Search

The long running controversy about the relative merits of hazard-based versus risk-based approaches has been investigated. There are three levels of hazard codification: level 1 divides chemicals into dichotomous bands of hazardous and non-hazardous; level 2 divides chemicals into bands of hazard based on severity and/or potency; and level 3 places each chemical on a continuum of hazard based on severity and/or potency. Any system which imposes compartments onto a continuum will give rise to issues at the boundaries, especially with only two compartments. Level 1 schemes are only justifiable if there is no variation in severity, or potency or if there is no threshold. This is the assumption implicit in GHS/EU classification for carcinogenicity, reproductive toxicity and mutagenicity. However, this assumption has been challenged. Codification level 2 hazard assessments offer a range of choices and reduce the built-in conflict inherent in the level 1 process. Level 3 assessments allow a full range of choices between the extremes and reduce the built-in conflict even more. The underlying reason for the controversy between hazard and risk is the use of level 1 hazard codification schemes in situations where there are ranges of severity and potency which require the use of level 2 or level 3 hazard codification. There is not a major difference between level 2 and level 3 codification, and they can both be used to select appropriate risk management options. Existing level 1 codification schemes should be reviewed and developed into level 2 schemes where appropriate.

Concern over substances that may cause cancer has led to various classification schemes to recognize carcinogenic threats and provide a basis to manage those threats. The least useful schemes have a binary choice that declares a substance carcinogenic or not. This overly simplistic approach ignores the complexity of cancer causation by considering neither how the substance causes cancer, nor the potency of that mode of action. Consequently, substances are classified simply as “carcinogenic”, compromising the opportunity to properly manage these kinds of substances. It will likely be very difficult, if not impossible, to incorporate New Approach Methodologies (NAMs) into binary schemes. In this paper we propose a new approach cancer classification scheme that segregates substances by both mode of action and potency into three categories and, as a consequence, provides useful guidance in the regulation and management of substances with carcinogenic potential. Examples are given, including aflatoxin (category A), trichlorethylene (category B), and titanium dioxide (category C), which demonstrate the clear differentiation among these substances that generate appropriate levels of concern and management options.

Developments in the understanding of the etiology of cancer have profound implications for the way the carcinogenicity of chemicals is addressed. This paper proposes a unified theory of carcinogenesis that will illuminate better ways to evaluate and regulate chemicals. In the last four decades, we have come to understand that for a cell and a group of cells to begin the process of unrestrained growth that is defined as cancer, there must be changes in DNA that reprogram the cell from normal to abnormal. Cancer is the consequence of DNA coding errors that arise either directly from mutagenic events or indirectly from cell proliferation especially if sustained. Chemicals that act via direct interaction with DNA can induce cancer because they cause mutations which can be carried forward in dividing cells. Chemicals that act via non-genotoxic mechanisms must be dosed to maintain a proliferative environment so that the steps toward neoplasia have time to occur. Chemicals that induce increased cellular proliferation can be divided into two categories: those which act by a cellular receptor to induce cellular proliferation, and those which act via non-specific mechanisms such as cytotoxicity. This knowledge has implications for testing chemicals for carcinogenic potential and risk management.

Over 50 years, we have learned a great deal about the biology that underpins cancer but our approach to testing chemicals for carcinogenic potential has not kept up. Only a small number of chemicals has been tested in animal-intensive, time consuming, and expensive long-term bioassays in rodents. We now recommend a transition from the bioassay to a decision-tree matrix that can be applied to a broader range of chemicals, with better predictivity, based on the premise that cancer is the consequence of DNA coding errors that arise either directly from mutagenic events or indirectly from sustained cell proliferation. The first step is in silico and in vitro assessment for mutagenic (DNA reactive) activity. If mutagenic, it is assumed to be carcinogenic unless evidence indicates otherwise. If the chemical does not show mutagenic potential, the next step is assessment of potential human exposure compared to the threshold for toxicological concern (TTC). If potential human exposure exceeds the TTC, then testing is done to look for effects associated with the key characteristics that are precursors to the carcinogenic process, such as increased cell proliferation, immunosuppression, or significant estrogenic activity. Protection of human health is achieved by limiting exposures to below NOEALs for these precursor effects. The decision tree matrix is animal-sparing, cost effective, and in step with our growing knowledge of the process of cancer formation.

