Chapter 35: Regulatory Toxicology

Regulatory toxicology is a subfield of regulatory science, which addresses the intersection of science with regulation, namely how is science developed, evaluated, and applied in regulatory decision making. Toxicology is one of the most common fields of scientific knowledge utilized in regulatory science, as regulatory agencies often are required to identify and quantify the health risks of the products and activities they seek to regulate. Regulatory toxicology, which must straddle traditional scientific methodologies with the public policy world of regulation, strains to adhere to the standards of good toxicological science while also providing relevant inputs to decision making, which often asks questions that science cannot answer completely or with any certainty. It is the tension between these two different worlds of science and policy that raise many of the most important and challenging issues in regulatory toxicology, many of which will be addressed in this chapter.

Over the past 40 years, the field of regulatory toxicology has grown enormously as the intersection between regulation and toxicology has expanded dramatically. With the enactment of a series of environmental, health, and safety statutes in the 1970s and 1980s, regulatory agencies increasingly rely on toxicological science to identify potential hazards, prioritize chemicals and other potentially toxic substances, and provide the data used for assessing risk. Regulators are not merely consumers of toxicological studies but also help shape toxicology science in important ways. Regulatory programs have provided a major impetus for improvements in toxicology methods, and they have stimulated a demand for toxicology studies that meet various regulatory requirements. Some programs, such as the programs of the Food and Drug Administration (FDA) for licensing drugs, devices, and food additives and that of the Environmental Protection Agency (EPA) for registering pesticides, explicitly demand toxicology studies as a condition for marketing products. Other statutory and regulatory requirements, such as the recent program requiring EPA to screen chemicals for their endocrine disruptor capability, or agencies to evaluate the safety of new nanotechnology applications are a major driver for the development of new types of toxicological assays.

Regulatory agencies have exercised important influence over the design and conduct of toxicology studies. For example, the EPA is empowered by the Toxic Substances Control Act (TSCA) to promulgate standards for different types of toxicology (and other scientific) investigations. Likewise, the FDA has long issued guidelines for laboratory studies submitted in support of food additives and drugs. Both agencies have adopted requirements governing laboratory operations and practice. Communication between government officials and laboratory scientists flows in both directions. Government testing standards are influenced strongly by the prevailing consensus among toxicologists, many of whom work in regulatory agencies.

There are critical differences between the objectives and methods of science and regulatory policy. Science seeks to understand natural phenomena through objective, empirical, and neutral methodologies, and is cautious, incremental, and evidence-based. In contrast, government regulation seeks to affect human behavior and resolve human disputes; it is episodic, peremptory, and normative, pursuing goals such as well-being, efficiency, and fairness.

Science is open-ended and can continue to study a problem indefinitely without coming to any final resolution. As the National Research Council noted, “the scientific process of seeking the truth, by design and to its credit, has no natural endpoint” (NRC, 2009). In contrast, regulation usually does not have the luxury of waiting until the science has matured, but rather must come to decisions based on legislative, political, and societal deadlines. This creates two problems. First, finality is essential for regulations, so that regulated entities and other stakeholders have certainty about the applicable regulatory requirements, which often involve significant resources and leadtime to achieve compliance. Consequently, regulations are adopted as final decisions, which can only be reconsidered at significant administrative burden. Science meanwhile continues to develop, resulting in regulations that often become increasingly out-dated in relationship to evolving scientific data and understanding.

The second problem created by the need to adopt regulations on a set timetable is often that the scientific inputs into that regulatory decision remain uncertain and underdeveloped. This leads to what Alvin Weinberg referred to as the “regulator’s dilemma”: “the regulator, by law, is expected to regulate even though science can hardly help him; this is the regulator’s dilemma” (Weinberg, 1985). Regulatory agencies are therefore required to make numerous assumptions to fill the gaps in the available scientific data. This then has the inevitable effect of making agency determinations have a mixed science–policy nature, which can lead to confusion and obfuscation.

Scientists are also often frustrated by the political and administrative requirements (often referred to colloquially as “red tape”) associated with regulatory utilization of toxicological data. A regulatory proposal incorporating scientific evidence has to pass through many layers within a regulatory office, including the agency scientists, economists, program officers, office of the general counsel, political appointees, various Centers, Programs and Offices within the agency that are affected by the proposed decision, external advisory committees, and finally the agency head, subject at every stage to a bevy of internal control requirements and processes. When the agency has finally made its proposed decision, the proposal is often then scrutinized by the Office of Information and Regulatory Affairs (OIRA) in the White House, courts conducting judicial review, Congressional oversight committees, and of course external stakeholders and the media. At every step along this pathway, the scientific findings of the agency are subject to challenge, prodding, and second-guessing.

This elaborate regulatory process has a number of implications. First, regulatory agencies tend to attempt to buttress their regulatory decisions with extensive analytical support. As a comparative study of US and European regulation observed, “American regulators, being more politically exposed than their European counterparts, have a greater need to support their actions through formal analytical arguments. Their demand for expertise necessitates a continuing build-up of technical capabilities and the development of more sophisticated and detailed analytical methodologies. The structure of the American rule-making process subjects the analytical case for regulation to intense political scrutiny. Any weaknesses are quickly exploited, and the uncertainties and shortcomings of the relevant scientific base are readily exposed.…” (Brickman et al., 1985). In response to these pressures, regulatory agencies will often compile extensive scientific assessments of the toxicological and other data supporting their regulatory decisions, sometimes thousands of pages in length.

