycogen, as well as interacting with lipids and other small molecules such as glutathione. These are often transient metabolites, e.g., arene oxides, and are not detectable as such in body fluids or excreta but must be sought for at their nucleophilic binding sites. The subject of covalently bound derivatives of test substances is exceedingly complex but it can provide answers to a number of important questions about a compound’s toxicity. The numerous details have been outlined in the report of the Scientific Committee of the Food Safety Council (1980). In general, they may be summarized as follows: (a) (b) (c) (d) (e) (f)

Tissue formation site; Influences of enzyme mechanism(s); Site and moiety of adduct if detected; Can toxic action be accounted for by this mechanism?; Does covalent binding account for genetic or nongenetic effects?; Is there a dose-response relationship?

84

CARR

AND

KOLBYE

While this knowledge is available at present for only a few substances, the methodology exists and such data will assist in the process of safety evaluation. Measurement of the formation of these reaction metabolites in man and animals will be enhanced by the application of general pharmacokinetic principles (Gillette, 1976). Formation of these metabolites can be related to the dose and conditions of administration of the parent compound. If overdosage causes dose-dependent nonlinear kinetics, the reserve capacity of the normal metabolic and other adaptive mechanisms has probably been surpassed. There is extensive literature on this subject. Eason et al. (1990) have pointed out the importance and use of pharmacokinetic and receptor studies in the clinical evaluations of candidate drugs in patients and they explain some of the techniques employed in their review. There is less concern about the toxicity of metabolites if they normally form in the body of man. However, metabolic degradation and transformation should be carefully evaluated. DOSES THAT

EXCEED

METABOLIC

CAPACITY

Doses in chronic toxicity tests that exceed the animal’s physiologic capacity can exceed the metabolic break point and as a consequence abnormally high tissue concentrations of the test material are produced. These can cause nonspecific toxic challenges that represent a series of phenomena substantially related to the process involved with tumor promotion by classical promoters. This effect has been characterized as “toxic hyperplasia.” Toxic hyperplasia or bioaccumulation can increase tissue susceptibility to the initiating influence of carcinogenic compounds by increasing the susceptibility of cells to electrophilic attack (Kolbye, 1982). This aspect of disturbed cell proliferation has been called “mitogenesis increases mutagenesis” (Ames and Gold, 1990). Further mechanistic information is required to elucidate these concepts (Cohen and Ellwein, 1990). Cellular injury of a nonspecific nature is known to impair protective cellular enzymes which in turn can facilitate further increases in the local concentrations of active chemical moieties. Proteins are denatured, membranes are destabilized, and normal cellular processes, such as the active and passive transport of cellular components, are impaired or cease to function. The net result is enhanced exposure to attack on the DNA and RNA by genotoxic agents. It is unfortunate that the role of toxicity per se in relation to carcinogenicity has been poorly appreciated in the entire field of cancer studies (Kolbye, 1982). This is also true of mutagenicity. The pharmacokinetics and metabolism of chemical substances may differ substantially depending on low or high doses (Watanabe and Gehring, 1976). The issues of appropriate dosage schedules have been reviewed with recommendations by numerous review groups (Moolenaar, 1983). Smyth and Hottendorf (1980) suggested that absorption, blood concentration, and elimination data be obtained on a test substance before the toxicological tests are conducted. Based on studies with a number of substances, they believe kinetic data are essential in interpreting toxicity studies and choosing a dosing schedule. For example, Smyth et al. ( 1980) reported that 8-methoxsalen is more rapidly metabolized in the rat than in man and that dose-dependent pharmacokinetics indicate that drug disposition is changed in the rat at doses as low as a 1O-fold multiple of the human dose. Therefore, the use of higher saturating doses

THE MTD CANCER ASSESSMENT

85

in a toxicological study would be useless and any extrapolation of toxic manifestations to man based on these high doses would be questionable. In a critique of the results of the approximately 200 NC1 carcinogenesis bioassays, Hottendorf and Pachter (1982) reported that the bioavailability, metabolism, and pharmacokinetics of a test substance should be determined in the animal at the proposed dose levels to be used in the selection of the MTD. Doses that produce “irrelevant aberrations in biotransformation and elimination should be avoided.” They recommend that some reasonable expression of a dose-response relationship be required in any positive carcinogenesis bioassay. After reviewing the issues related to the use of the MTD the American Industrial Health Council (AIHC) (Moolenaar, 1983) stated its position: Bioassaysusing the maximum tolerated dose administered via unexpected routes of exposure have not been selective in distinguishing chemicals for regulation as carcinogens, and furthermore such studies provide very little guidance for risk assessment.