The evolved World Health Organization/International Programme on Chemical Safety mode of action (MOA) framework provides a structure for evaluating evidence in pathways of causally linked key events (KE) leading to adverse health effects. Although employed globally, variability in use of the MOA framework has led to different interpretations of the sufficiency of evidence in support of hypothesized MOAs. A proof of concept extension of the MOA framework is proposed for scoring confidence in the supporting data to improve scientific justification for MOA use in characterizing hazards and selecting dose-response extrapolation methods for specific chemicals. This involves selecting hypothesized MOAs, and then, for each MOA, scoring the weight of evidence (WOE) in support of causality for each KE using evolved Bradford Hill causal considerations (biological plausibility, essentiality, dose-response concordance, consistency, and analogy). This early proof of concept method is demonstrated by comparing two potential MOAs (mutagenicity and peroxisome proliferator activated receptor-alpha) for clofibrate, a rodent liver carcinogen. Quantitative confidence scoring of hypothesized MOAs is shown to be useful in characterizing the likely operative MOA. To guide method refinement and future confidence scoring for a spectrum of MOAs, areas warranting further focus and lessons learned, including the need to incorporate a narrative discussion of the weights used in the evaluation and an overall evaluation of the plausibility of the outcome, are presented.

Background: Many reports have been published that contain recommendations for improving the quality, transparency, and usefulness of decision making for risk assessments prepared by agencies of the U.S. federal government. A substantial measure of consensus has emerged regarding the characteristics that high-quality assessments should possess. Objective: The goal was to summarize the key characteristics of a high-quality assessment as identified in the consensus-building process and to integrate them into a guide for use by decision makers, risk assessors, peer reviewers and other interested stakeholders to determine if an assessment meets the criteria for high quality. Discussion: Most of the features cited in the guide are applicable to any type of assessment, whether it encompasses one, two, or all four phases of the risk-assessment paradigm; whether it is qualitative or quantitative; and whether it is screening level or highly sophisticated and complex. Other features are tailored to specific elements of an assessment. Just as agencies at all levels of government are responsible for determining the effectiveness of their programs, so too should they determine the effectiveness of their assessments used in support of their regulatory decisions. Furthermore, if a nongovernmental entity wishes to have its assessments considered in the governmental regulatory decision-making process, then these assessments should be judged in the same rigorous manner and be held to similar standards. Conclusions: The key characteristics of a high-quality assessment can be summarized and integrated into a guide for judging whether an assessment possesses the desired features of high quality, transparency, and usefulness. Citation: Fenner-Crisp PA, Dellarco VL. 2016. Key elements for judging the quality of a risk assessment. Environ Health Perspect 124:1127–1135; http://dx.doi.org/10.1289/ehp.1510483

Developments in the understanding of the etiology of cancer have undermined the 1970s concept that chemicals are either "carcinogens" or "non-carcinogens". The capacity to induce cancer should not be classified in an inflexible binary manner as present (carcinogen) or absent (non-carcinogen). Chemicals may induce cancer by three categories of mode of action: direct interaction with DNA or DNA replication including DNA repair and epigenetics; receptor-mediated induction of cell division; and non-specific induction of cell division. The long-term rodent bioassay is neither appropriate nor efficient to evaluate carcinogenic potential for humans and to inform risk management decisions. It is of questionable predicitiveness, expensive, time consuming, and uses hundreds of animals. Although it has been embedded in practice for over 50 years, it has only been used to evaluate less than 5% of chemicals that are in use. Furthermore, it is not reproducible because of the probabilisitic nature of the process it is evaluating combined with dose limiting toxicity, dose selection, and study design. The modes of action that lead to the induction of tumors are already considered under other hazardous property categories in classification (Mutagenicity/Genotoxicity and Target Organ Toxicity); a separate category for Carcinogenicity is not required and provides no additional public health protection.

The identification of adverse health effects has a central role in the development and risk/safety assessment of chemical entities and pharmaceuticals. There is currently a need for better alignment regarding how nonclinical adversity is determined and characterized. The European Society of Toxicologic Pathology (ESTP) therefore coordinated a workshop to review available definitions of adversity, weigh determining and qualifying factors of adversity based on case examples, and recommend a practical approach to define and characterize adversity in toxicology reports, to serve as a valuable prerequisite for future organ- or lesion-specific workshops planned by the ESTP.