Another consequence of the highly politicized and adversarial nature of regulatory decision making is that regulatory agencies have an incentive to overstate the scientific basis of their decisions. Although scientific data are a critical input into sound-regulatory decision making, the actual decision of how much protection to provide and which requirements or limits to impose is inevitably a normative policy decision. Yet, agencies receive greater deference from Congress, the White House, reviewing courts, and even the media and public if they justify their decision on scientific rather than policy grounds. Accordingly, regulatory agencies are inclined to give disproportionate weight to scientific arguments and evidence in trying to justify their decisions, a tendency that has been labeled in the literature as the “science charade” (Wagner, 1995).

Different agencies and even different regulatory programs within the same agency use toxicological data in different ways. One important difference relates to the extent of premarket analysis or approval required before a new product may be commercialized. Some products require premarket approval from a regulatory agency (eg, drugs and medical devices by FDA and pesticides by EPA), other products require only premarket notification to the agency (eg, new chemical substances), whereas still other products require no premarket activity with a regulatory agency (eg, foods, cosmetics) before they can be sold commercially. There are also differences in who has the burden of proof to demonstrate safety or hazard. For example, the manufacturer of a new food additive has the burden to demonstrate to the FDA the lack of hazard before humans may be exposed, while laws such as the Occupational Safety and Health Act (1970) put the burden on regulators to show that a substance is hazardous before exposures can be restricted. The approach chosen by Congress for a particular type of product or activity greatly influences an agency’s ability to require comprehensive toxicological investigation prior to commercialization.

In determining what level of a substance to allow, regulatory agencies generally use one of three different approaches: (i) acceptable risk, (ii) balancing, or (iii) feasibility. The determination of which of these three approaches is used for a particular hazard is generally determined by the applicable statutory framework.

Acceptable Risk

This approach determines the level of risk a substance presents, and then seeks to lower that risk (if necessary) to a level that is “acceptable.” Because these risk-based programs only consider health evidence, toxicological data play a central role in these regulatory programs. In making the determination of acceptable risk, agencies have traditionally distinguished between carcinogens and noncarcinogens in their approach. For noncarcinogenic chemicals, regulators have generally embraced a standard safety assessment formula, built around the concept of acceptable daily intake (ADI) for the FDA or reference dose (RfD) or reference concentration (RfC) in the case of EPA. These levels are derived by applying a series of safety factors to the lowest “no observed adverse effect level” from animal experiments. Human exposures to a chemical that fall below the resulting ADI, RfD, or RfC are assumed to be “safe”or acceptable. This traditional approach to noncarcinogenic toxicants has not been considered appropriate for carcinogens. Regulators in the United States and many of their counterparts in other countries have operated on the premise that carcinogens as a class cannot be assumed to have “safe” or threshold doses. Furthermore, they have assumed that any chemical shown convincingly in animal studies to cause cancer should be considered a potential human carcinogen. Accordingly, for this group of compounds, regulators have generally assumed that no finite level of human exposure can be considered risk-free. As research has begun to illuminate the different mechanisms by which chemicals may cause cancer, however, regulatory agencies have cautiously accepted the possibility that “safe” thresholds may be established for some nongenotoxic carcinogens.

An example of an “acceptable” risk-regulatory program is the national ambient air quality program under the Clean Air Act (CAA). Under this program, EPA must set standards for criteria air pollutants such as particulate matter, ozone, and carbon monoxide that protect against “adverse effects” in the most susceptible subgroup with an adequate margin of safety. EPA may only consider risk data in setting these standards; no consideration of costs or feasibility is permitted (Whitman v ATA, 2001). A more stringent form of the acceptable risk approach is a “zero” or “negligible” risk requirement, epitomized by the famous Delaney clause, enacted in 1958 as part of the Food Additives Amendment. The amendment itself requires that any food additive be found “safe” before the FDA may approve its use [Food Additive Amendment, 1958]. The Delaney clause stipulated that this finding may not be made for a food additive that has been shown to induce cancer in humans or in experimental animals. In 1996, Congress amended the provisions of the Food, Drug, and Cosmetics Act applicable to pesticide residues in food, revoking the Delaney proviso as it applies to food additives (but not color addictives), and adopting a standard that, though not expressed in these words, is understood to permit a tolerance for a carcinogenic pesticide if the estimated cancer risk is extremely small, on the order of one in one million (Food Quality Protection Act, 1996).

Balancing Approaches

Other regulatory programs require the agency to balance the health benefits of risk reduction against the costs of such reductions. These approaches known as cost-benefit or risk-benefit analysis generally require a quantified estimate of both the health risks and costs associated with various regulatory options, and the agency then chooses the regulatory option that offers the biggest estimated net benefit. Both toxicological and economic data play key roles in these statutory programs for calculating the benefits and costs of regulation, respectively. The TSCA (1976) is an example of a statute utilizing a balancing approach, requiring EPA to set standards for toxic substances based on a quantified cost-benefit analysis (Corrosion Proof Fittings v EPA, 1991).