Admittedly the acquisition of pharmacokinetic or “toxicokinetic” data is difficult and time consuming. However, when so much is at stake it is certainly prudent and necessary to critically examine how these data are obtained and their validity. SUMMARY Based on this review and analysis of the issues related to the problems created by the traditional use of the maximum tolerated dose, it is suggested that the concept be discontinued and that a new approach be selected based on the measurement of the pharmacokinetic parameters of the individual test substance. Metabolic criteria, target cell identification, excretion factors, and related in viva data are more accurate indicators of subliminal toxicity that require evaluation as one develops the dosage protocol for chronic toxicity studies. The animal dosage selected based on the pharmacokinetics of the test substance should give a linear dose-response relationship. This may be called the pharmacokinetic adjusted dose. Doses exceeding the maximum PAD would be recognized as exceeding the metabolic break point and would not be included in the regulatory decision process. This concept was suggested also by the AIHC (Moolenaar, 1983). Ideally, the lowest dose for animals should not be less than the pharmacokinetical equivalent of the human exposure level. The high dose in chronic studies should not exceed the toxic effect dose determined in the subchronic studies described, but it is expected to elicit some signs of toxicity. However, homeostasis and physiological defense mechanisms should not be disturbed; nor should there be any likelihood of inducing secondary carcinogenesis. It is important to recognize that doses above the threshold (the metabolic break point) may provide insight into toxicity mechanisms, but these doses exceed the physiological capacity of the animal to metabolize the chemical. Therefore, this fact must be taken into account when assessing the toxicity of the substance and when extrapolating the findings to man. In many tests, especially in carcinogenicity studies, these issues have not been properly reconciled. Conversely, for those substances shown to lack substantial mutagenic activity and for which clearly subtoxic doses are not carcinogenic, a nonlinear or safety factor approach should be taken, predicated on judgments by qualified scientists, taking into

86

CARR AND KOLBYE

account all available data from short-term toxicokinetics bioassays (Kolbye, 1983).

and prechronic and chronic

CONCLUSIONS With more reasonable and verifiable scientific data, public health decisions concerning estimations of risk to humans from ingestive exposures could be made on a much sounder scientific basis. Chronic tests on animals using the MTD should be replaced by the more realistic pharmacokinetic adjusted dose that equates the individual metabolic data of the test substance with the toxicity as detected in the subchronic studies. This approach would materially assist in estimating the health risk of chemicals on a rational basis. This critique has been restricted to the present established use of the MTD in carcinogenicity studies in rodents. These health risk assessments using animal tests have advantages and disadvantages recognized by toxicologists but not by the public at large. The decisions whether to ban chemicals on these evaluations have led to controversy, polemics, disputes, and litigation that has reached the point that some suggest an alternative be sought. We believe the utilization of the proper toxicologic methodology would avoid the disagreements and provide a rational basis for public health decisions. REFERENCES AMES, B. N., et al. (1987). Ranking possible AMES, B. N., AND GOLD, L. S. (1990). Too Science 249,970-97 1.

carcinogenic hazards. Science 236 (April 17). many rodent carcinogens: Mitogenesis increases mutagenesis.

APOSTOLOU, A. ( 1990). Relevance of maximum tolerated dose to human carcinogenic risk. Regul. Toxicof. Pharmacol.

11,68-80.

BARROWS,L. R., AND SHANK, R. C. (198 1). Aberrant methylation of liver DNA in rats during hepatoxicity. Toxicol.

Appl. Pharmacol.

80,334-335.