To advance the utility and predictability of safety evaluation, an integrated approach that relies on all existing knowledge to understand how agents perturb normal biological function or structure is needed to progress more focused evaluation strategies. The mode of action (MOA)-human relevance framework developed by the International Program for Chemical Safety and The International Life Sciences Institute provides a useful analytical approach where different lines of evidence (e.g., in vitro, in vivo) can be organized, linked, and integrated at different levels of biological organization into a more efficient, hypothesis-driven approach to safety evaluation. This framework provides a weight-of-evidence approach based on considerations for causality (as originally articulated by Bradford Hill), including dose response and temporal concordance, consistency, specificity, and biological plausibility and coherence. Once an animal MOA and its key events are established, qualitative and quantitative comparisons between experimental animals and humans are made based on the key events. This comparison enables a conclusion as to whether the MOA is likely operative in humans and, if so, whether it can result in a more refined hazard and dose-response assessment. This framework provides an important tool to promote and formalize the use of MOA data in safety evaluation regardless of whether the information comes from traditional or novel approaches, such as those recommended by the NRC in its 2007 report "Toxicity Testing in the 21st Century," which recommends moving away from traditional approaches of measuring adverse endpoints by using newer technologies to identify ways agents may considerably perturb cellular pathways to produce their toxicity.

Significant advances have been made in human health and ecological risk assessment over the last decade. Substantial challenges, however, remain in providing credible scientific information in a timely and efficient manner to support chemical risk assessment and management decisions. A major challenge confronting risk managers is the need for critical information to address risk uncertainties in large chemical inventories such as high- and medium-production-volume industrial chemicals or pesticide inert ingredients. From a strategic and tactical viewpoint, an integrated approach that relies on all existing knowledge and uses a range of methods, including those from emerging and novel technologies, is needed to advance progressive and focused testing strategies, as well as to advance the utility and predictability of the risk assessment by providing more relevant information. A hypothesis-based approach that draws on all relevant information is consistent with the vision articulated in the 2007 report by the National Research Council, Toxicity Testing in the 21st Century: A Vision and a Strategy. This article describes the current practices in evaluating chemical risks and ongoing efforts to enhance the quality and efficiency of risk assessment and risk management decisions within the Office of Prevention, Pesticides, and Toxic Substances at the U.S. Environmental Protection Agency.

Prior to October 2007, the US Environmental Protection Agency (EPA) required both 13-week and 1-year studies in Beagle dogs be submitted in support of registration for pesticides. Following an extensive retrospective analysis, we (the authors) determined that the 1-year toxicity dog study should be eliminated as a requirement for pesticide registration. The present work presents this retrospective analysis of results from 13-week and 1-year dog studies for 110 conventional pesticide chemicals, representing more than 50 classes of pesticides. The data were evaluated to determine if the 13-week dog study, in addition to the long-term studies in two rodent species (mice and rats), were sufficient for the identification of no observed adverse effect levels (NOAELs) and lowest observed adverse effect levels (LOAELs) for the derivation of chronic reference doses (RfD). Only pesticides with adequate 13-week and 1-year duration studies were included in the present evaluation. Toxicity endpoints and dose-response data from 13-week and 1-year studies were compared. The analysis showed that 70 of the 110 pesticides had similar critical effects regardless of duration and had NOAELs and LOAELs within a difference of 1.5-fold of each other. For the remaining 40 pesticides, 31 had lower NOAELs and LOAELs in the 1-year study, primarily due to dose selection and spacing. In only 2% of the cases were additional toxic effects identified in the 1-year study that were not observed in the 13-week study and/or in the rodent studies. In 8% of the cases, the 1-year dog had a lower NOAEL and/or LOAEL than the 13-week study, but there would have been no regulatory impact if the 1-year dog study had not been performed because adequate data were available from the other required studies. A dog toxicity study beyond 13-weeks does not have significant impact on the derivation of a chronic RfD for pesticide risk assessment.

Although rarely occurring in humans, hemangiosarcomas (HS) have become important in evaluating the potential human risk of several chemicals, including industrial, agricultural, and pharmaceutical agents. Spontaneous HS arise frequently in mice, less commonly in rats, and frequently in numerous breeds of dogs. This review explores knowledge gaps and uncertainties related to the mode of action (MOA) for the induction of HS in rodents, and evaluates the potential relevance for human risk. For genotoxic chemicals (vinyl chloride and thorotrast), significant information is available concerning the MOA. In contrast, numerous chemicals produce HS in rodents by nongenotoxic, proliferative mechanisms. An overall framework is presented, including direct and indirect actions on endothelial cells, paracrine effects in local tissues, activation of bone marrow endothelial precursor cells, and tissue hypoxia. Numerous obstacles are identified in investigations into the MOA for mouse HS and the relevance of the mouse tumors to humans, including lack of identifiable precursor lesions, usually late occurrence of the tumors, and complexities of endothelial biology. This review proposes a working MOA for HS induced by nongenotoxic compounds that can guide future research in this area. Importantly, a common MOA appears to exist for the nongenotoxic induction of HS, where there appears to be a convergence of multiple initiating events (e.g., hemolysis, decreased respiration, adipocyte growth) leading to either dysregulated angiogenesis and/or erythropoiesis that results from hypoxia and macrophage activation. These later events lead to the release of angiogenic growth factors and cytokines that stimulate endothelial cell proliferation, which, if sustained, provide the milieu that can lead to HS formation.