Feasibility/Best Available Technology

The third major approach requires the agency to reduce exposures to the lowest feasible level, or to require companies to install the best available technology. This approach does not explicitly rely on any health data, and this skirting of the complexities and uncertainties of toxicological evidence is seen as a major advantage of this approach. The consequence of ignoring health impacts, though, is that the resulting standards may overprotect or underprotect health. An example of such a feasibility-based approach is the Clean Water Act’s effluent standards, which requires dischargers of pollutants to install the best available technology determined to be feasible for that industrial category. A program using a hybrid feasibility/acceptable risk approach is the Occupational Safety and Health Act, which directs the Occupational Safety and Health Administration (OSHA) to set workplace standards for hazardous exposures to the level “which most adequately assures, to the extent feasible… that no employee will suffer material impairment of health or functional capacity” (OSHA, 1970). This language has been interpreted by OSHA and the courts as requiring the agency to set occupational health standards at the lowest exposure level that is technologically and economically feasible, to the extent necessary to reduce a “significant” risk (American Textile Manufacturers Institute v Donovan, 1982).

At the federal level, four agencies are chiefly responsible for regulating human exposure to toxic materials: the FDA, EPA, OSHA, and the Consumer Product Safety Commission (CPSC). Together they administer over two-dozen statutes whose primary goal is the protection of health. The statutes administered by these four agencies utilize different regulatory approaches and require divergent safety benchmarks. This diversity has several explanations. The statutes were enacted in different eras. They originated with different political constituencies and remain under their influence. Perhaps most significant, statutory standards often reflect differences in the technical capacity to control different types of exposures, and they embody different Congressional judgments about the economic implications of limiting exposures.

Food and Drug Administration

The oldest of the major health regulation laws, the FFD&C Act, was originally enacted in 1906 and, as amended over time, now covers food for humans and animals, human and veterinary drugs, medical devices, and cosmetics.

Food

The initial Food and Drug Act enacted in 1906 prohibited “adulterated” foods that contain “any poisonous or deleterious substance, which may render it injurious to health.” This provision, which remains in the statute to date [FFD&C Act § 402(a)], does not require premarket approval of foods and puts the burden of proof on FDA to demonstrate that a food meets the definition of “adulterated.”

Congress has since amended the act several times to strengthen the FDA’s ability to ensure the safety of foods (FFD&C, 1938). The most important of these amendments was the 1958 Food Additives Amendments that require the safety of food additives to be demonstrated prior to marketing (Food Additive Amendments, 1958). The manufacturer of a food additive must submit a petition demonstrating that the substance is “reasonably certain to be safe”; no inquiry into the benefits of an additive is undertaken or authorized (Cooper, 1978). Two categories of nonnaturally occurring food ingredients were exempted from this food additive requirement: (a) substances generally recognized as safe among experts qualified by scientific training and experience to evaluate safety; and (b) substances that either FDA or the US Department of Agriculture (USDA) had sanctioned for use in food prior to 1958 (so-called “prior sanction” substances).

Separate regulatory standards have been developed for several classes of indirect food constituents. For example, a food-contact substance requires approval as a food additive if, when used as intended, it “may reasonably be expected to become a component of food.” In 1997 omnibus legislation that addressed most of FDA’s regulatory programs, Congress created a new premarket notification system for food-contact materials that permits the agency to delay introduction if it questions a material’s safety, but does not require affirmative agency approval (FDA Modernization Act, 1997).

Human Drugs

Preclinical studies in animals play an important role in the FDA’s evaluation of human drugs. The current law requires premarket approval, for both safety and efficacy, of all “new” drugs, a category that embraces virtually all prescription drug ingredients introduced since 1938 (Hutt et al., 2007). Premarket approval of therapeutic agents for commercial use primarily relies on randomized controlled trials in human subjects to establish safety and efficacy. However, animal and in vitro studies are the primary source of information about a substance’s biological effects before human trials are begun, and their results influence not only the decision whether to expose human subjects but also the design of clinical protocols. The FDA reviews the clinical evidence before authorizing human clinical trials through review of an Investigational New Drug (IND) application.

Medical Devices

In 1976, Congress overhauled the FFD&C Act’s requirements for medical devices, providing the FDA major new authority to regulate their testing, marketing, and use. The elaborate new scheme contemplates three tiers of control, the most restrictive of which (class III) is premarket approval similar to that required for new drugs. To obtain FDA premarket approval of a class III device, the sponsor must demonstrate safety and efficacy. The bulk of the data supporting such applications will be derived from clinical studies but also will include toxicology studies of any constituents likely to be absorbed by the patient.

Cosmetics

The statutory provisions governing cosmetics do not require premarket approval of any ingredient, or demand that manufacturers test their products for safety, though many manufacturers routinely do so. The basic safety standard for cosmetics is similar to that for food ingredients: no product may be marketed if it contains “a poisonous or deleterious substance, which may render it injurious to health” [FFD&C Act § 601(a)]. The case law establishes that this language, too, bars distribution of a product that poses any significant risk of more than transitory harm when used as intended, but it places on the FDA the burden of proving a violation (Hutt et al., 2007). The FDA has brought few cases under this standard, in part because acute toxic reactions to cosmetics are readily detected and immediately result in abandonment of the offending ingredient. The FDA also relies on the Cosmetic Ingredient Review (CIR), a private expert assessment body established in 1976 and operated by the Personal Care Products Council. The CIR appoints an expert panel that conducts safety assessments of cosmetics ingredients that are subject to public review and comment and then published as monographs.

Environmental Protection Agency

The EPA is responsible for administering most of the nation’s environmental laws, which include protection of both human health and the environment. A comprehensive review of the EPA’s numerous programs is not possible here; the following summary focuses on those EPA activities in which toxicology evidence plays a central role: the agency-wide Integrated Risk Information System (IRIS) program, pesticide regulation, regulation of industrial chemicals, regulation of drinking water supplies, hazardous waste control, and regulation of toxic pollutants of water and of air.