BURCHFIELD,H. P., STORRS,E. E., AND KRAYBILL, H. F. (1974). The maximum tolerated dose in pesticide carcinogenicity studies. Environ. Qual. Saf: Suppl. 3, 599-603. COHEN, S. M., AND ELLWEIN, L. B. (1990). Cell proliferation in carcinogenesis. Science 249, 1007- 1011. CONTI, A., AND BICKEL, M. H. (1977). History of drug metabolism: Discoveries of the major pathways in the 19th Century. Drug Metab. Rev. 6. 1. EASON,C. T., et al. (1990). The importance of pharmacokinetic and receptor studies in drug safetyevaluation. Regul.

Toxicol.

Pharmacol.

11, 288-307.

Environmental Protection Agency (EPA) (1989). Risk Assessment Forum. EPA/625/3-89/015. Environmental Protection Agency Washington, DC. (March 2 1-26). Food SafetyCouncil (1980). Proposed System for Food Safety Assessment: Report of the Scientific Committee. Food Safety Council, Washington, DC (June). GILLETTE, J. R. (1976). Application of pharmacokinetic principles in the extrapolation of animal data to humans. Clin. Toxicol. 9, 709. GREGORY, A. R. (1990). Uncertainty in health risk assessments.Regul. Toxicol. Pharmacol. 11, 191-200. HART, J. W., AND JENSEN,N. J. (1990). The myth ofthe final hazard assessment.Regal. Toxicol. Pharmacol. 11, 123-131. HOTTENWRF, G. H., AND PACHTER, I. J. (1982). An analysis of the carcinogenesis testing experience of the National Cancer Institute. Toxicol. Pathol. 10,22-26. KAVLOCK, R. J. (199 1). Symposium on application of pharmacokinetics in developmental toxicity risk assessments.Fundam. Appl. Toxicol. 16, 2 13-232. KOLBYE, A. C., JR. (1982). Proposed research and regulatory plan to improve risk evaiuation concerning factors contributing to the incidence of cancer and genetic damage. Regul. Toxicol. Pharmacol. 2, 232237.

THE MTD CANCER ASSESSMENT

87

KOLBYE, A. C., JR. (1983). Regulatory considerations regarding safety testing. Prod. Suf Eva/. 1,45-49. LAVE, L. B., et al. (1988). Information value of the rodent bioassay. Nature 336,631-633. MOODY, D. E., et al. (1991). Peroxisome proliferation and nongenotoxic carcinogenesis: Commentary on a symposium. Fundam. Appl. Toxicol. 16,233-248. MOOLENAAR, R. J. (1983). American Industrial Health Council view of current policy direction in the federal establishment. Regtd. Toxicol. Pharmacol. 3, 381-388. The Nutrition Foundation (1983). Ad Hoc Working Group Report on Oil/Gavage in Toxicology. The Nutrition Foundation, Washington, DC (July 14 and 15). Office of Management and Budget (OMB) (1990). Regulatory Program of the United States Government: April 1, I990-March 31, 1991. Office of Management and Budget, Washington, DC. SHEEHAN, D. M., et al. (1989). Workshop on risk assessmentin reproductive and developmental toxicology: Addressing the assumptions and identifying the research needs. Regul. Toxicol. Pharmacol. 10, 110-122. SMYTH, R. D., AND HOTTENDORF, G. H. (1980). Application of pharmacokinetics and biopharmaceutics in the design of toxicologica studies. Toxicol. Appl. Pharmacol. 53, 179-195. SMYTH. R. D., et al. (1980). Biological disposition of 8-methoxsalen in rat and man. Arzneim.-Forsch. 30(11)9, 1725-1730. SOMERS, E. (1987). Making decisions from numbers. Regul. Toxicol. Pharmacol. 7, 35-42. SONTAG, J. M., PAGE, N. P., AND SA~OTTI, U. (1976). Guidelines for Carcinogens Bioassay in Small Rodents. DHEW Publication No. (NIH) 76-801. NIH, Washington, DC. VOCCI, F., AND FARBER,T. (1988). Extrapolation of animal toxicity data to man. Regul. Toxicol. Pharmucol. 8,389-398.

WATANABE, P. G., AND GEHRING, P. J. (1976). Dose dependent fate of vinyl chloride and its possible relationship to oncogenicity in rats. Environ. Health Perspect. 17, 145-l 52. WATANABE, P. G.. et al. ( 1988). Toxicokinetics in the evaluation oftoxicity data. Regul. Toxicol. Pharmacol. 8,389-398.