Classification schemes for carcinogenicity based solely on hazard-identification such as the IARC monograph process and the UN system adopted in the EU have become outmoded. They are based on a concept developed in the 1970s that chemicals could be divided into two classes: carcinogens and non-carcinogens. Categorization in this way places into the same category chemicals and agents with widely differing potencies and modes of action. This is how eating processed meat can fall into the same category as sulfur mustard gas. Approaches based on hazard and risk characterization present an integrated and balanced picture of hazard, dose response and exposure and allow informed risk management decisions to be taken. Because a risk-based decision framework fully considers hazard in the context of dose, potency, and exposure the unintended downsides of a hazard only approach are avoided, e.g., health scares, unnecessary economic costs, loss of beneficial products, adoption of strategies with greater health costs, and the diversion of public funds into unnecessary research. An initiative to agree upon a standardized, internationally acceptable methodology for carcinogen assessment is needed now. The approach should incorporate principles and concepts of existing international consensus-based frameworks including the WHO IPCS mode of action framework.

In vitro genotoxicity assays are often used to screen and predict whether chemicals might represent mutagenic and carcinogenic risks for humans. Recent discussions have focused on the high rate of positive results in in vitro tests, especially in those assays performed in mammalian cells that are not confirmed in vivo. Currently, there is no general consensus in the scientific community on the interpretation of the significance of positive results from the in vitro genotoxicity assays. To address this issue, the Health and Environmental Sciences Institute (HESI), held an international workshop in June 2006 to discuss the relevance and follow-up of positive results in in vitro genetic toxicity assays. The goals of the meeting were to examine ways to advance the scientific basis for the interpretation of positive findings in in vitro assays, to facilitate the development of follow-up testing strategies and to define criteria for determining the relevance to human health. The workshop identified specific needs in two general categories, i.e., improved testing and improved data interpretation and risk assessment. Recommendations to improve testing included: (1) re-examine the maximum level of cytotoxicity currently required for in vitro tests; (2) re-examine the upper limit concentration for in vitro mammalian studies; (3) develop improved testing strategies using current in vitro assays; (4) define criteria to guide selection of the appropriate follow-up in vivo studies; (5) develop new and more predictive in vitro and in vivo tests. Recommendations for improving interpretation and assessment included: (1) examine the suitability of applying the threshold of toxicological concern concepts to genotoxicity data; (2) develop a structured weight of evidence approach for assessing genotoxic/carcinogenic hazard; and (3) re-examine in vitro and in vivo correlations qualitatively and quantitatively. Conclusions from the workshop highlighted a willingness of scientists from various sectors to change and improve the current paradigm and move from a hazard identification approach to a "realistic" risk-based approach that incorporates information on mechanism of action, kinetics, and human exposure..

Thiazopyr increases the incidence of male rat thyroid follicular-cell tumors; however, it is not carcinogenic in mice. Thiazopyr is not genotoxic. Thiazopyr exerts its carcinogenic effect on the rat thyroid gland secondary to enhanced metabolism of thyroxin leading to hormone imbalance. The relevance of these rat tumors to human health was assessed by using the 2006 IPCS Human Relevance Framework. The postulated rodent tumor mode of action was tested against the Bradford Hill criteria and was found to satisfy the conditions of dose and temporal concordance, biological plausibility, coherence, strength, consistency, and specificity that fits with a well-established mode of action for thyroid follicular-cell tumors. Although the postulated mode of action could theoretically operate in humans, marked quantitative differences in the inherent susceptibility for neoplasia to thyroid hormone imbalance in rats allows for the conclusion that thiazopyr does not pose a carcinogenic hazard to humans.