IRIS Program

EPA’s IRIS provides an agency-wide scientific assessment program to evaluate the carcinogenic and noncarcinogenic human health risks of various chemicals. EPA prepares a Toxicological Review for a chemical, which goes through a series of internal and external reviews before being finalized, that is then used to derive a RfD and/or inhalation RfC for noncarcinogenic effects, and a weight of evidence determination as well as oral and inhalation unit risks for carcinogenic effects. IRIS values are used to support a variety of agency regulatory programs, including many of those described below. As of 2011, the agency has established IRIS values for over 550 chemical substances. EPA is in the process of reforming its IRIS process in response to criticisms of recent reviews, including a critical National Research Council study of EPA’s IRIS assessment of formaldehyde (NRC, 2011).

Pesticides

Pesticides are subject to two types of regulation: (i) registration under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA, 1972), which is required for a pesticide to be marketed in the United States; and (ii) tolerance (under the FFD&C Act), which sets the allowable level of pesticide residue in food for human consumption. Pesticide residues on raw agricultural commodities for years were initially regulated by the EPA under a 1954 amendment to the FFD&C Act, which allowed the agency to consider in setting tolerances both the potential adverse health effects of residues and the benefits of pesticide uses. If the concentration of a pesticide in a processed food exceeded the established tolerance, or the pesticide was a carcinogen, the Delaney clause prohibited its approval for that use (Les v Reilly, 1992). This framework was revised by Congress in 1996 by the Food Quality Protection Act (FQPA) to exempt pesticide residues from the operation of the Delaney clause but also restricted the consideration of pesticide benefits in setting tolerances.

The FQPA integrated the standard for registration and tolerance of a pesticide to require a pesticide to be shown to be safe, a standard defined as “reasonable assurance of no harm” (FQPA, 1996). The 1996 amendments also introduced some additional novel elements to the tolerance process. First, it required EPA to determine the safe level of a pesticide residue in a cumulative context, considering “all anticipated dietary exposures and all other exposures for which there is reliable information” [21 USC § 346a(b)(2)(A)(ii)]. This includes other nonoccupational sources of exposure, including food, drinking water, and residential exposures [21 USC § 346a(b)(2)(D)(vi)]. EPA must also base its safety determination on a consideration of exposures to other pesticides that share the same mechanism of toxicity [21 USC § 346a(b)(2)(C)(i)(III)].

Second, the 1996 FQPA required EPA to give special consideration to the health of infants and children, by ensuring “that there is a reasonable certainty that no harm will result to infants and children from aggregate exposure to the pesticide chemical residue.…” [21 USC § 346a(b)(2)(C)(ii)(I)]. To implement this requirement, the statute directs EPA to “use an additional 10-fold margin of safety in assessing the risks to infants and children to take into account potential for pre- and postnatal toxicity and the completeness of the toxicology and exposure databases.” EPA can deviate from the 10-fold margin of safety “only if, based on reliable data, the resulting margin would be safe for infants and children” [21 USC § 346a(b)(2)(C)(ii)(I)]. EPA’s failure to provide a legally sufficient rationale for not applying the presumptive 10-fold extra safety factor for children has been found to be arbitrary and capricious by reviewing courts (NRDC v EPA, 2011; Northwest Coalition for Alternatives to Pesticides v EPA, 2008).

Third, EPA must develop a program to test for endocrine disrupting effects of pesticides. There has been a flurry of research over the past couple of decades addressing the identification and mechanisms of toxicity from endocrine-disrupting chemicals (Marty et al., 2011). The FQPA required EPA to “develop a screening program, using appropriate validated test systems and other scientifically relevant information, to determine whether certain substances may have an effect in humans that is similar to an effect produced by a naturally occurring estrogen, or such other endocrine effect as the Administrator may designate” [21 USC § 346a(p)(1)]. Although EPA was required to establish this program by 1999, it did not meet this deadline.

EPA has been moving forward nevertheless and has created an Endocrine Disruptor Screening Program, which involves a two-tier screening and testing process: In Tier 1 testing, EPA seeks to identify chemicals that have the potential to interact with the endocrine system. Tier 1 consists of 11 different assays, which are being applied to 67 pesticide ingredients on EPA’s initial list of chemicals to be tested, which was released in 2009. This initial round of testing was scheduled to be completed in February 2012. EPA proposed in 2010 a second list of 134 additional chemicals to be tested in Tier 1. In Tier 2, EPA will determine the endocrine-related effects caused by each chemical that indicated endocrine disruption activity in Tier 1, and obtain information about effects at various doses in order to enable risk assessment. The agency is in the process of developing and validating Tier 2 tests. One issue is whether in Tier 2 EPA should use traditional animal assays or in vitro and computational assays that are part of EPA’s ToxCast program.

After the FQPA of 1996, the pesticide registration process now includes a broader review of the human health impacts of the pesticide from commercial use of the pesticides on applicators and other individuals, as well as effects on the environment. The manufacturer of a pesticide must produce a variety of data from tests done according to EPA guidelines. EPA’s test guidelines, most recently revised in 2007, now encompass guidelines for 340 types of tests recommended for pesticides (40 CFR 158 and 161). These tests evaluate whether a pesticide has the potential to cause adverse effects on humans, wildlife, fish, and plants, as well as possible contamination of surface water or ground water from leaching, runoff, and spray drift. The potential human risks that are evaluated include both acute and chronic toxicity. The EPA then conditions the use of the pesticide on various protections to ensure that the pesticide is used safely, which are summarized on the pesticide label.

The EPA has been engaged in a comprehensive review of previously registered pesticides and “reregistration” of those that meet contemporary standards for marketing for approximately the past three decades. Under this program, many older pesticides have been subjected to comprehensive toxicology testing, including carcinogenicity testing, for the first time, and the results have required modification of the terms of approved use and, in some instances, cancellation for several agents. The FQPA required EPA to review the registration (“reregistration”) and tolerance of older pesticides. In addition, EPA initiated a new program in 2006 called “registration review” in which it intends to re-evaluate all pesticides every 15 years. Finally, EPA may initiate a “Pesticide Special Review” process whenever it discovers that the use of a registered pesticide may result in unreasonable adverse effects on people or the environment.

Industrial Chemicals

The TSCA was promulgated in 1976 (TSCA, 1976) to regulate all chemical substances manufactured or processed in or imported into the United States, except for substances already regulated under other laws. A chemical substance is defined broadly as “any organic or inorganic substance of a particular molecular identity.”

TSCA gives the EPA three main powers. First, the agency is empowered to restrict or even ban the manufacturing, processing, distribution, use, or disposal of a chemical substance when there is a reasonable basis to conclude any such activity poses an “unreasonable risk of injury to health or environment.” In determining whether a chemical substance presents an unreasonable risk, the agency is instructed to balance the costs of restricting the substance against the health benefits of the proposed regulation [TSCA § 6]. This trade-off approach to regulation has proved a major challenge to the EPA. The agency has exercised its regulatory authority under section 6 against only a handful of substances, and it’s most significant regulations, such as its attempted ban on asbestos products, have been struck down by the courts for failing to do an adequate cost-benefit analysis (Corrosion Proof Fittings v EPA, 1991).

Second, if the EPA suspects that a chemical may pose an unreasonable risk but lacks sufficient data to take action, TSCA empowers the agency to require testing to develop the necessary data. Similarly, it may order testing if the chemical will be produced in substantial quantities that may result in significant human exposure whose effects cannot be predicted on the basis of existing data. In either case, the EPA must consider the “relative costs of the various test protocols and methodologies” and the “reasonably foreseeable availability of the facilities and personnel” needed to perform the tests [TSCA § 4].

Finally, to enable the EPA to evaluate chemicals before humans are exposed, TSCA requires the manufacturer of a new chemical substance to notify the agency 90 days prior to production or distribution [TSCA § 5(a)(1)]. The manufacturer’s or distributor’s notice (called a pre-manufacturing notice or PMN) must include any health effects data it possesses. However, the EPA is not empowered to require that manufacturers routinely conduct testing of all new chemicals to permit an evaluation of their risks; Congress declined to confer the kind of premarket approval authority that the FDA exercises for drugs and food additives and the EPA exercises for pesticides.

In contrast, the European Union (EU) is implementing a much more aggressive regulatory program for chemical substances, known as the Registration, Evaluation, Authorization, and Restriction of Chemical substances (REACH) program. REACH puts the responsibility for ensuring the safety of chemicals on the companies that produce, import, and use chemicals. The program applies to both new and existing chemicals, and is phased-in over an 11-year period based primarily on the production volume of individual chemicals. REACH employs a tiered-testing approach, with the amount of initial testing determined by production volume. A technical dossier is required for all registered chemicals; however, only chemicals with a production volume of 10 or more tons per year require a chemical safety assessment documented in a chemical safety report. REACH allows the use of alternative testing strategies to evaluate specific endpoints in lieu of experimental testing; however, whole animal testing requirements exist for each level based on production volume.

In light of the limitations of TSCA and the more proactive approach of the EU REACH program, there are active proposals in Congress to modify and update TSCA. Although the details of TSCA reform remain undecided at the time of this writing, there is broad consensus that some type of reform is needed.

Hazardous Wastes

The two primary statutes regulating hazardous wastes in the United States are the Resource Conservation and Recovery Act (RCRA, 1976) and the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA, 1980). With some exceptions, the primary purpose of CERCLA is the cleanup of contamination from pre-existing hazardous waste sites, whereas RCRA regulates the generation, transport, and disposal of new hazardous wastes. Both RCRA and CERCLA have their own extensive lists of substances and waste streams of toxic materials that are defined as hazardous wastes. In addition, under RCRA, wastes that meet any of four “characteristics” are also treated as hazardous wastes. The four characteristics are ignitability, corrosivity, reactivity, and toxicity, and EPA has identified standardized protocols for determining whether a particular waste stream meets one or more of these characteristics. Under RCRA, EPA has developed standards for generators, transporters, and those who treat, store, or dispose hazardous wastes that are required to “protect human health and the environment.” The EPA’s regulations applicable to generators and transporters establish a manifest system that is designed to create a paper trail for every shipment of waste, from generator to final destination, to ensure proper handling and accountability. The agency has the broadest authority over persons who own or operate hazardous waste treatment, storage, or disposal facilities. Pursuant to the RCRA, it has issued regulations prescribing methods for treating, storing, and disposing of wastes; governing the location, design, and construction of facilities; mandating contingency plans to minimize negative impacts from such facilities; setting qualifications for ownership, training, and financial responsibility; and requiring permits for all such facilities (RCRA, 1976).