As our scientific technology grows, risk assessment methods become more complex and, therefore, open to greater scientific debate. Risk assessment has always been a part of the regulatory notification and approval process for foods. However, the methodologies applied to risk assessment and decision-making have become diverse, dependent on a number of features, including the areas of the world in which one operates, the need to use cumulative risk assessment for pesticides and otheringredients or alternative risk assessment considerations for evaluating nontraditional or bioengineered foods. Diverse institutional structures within a single federal regulatory authority may tend to lead to diversity in risk outcomes that creates policy decisions that complicate and confuse the risk management process. On top of this challenge, decisions become more complicated by the need to examine beneficial factors of foods rather than the adverse effects of foods and food additives. Foods are a complex mixture ofingredients. Regulatory groups recognize the need to use new approaches for evaluating the safety and risks associated with foods and food additives, and to do so in a timely manner. The United States Food and Drug Administration (US FDA) in its need to ensure standards of “reasonable certainty of no harm” continues to explore alternative means to be responsive to petitioners as well as continue to examine scientifically validated means, e.g., quantitative structure-activity relationship (QSAR), and computer-assisted programs, within the approval process to assist in the evaluation of risks. Another means to improve the risk management process would include the cumulative risk assessment of pesticides that will, no doubt, be the beginning of more intensive efforts to understand cumulative exposures and the inherent risks from multiple pathways of exposure. The passage of the Food Quality Protection Act (FQPA) resulted in developing additional risk assessment methodologies and approaches to assess the potential for multiple exposures and risks. Addressing the international criteria used in decision-making related to foods safety assessment has resulted in acceptable intake values for foodingredients for carcinogens and noncarcinogens that, in general, tend to be more stringent in the United States compared to Europe. Clearly, the need for harmonization of risk assessment criteria and the impact of the decision process on regulatory approvals and safety assessment is a future need for the continued assurances of food safety. The topics presented in this paper are based on a symposium held in November 2002 at the annual meeting of the American College of Toxicology.

The human relevance framework (HRF) outlines a four-part process, beginning with data on the mode of action (MOA) in laboratory animals, for evaluating the human relevance of animal tumors. Drawing on U.S. EPA and IPCS proposals for animal MOA analysis, the HRF expands those analyses to include a systematic evaluation of comparability, or lack of comparability, between the postulated animal MOA and related information from human data sources. The HRF evolved through a series of case studies representing several different MOAs. HRF analyses produced divergent outcomes, some leading to complete risk assessment and others discontinuing the process, according to the data available from animal and human sources. Two case examples call for complete risk assessments. One is the default: When data are insufficient to confidently postulate a MOA for test animals, the animal tumor data are presumed to be relevant for risk assessment and a complete risk assessment is necessary. The other is the product of a data-based finding that the animal MOA is relevant to humans. For the specific MOA and endpoint combinations studied for this article, full risk assessments are necessary for potentially relevant MOAs involving cytotoxicity and cell proliferation in animals and humans (Case Study 6, chloroform) and formation of urinary-tract calculi (Case Study 7, melamine). In other circumstances, when data-based findings for the chemical and endpoint combination studied indicate that the tumor-related animal MOA is unlikely to have a human counterpart, there is little reason to continue the risk assessment for that combination. Similarly, when qualitative considerations identify MOAs specific to the test species or quantitative considerations indicate that the animal MOA is unlikely to occur in humans, such hazard findings are generally conclusive and further risk assessment is not necessary for the endpoint-MOA combination under study. Case examples include a tumor-related protein specific to test animals (Case Study 3, d-limonene), the tumor consequences of hormone suppression typical of laboratory animals but not humans (Case Study 4, atrazine), and chemical-related enhanced hormone clearance rates in animals relative to humans (Case Study 5, phenobarbital). The human relevance analysis is highly specific for the chemical-MOA-tissue-endpoint combination under analysis in any particular case: different tissues, different endpoints, or alternative MOAs for a given chemical may result in different human relevance findings. By providing a systematic approach to using MOA data, the HRF offers a new tool for the scientific community's overall effort to enhance the predictive power, reliability and transparency of cancer risk assessment.

There are new opportunities to using data from molecular and cellular studies in order to bring together a fuller biological understanding of how chemicals induce neoplasia. In 1996, the Environmental Protection Agency (EPA) published a proposal to replace its 1986 Guidelines for Carcinogen Risk Assessment to take advantage of these new scientific advances in cancer biology. The analytical framework within the new guidelines focuses on an understanding of the mode of carcinogenic action. Mode of action data come into play in a couple of ways in these new guidelines. For example, such information can inform the dose–response relationship below the experimental observable range of tumours. Thus, mode of action data can be useful in establishing more appropriate guidance levels for environmental contaminants. It is the understanding of the biological processes that lead to tumour development along with the response data derived from experimental studies that can help discern the shape of the dose–response at low doses (linear vs. nonlinear). Because it is experimentally difficult to establish “true thresholds” from others with a nonlinear dose–response relationship, the proposed guidelines take a practical approach to depart from low-dose linear extrapolation procedures when there is sufficient experimental support for a mode of action consistent with nonlinear biological processes (e.g., tumours resulting from the disruption of normal physiological processes).