CERCLA, often referred to as Superfund, authorizes the cleanup of hazardous waste sites and provides for liability of persons responsible for releases of hazardous waste at these sites. EPA can conduct cleanup of hazardous waste sites itself, and then seek recovery for the costs from potentially responsible parties (PRPs), or it can direct one or more PRPs to conduct cleanup, who may then seek contribution from other PRPs for their share of the cleanup. Under CERCLA, EPA has created the National Priorities List (NPL), which is a list of hazardous waste sites that are deemed the most hazardous sites and eligible for use of federal funds for long-term clean-up, and the National Contingency Plan (NCP), which provides the guidelines and procedures needed to respond to releases and threatened releases of hazardous substances, pollutants, or contaminants. For sites on the NPL, EPA conducts a Remedial Investigation and Feasibility Study to support the risk management process, which includes a site-specific human health and ecological risk assessment.

Toxic Water Pollutants

The EPA has had responsibility for regulating toxic water pollutants since 1972. As originally enacted, Section 307 of the Federal Water Pollution Control Act required the EPA to adopt standards for toxic water pollutants that provided an “ample margin of safety”—a difficult criterion to meet for most toxic pollutants and arguably impossible for any known to be carcinogenic. Given EPA’s difficulties in complying with this statutory directive, Congress amended the statute in 1977 to focus standards on economic cost and technologic feasibility. In 1987, Congress amended the statute again to toughen standards for toxic pollutants. Under the prior law, the EPA had developed health-based “water quality criteria” for 126 compounds it had identified as toxic. These criteria essentially described desirable maximum contamination levels, which, because the EPA’s discharge limits were technology-based, generally were substantially lower than the levels actually achieved. The 1987 amendments gave what had been advisory criteria real bite by requiring that states incorporate them in their own mandatory standards for water quality and impose additional effluent limits on operations discharging into below-standard waterways (Heineck, 1989).

Drinking Water

The 1974 Safe Drinking Water Act (SDWA) was enacted to ensure that public water supply systems “meet minimum national standards for the protection of public health.” Under the SDWA, the EPA is required to regulate any contaminants “which may have an adverse effect on human health.” Over time and several Congressional amendments, this system has evolved into a two-tier system in which EPA adopts for each contaminant a maximum contaminant level goal (MCLG) that represents the optimal level to be achieved to ensure safety and the usually less stringent, feasibility-based maximum contaminant level (MCL). Only the MCL is legally enforceable. As part of the 1996 Safe Drinking Water Amendments, Congress directed EPA to use the “best available science” in setting MCLGs and MCLs.

Toxic Air Pollutants

The original 1970 version of the CAA provision for regulation of hazardous air pollutants (section 112) required the agency to set emission standards for individual toxic pollutants that protect public health with “an ample margin of safety.” The implication of this language—that standards were to provide absolute safety without regard to the costs of emissions control—generated intense debate from the beginning and contributed to the EPA’s glacial pace of implementation. The EPA finally attempted to escape the strict language of section 112 when it issued a standard for vinyl chloride in 1986 in which the agency claimed it could consider costs and declined to adopt a standard dictated solely by safety. A court ruled that the EPA had improperly considered costs, but it did make clear that in determining what emissions level was safe, even for a carcinogen, the EPA was not obligated to eliminate exposure but rather was to “decide what risks are acceptable in the world in which we live” (NRDC v EPA, 1987).

Amendments to the CAA in 1990 responded to the difficulties presented by the strict health-based approach of old section 112. The statute now provides a list of 188 hazardous air pollutants, which EPA may modify by adding or deleting items. The EPA must establish national emissions standards for sources that emit any of the listed pollutants with a two-tier system of regulation. The EPA must first issue standards that are technology-based, designed to require the “maximum achievable control technology” (MACT) [CAA § 112(d)(2)]. If the MACT controls are insufficient to protect human health with an “ample margin of safety,” the EPA must issue residual risk standards [CAA § 112(f)]. The 1990 Amendments essentially define “ample margin of safety” for carcinogens by requiring the EPA to establish added residual risk limits for any pollutant that poses a lifetime excess cancer risk of greater than one in one million. However, as subsequently clarified in a judicial decision, while EPA must promulgate residual risk standards if the MACT standards do not reduce risks to one in one million, the residual risk standards are not required to reduce the risks to one in one million, but rather to 100 in one million (NRDC v EPA, 2008).

Occupational Safety and Health Administration

The 1970 Occupational Safety and Health Act requires employers to provide employees with safe working conditions and empowers OSHA to prescribe mandatory occupational safety and health standards (OSHA, 1970). OSHA’s most controversial actions involved its attempts to set occupational exposure limits for hazardous chemicals.

The Act specifies that in regulating hazardous occupational exposures, OSHA shall adopt the standard “which most adequately assures, to the extent feasible, on the basis of the best available evidence, that no employee will suffer material impairment of health or physical capacity” [OSHA § 6 (b)(5)]. Judicial challenges to OSHA standards have clarified OSHA’s responsibilities. In a famous case, the US Supreme Court overturned OSHA’s benzene standard because the agency had not shown that prevailing worker exposure levels posed a “significant” health risk (Industrial Union Department, AFL-CIO, v American Petroleum Institute, 1980). This prerequisite proved a major obstacle when OSHA attempted to establish standards for 428 air contaminants in a single proceeding in 1989. Although it found that OSHA’s generic approach to regulation was permissible in theory, a court vacated the standards because OSHA failed to show that each individual contaminant posed a “significant risk” at current levels (American Federation of Labor v OSHA, 1992). However, the Supreme Court earlier upheld OSHA’s cotton dust standard, rejecting arguments that the agency was obligated to balance the costs of individual standards against the benefits of reducing hazardous workplace exposures (American Textile Manufacturers Institute v Donovan, 1982).

Consumer Product Safety Commission

Of the four agencies discussed here, the CPSC has played the least important role in federal efforts to control toxic chemicals to date. The commission was created in 1972 by the Consumer Product Safety Act (CPSA, 1972) with authority to regulate products that pose an unreasonable risk of injury or illness to consumers. The commission is empowered to promulgate safety standards “to prevent or reduce an unreasonable risk of injury” associated with a consumer product. If no feasible standard “would adequately protect the public from the unreasonable risk of injury” posed by a consumer product, the commission may ban the product [CPSA § 8]. In assessing the need for a standard or ban, the agency must balance the likelihood that a product will cause harm, and the severity of harm it will likely cause, against the effects of reducing the risk on the product’s utility, cost, and availability to consumers.

The CPSC also administers the older Federal Hazardous Substances Act (FHSA, 1976). The FHSA authorizes the CPSC to regulate, primarily through prescribed label warnings, products that are toxic, corrosive, combustible, or radioactive or that generate pressure. The FHSA is unusual among federal health laws because it contains detailed criteria for determining toxicity. It defines “highly toxic” in terms of a substance’s acute effects in specified tests in rodents; substances capable of producing chronic effects thus fall within the “toxic” category. The FHSA contains another unique provision [FHSA § 2(h)(2)] specifically addressing the probative weight of animal and human data on acute toxicity: “If the [commission] finds that available data on human experience with any substance indicates results different from those obtained on animals in the above-named dosages or concentrations, the human data shall take precedence.”

The CPSC has prescribed labeling for products containing numerous substances that are acutely toxic. It has also acted to ban from consumer products several substances that pose a cancer risk, including asbestos, vinyl chloride as a propellant, benzene, tris(hydroxymethyl)aminomethane, and formaldehyde (Merrill, 1981).

In 2008, Congress adopted the Consumer Product Safety Improvement Act (2008), which significantly strengthened the CPSC’s enforcement authority, especially for children’s products, in response to a number of high-profile recalls of children’s toys. The Act imposes bans on products containing lead or phthalates that exceed stringent levels, and requires third-party testing of certain children’s products.

As the previous subsection describes, regulatory agencies are required to evaluate and apply toxicological and other scientific data in a variety of statutory contexts. In addition to the statutory language, a number of internal and external institutional control mechanisms and processes greatly influence the agencies’ treatment of toxicological data.

Advisory Committees

Regulatory agencies frequently rely on scientific advisory committees composed of external scientific experts to help guide their scientific determinations, including toxicological evaluations. For example, the EPA has a Scientific Advisory Board (SAB), established by Congress in 1978, to advise the agency on the quality and relevance of the scientific and technical information being used in regulatory programs as well as to review the agency’s research programs. The SAB’s work is done by a relatively large number of subcommittees and panels that focus on special subjects. The EPA also has two other independent scientific advisory committees associated with specific regulatory programs—the Clean Air Science Advisory Committee for air pollution, and the Science Advisory Panel for pesticides. These committees consist of external scientists who are appointed to serve on the committees for specified terms. Similarly, the FDA has over 50 external advisory committees that provide advice on the agency’s significant product approvals and other scientific issues.

Regulatory agencies are not required to automatically follow the advice of their external advisory committees, but frequently do. Failure to follow the independent advice of their own advisory committee tends to call into question the soundness of the agency’s scientific findings, whereas advisory committee endorsement of the agency’s position can provide important scientific credibility and political cover to the agency.

Peer Review

Beyond peer review by agency advisory committees, regulatory agencies frequently utilize external peer review of scientific reports and assessments. In December 2004, the Office of Management and Budget (OMB) formalized the requirement for peer review by issuing its Final Information Quality Bulletin for Peer Review (OMB, 2004). This Peer Review Bulletin requires federal regulatory agencies to ensure that all “influential scientific information” they disseminate is peer-reviewed. Influential scientific information is defined as “scientific information the agency reasonably can determine will have or does have a clear and substantial impact on important public policies or private sector decisions.” The Bulletin specifically identifies “scientific assessments” as one type of information subject to the peer review requirement, which are defined as “an evaluation of a body of scientific or technical knowledge, which typically synthesizes multiple factual inputs, data, models, assumptions, and/or applies best professional judgment to bridge uncertainties in the available information.” These assessments include, but are not limited to, state-of-science reports; technology assessments; weight-of-evidence analyses; meta-analyses; health, safety, or ecological risk assessments; toxicological characterizations of substances; integrated assessment models; hazard determinations; or exposure assessments.

The Peer Review Bulletin directs agencies to choose a peer review mechanism that is appropriate and commensurate with how the information will be used, giving due consideration to the novelty and complexity of the science involved, the relevance of the information to regulatory decision making, the extent of prior peer reviews, and the expected benefits and costs of additional review. The Peer Review Bulletin requires the most rigorous peer review requirements for highly influential scientific assessments.

Risk Assessment Guidance/Guidelines

Regulatory agencies often adopt guidelines to promote consistency and predictability in their treatment of toxicological data and the conduct of risk assessment. For example, EPA has adopted risk assessment guidelines for carcinogenicity, mutagenicity, developmental toxicity, chemical mixtures, exposure analysis, reproductive toxicity, neurotoxicity, and ecotoxicity. The genesis of EPA’s risk assessment guidelines was a series of “cancer principles” prepared largely at the instigation of EPA lawyers in the mid-1970s (Albert, 1994). The agency lawyers were frustrated by the delays in regulating pesticides as a result of scientific disagreement about the mechanisms and causes of human cancer. In an attempt to circumvent such scientific disputes, the lawyers developed a series of 17 “cancer principles” with the assistance of EPA scientists during hearings on the pesticides heptachlor and chlordane. The 17 principles were intended to be nonrebuttable guidelines that would govern and simplify carcinogenicity determinations.