When the United States Environmental Protection Agency (USEPA) published guidelines for carcinogen risk assessment in 1986, it provided essentially no guidance on how to consider mechanistic information. The guidelines simply pointed out that ‘‘The Agency will review each assessment as to the evidence on carcinogenesis mechanisms and other biological or statistical evidence that indicates the suitability of a particular extrapolation model.’’ (USEPA, 1986). At that time, mechanisms of carcinogenesis were largely unknown and empirically difficult to approach. Nonetheless, as mechanistic data became available, the Agency considered it in formulating science policies, e.g., alpha2u-globulin–induced rat renal tumors (USEPA, 1991) and thyroid follicular cell rodent tumors and disruption of thyroid homeostasis (USEPA, 1998) or in chemical-specific assessments, e.g., bladder tumors induced by melamine through formation of urinary calculi (USEPA, 1988). In 1996, the Agency formally proposed to make information on the mode of carcinogenic action a pivotal point of the cancer risk assessment process (USEPA, 1996). Because complete knowledge of how an agent causes cancer is unlikely to exist (certainly for the near term), the 1996 guidelines put forth the notion of understanding mode of action versus mechanism of action. The former being a less detailed biochemical description of events than is meant by mechanism of action. The mode of action is sufficient evidence to draw a reasonable working conclusion concerning the agent’s influence on key processes. The mode of action concept permits information on precursor events to be evaluated and incorporated into the risk assessment process in a realistic way. The most frequent comment about the Agency’s 1996 proposed revision of its cancer risk assessment guidelines was that more guidance was needed on how to evaluate an agent’s mode of action. A few comments raised concern that the new USEPA guidelines would not be public health protective and opposed using mode of action data. In response, in 1999, EPA incorporated a mode of action framework analysis (USEPA, 1999) in conjunction with work by the International Programme for Chemical Safety (IPCS) (SonichMullin et al., 2001). This framework provided assurance that a rigorous and transparent approach would be taken to using mode of action data in the Agency’s cancer risk assessments, and it was well received by the scientific community. The mode of action framework is derived from considerations for causality used in epidemiologic investigations, which were originally articulated by Hill (1965). The Hill criteria were later modified by the U.S. Department of Health and Human Services (USHHS, 1982) and by Faustman et al. (1997) to extend to experimental animal studies. The publication of the USEPA 1999 framework provided an important tool with which the Agency can promote and formalize the use of mode of action data in cancer assessment. Mode of action data have come into play in several ways in EPA risk assessments. It has been critical in informing the dose–response relationship below the experimental observable range of tumors, and thus it is useful in establishing more appropriate guidance levels for environmental contaminants. The Agency’s traditional linear default extrapolation procedure for modeling tumor incidence has been a longstanding and often controversial issue. The 1986 USEPA guidelines provided only one dose–response default, namely the linear extrapolation procedure. The 1999 (and 1996) guidelines advanced this issue by providing two default extrapolation approaches—one linear and and one nonlinear. Application of the nonlinear default extrapolation procedure (typically a margin of exposure analysis or determination of a reference dose/ concentration) is based on the mode of action understanding about the agent. It was acknowledged in the 1999 (1996) guidelines that ‘‘true thresholds’’ are experimentally difficult to This paper does not necessarily reflect the views and policies of the U.S. Environmental Protection Agency. The opinions expressed within this paper reflect the views of the authors. 1 To whom correspondence should be addressed. Fax: þ1-703-5147; E-mail: dellarco.vicki@epamail.epa.gov.

The U.S. Environmental Protection Agency (USEPA) recently proposed new guidelines to update and replace the 1986 USEPA Guidelines for Carcinogen Risk Assessment. Today, there is a better understanding of the variety of modes by which carcinogens can operate that did not exist when the 1986 USEPA guidelines were published. Many laboratories are adding new test protocols in their programs directed at questions concerning the mechanisms of action of carcinogens. In response to the evolving science of carcinogenesis, the new guidelines provide an analytical framework for incorporating all relevant biological information and recognizing a variety of situations regarding cancer risk. In addition, the guidelines are flexible enough to allow consideration of future scientific advances.

This paper explores how quantitative risk assessment methods might be extended to analysis of risks to the human germ line. High inhalation exposures to ethylene oxide are reported to cause heritable translocations in male mice with a steep and nonlinear dose-response-curve. We explore quantitative estimation of risk to humans from low exposures based on these animal data, addressing questions of tissue dosimetry for this alkylating agent, expected equivalency of doses across species, germ-cell sensitivity, and extrapolation of dose-response relationship to low exposure levels. Various dose-response models are discussed in terms of their applicability to genetic end points and their ability to reflect the underlying basis of induced heritable translocations.

Better understanding of toxicological mechanisms, enhanced testing capabilities, and demands for more sophisticated data for safety and health risk assessment have generated international interest in improving the current testing paradigm for agricultural chemicals. To address this need, the ILSI Health and Environmental Sciences Institute convened a large and diverse group of international experts to develop a credible and viable testing approach that includes scientifically appropriate studies that are necessary without being redundant, and that emphasize toxicological endpoints and exposure durations that are relevant for risk assessment. Benefits of the proposed approach include improved data for risk assessment, greater efficiency, use of fewer animals, and better use of resources. From the outset of this endeavor, it was unanimously agreed that a tiered approach should be designed to incorporate existing knowledge on the chemistry, toxicology, and actual human exposure scenarios of the compound, with integration of studies on metabolism/kinetics, life stages, and systemic toxicities. Three international task forces were charged with designing study types and endpoints on metabolism/kinetics, life stages, and systemic toxicities to be used in the tiered approach. This tiered testing proposal departs from the current standardized list of hazard studies used by many national authorities, and represents the first comprehensive effort of its kind to scientifically redesign the testing framework for agricultural chemicals.

A proposal has been developed by the Agricultural Chemical Safety Assessment (ACSA) Technical Committee of the ILSI Health and Environmental Sciences Institute (HESI) for an improved approach to assessing the safety of crop protection chemicals. The goal is to ensure that studies are scientifically appropriate and necessary without being redundant, and that tests emphasize toxicological endpoints and exposure durations that are relevant for risk assessment. The ACSA Systemic Toxicity Task Force proposes an approach to systemic toxicity testing as one part of the overall assessment of a compound's potential to cause adverse effects on health. The approach is designed to provide more relevant data for deriving reference doses for shorter time periods of human exposure, and includes fewer studies for deriving longer term reference doses-that is, neither a 12-month dog study nor a mouse carcinogenicity study is recommended. All available data, including toxicokinetics and metabolism data and life stages information, are taken into account. The proposed tiered testing approach has the potential to provide new risk assessment information for shorter human exposure durations while reducing the number of animals used and without compromising the sensitivity of the determination of longer term reference doses.

Hazard identification and risk assessment paradigms depend on the presumption of the similarity of rodents to humans, yet species specific responses, and the extrapolation of high-dose effects to low-dose exposures can affect the estimation of human risk from rodent data. As a consequence, a human relevance framework concept was developed by the International Programme on Chemical Safety (IPCS) and International Life Sciences Institute (ILSI) Risk Science Institute (RSI) with the central tenet being the identification of a mode of action (MOA). To perform a MOA analysis, the key biochemical, cellular, and molecular events need to first be established, and the temporal and dose-dependent concordance of each of the key events in the MOA can then be determined. The key events can be used to bridge species and dose for a given MOA. The next step in the MOA analysis is the assessment of biological plausibility for determining the relevance of the specified MOA in an animal model for human cancer risk based on kinetic and dynamic parameters. Using the framework approach, a MOA in animals could not be defined for metal overload. The MOA for phenobarbital (PB)-like P450 inducers was determined to be unlikely in humans after kinetic and dynamic factors were considered. In contrast, after these factors were considered with reference to estrogen, the conclusion was drawn that estrogen-induced tumors were plausible in humans. Finally, it was concluded that the induction of rodent liver tumors by porphyrogenic compounds followed a cytotoxic MOA, and that liver tumors formed as a result of sustained cytotoxicity and regenerative proliferation are considered relevant for evaluating human cancer risk if appropriate metabolism occurs in the animal models and in humans.

This paper summarises and extends the work developed over the last decade by the Joint FAO/WHO Meeting on Pesticide Residues (JMPR) for acute health risk assessment of agricultural pesticides. The general considerations in setting of acute reference doses (ARfDs) in a step-wise process, as well as specific considerations and guidance regarding selected toxicological endpoints are described in detail. The endpoints selected are based on the practical experience with agricultural pesticides by the JMPR and are not a comprehensive listing of all possible relevant endpoints. Haematotoxicity, immunotoxicity, neurotoxicity, liver and kidney toxicity, endocrine effects as well as developmental effects are taken into account as acute toxic alerts, relevant for the consideration of ARfDs for pesticides. The general biological background and the data available through standard toxicological testing for regulatory purposes, interpretation of the data, conclusions and recommendations for future improvements are described for each relevant endpoint. The paper also considers a single dose study protocol. This type of study is not intended to be included in routine toxicological testing for regulatory purposes, but rather to guide further testing when the current database indicates the necessity for an ARfD but does not allow a reliable derivation of the value.

Defining the mode(s) of action by which chemicals induce tumors in laboratory animals has become a key to judgments about the relevance of such tumor data for human risk assessment. Frameworks for analyzing mode of action information appear in recent U.S. EPA and IPCS publications relating to cancer risk assessment. This FORUM paper emphasizes that mode of action analytical frameworks depend on both qualitative and quantitative evaluations of relevant data and information: (1) presenting key events in the animal mode of action, (2) developing a "concordance" table for side-by-side comparison of key events as defined in animal studies with comparable information from human systems, and (3) using data and information from mode of action analyses, as well as information on relative sensitivity and exposure, to make weight-of-evidence judgments about the relevance of animal tumors for human cancer assessments. The paper features a systematic analysis for using mode of action information from animal and human studies, based in part on case examples involving environmental chemicals and pharmaceuticals.

Disinfection by-products (DBPs) are formed when disinfectants such as chlorine, chloramine, and ozone react with organic and inorganic matter in water. The observations that some DBPs such as trihalomethanes (THMs), di-/trichloroacetic acids, and 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) are carcinogenic in animal studies have raised public concern over the possible adverse health effects of DBPs. To date, several hundred DBPs have been identified. To prioritize research efforts, an in-depth, mechanism-based structure-activity relationship analysis, supplemented by extensive literature search for genotoxicity and other data, was conducted for ranking the carcinogenic potential of DBPs that met the following criteria: a) detected in actual drinking water samples, b) have insufficient cancer bioassay data for risk assessment, and c) have structural features/alerts or short-term predictive assays indicative of carcinogenic potential. A semiquantitative concern rating scale of low, marginal, low-moderate, moderate, high-moderate, and high was used along with delineation of scientific rationale. Of the 209 DBPs analyzed, 20 were of priority concern with a moderate or high-moderate rating. Of these, four were structural analogs of MX and five were haloalkanes that presumably will be controlled by existing and future THM regulations. The other eleven DBPs, which included halonitriles (6), haloketones (2), haloaldehyde (1), halonitroalkane (1), and dialdehyde (1), are suitable priority candidates for future carcinogenicity testing and/or mechanistic studies.

The use of structured frameworks can be invaluable in promoting harmonization in the assessment of chemical risk. IPCS has therefore updated and extended its mode of action (MOA) framework for canc...

A risk estimate of the heritable effects of ethylene oxide exposure, using the parallelogram approach, as suggested by Frits Sobels, is described. The approach is based on available data on the ethylene oxide-induced responses for the same genetic endpoint in somatic cells of both laboratory animals and humans, and for germ cell mutations in the same laboratory animal. Human germ cell effects are estimated. The available data sets for this approach were evaluated. We consider this as complementary to the genetic risk assessment carried out by U.S. EPA scientists, in which the risk from heritable (reciprocal) translocations induced by ethylene oxide was estimated. In the present study we restricted our assessment to dominant mutations. The sensitivity factor relating mouse to man was based on ethylene oxide-induced HPRT mutant frequencies in lymphocytes in vivo. From this comparison, it could be concluded that occupational exposure for 1 year to 1 ppm ethylene oxide would lead to a risk of a dominantly inherited disease in the offspring of 4 × 10−4 above the background level. The uncertainty interval of this figure is quite large (0.6–28) × 10−4. The values are compatible with the existing estimates of the corresponding risk from exposure to low LET radiation when the genotoxic potency ratio of ethylene oxide and radiation is considered. This risk estimation approach has allowed us to identify additional data that are required for a more complete risk estimation of the heritable effects of ethylene oxide, or indeed any mutagenic chemical.

The U.S. Environmental Protection Agency (EPA) recently sponsored a workshop† to discuss (1) the contribution of aneuploidy to human disease and disability, (2) the development of tests for detecting chemicals that induce aneuploidy and the relevance of these tests to human risk, and (3) the current understanding of mechanisms by which aneuploidy arises. This summary is based on the presentations given at the workshop. It is hoped that this summary will stimulate thinking in this vitally important area of risk assessment and contribute to the establishment of priorities for basic research, development of new test methods, and validation of existing test approaches. Such research is needed to enhance the scientific basis of risk assessment for aneuploidy-producing chemicals.