Shortly thereafter, EPA began to develop a more comprehensive set of guidelines for evaluating carcinogens. This led to EPA publishing its “interim” guidelines for carcinogenic risk assessment in 1976, which were subsequently updated and finalized in 1986 (EPA, 1986). The 1986 guidelines primarily consisted of a series of “default options” that were intended as presumptive principles that would apply in the absence of sufficient data to establish an alternative assumption. Some of the most important default options included in EPA’s 1986 carcinogen guidelines included: (i) laboratory animals are assumed to be a valid surrogate for humans in assessing risk; thus, positive cancer results in animal bioassays are taken as evidence of human carcinogenicity; (ii) humans are assumed to be as sensitive as the most sensitive animal species, strain, or sex evaluated in an appropriately designed animal bioassay; (iii) benign tumors are assumed to be as significant as malignant tumors; (iv) the dose–response of humans to potential carcinogens is assumed to be linear all the way to the zero exposure levels with no threshold; and (v) a given intake of a substance is assumed to have the same effect regardless of the rate or route of intake (EPA, 1986). Most of the default options selected by EPA were deliberately chosen to be “conservative,” in that they are intended to estimate the plausible upper-bound of actual risk (NRC, 1994).

EPA’s risk assessment guidelines not only promoted consistency in risk assessment approach and procedure across the broad array of EPA regulatory programs, which use risk assessment, but also furthered the objective of efficiency, by sparing the agency the need to revisit the same controversial issues in each successive rule-making proceeding (Wiltse and Dellarco, 1996). The courts were generally supportive of the predictability and consistency provided by the agency’s risk assessment guidelines, giving greater deference to EPA risk assessments that comported with the guidelines. As one court stated, “EPA’s specific enunciation of its underlying analytical principles, derived from its experience in the area, yields meaningful notice and dialogue, enhances the administrative process, and furthers reasoned agency decision making” (Environmental Defense Fund, Inc, v Environmental Protection Agency, 1976).

Notwithstanding these benefits of risk assessment guidelines, EPA’s implementation of its guidelines was criticized in a 1994 NRC report that found that EPA applied the default principles too rigidly and had failed to articulate clear criteria for departing from the defaults (NRC, 1994). EPA itself acknowledged that “[i]n practice, the agency’s assessments routinely have employed defaults and, until recently, only occasionally departed from them. (EPA, 1996). EPA published draft revisions to its carcinogen guidelines in 1996 (EPA, 1996) that were finalized in 2005 (EPA, 2005) that take a more flexible, data-based approach to default options. As explained by EPA, the revised guidelines “are intended to be both explicit and more flexible than in the past concerning the basis for making departures from defaults, recognizing that expert judgment and peer review are essential elements of the process” (EPA, 1996). The revised guidelines incorporate “a weight-of-the-evidence approach that considers all relevant data in reaching conclusions about the potential human carcinogenicity of an agent” (EPA, 1996).

Judicial Review

Agency regulations are often subject to legal challenges by stakeholders who disagree with the agency’s decision. Under the Administrative Procedure Act, such judicial review tends to be deferential to the agency’s decision, overturning the decision only if it is “arbitrary and capricious” or lacking “substantial evidence.” Nevertheless, a significant proportion of major agency regulations are overturned in whole or in part, or sent back to the agency for further explanation. Scientific determinations, such as toxicological findings, are generally given the highest level of deference under a decision by the US Supreme Court, which held that “[w]hen examining this kind of scientific determination, as opposed to simple findings of fact, a reviewing court must generally be at its most deferential (Baltimore Gas & Elec v NRDC, 1983).

Even with this deference to scientific determinations, there have been many important judicial decisions impacting the way regulatory agencies address toxicological information. One prominent example was a challenge to the EPA’s decision to retain a MCLG of zero under the Safe Drinking Water Act for chloroform in the face of evidence, accepted by its own scientific advisory committee, that the mechanism by which chloroform produces tumors in rodents is threshold-limited. The reviewing court held that given its acceptance of the evidence of a threshold, the EPA was required by the SDWA’s “best available” evidence provision to base its decision on that evidence (Chlorine Chemistry Council v EPA, 2000). This ruling marks the first time that a court has directed a federal regulatory agency to recognize a “safe” finite level of exposure for an animal carcinogen. Another example of proactive judicial oversight is two recent decisions where the court second-guessed EPA’s determination that there was sufficient toxicological evidence to justify departing from the presumptive 10-fold extra safety factor to protect children in pesticide regulatory decisions (NRDC v, EPA, 2011; Northwest Coalition for Alternatives to Pesticides v EPA, 2008).

Congressional Oversight

Congress has ultimate oversight of regulatory agencies, both by delegating the authority to regulate in the form of statutes, and by controlling the purse strings for the agencies. Regulatory statutes determine many of the fundamental choices for regulatory programs by specifying what factors may be considered in adopting regulatory standards and the objectives of the standards in terms of the level of safety to be attained. Statutes differ greatly on these questions, with some mandating standards be based on technological feasibility, whereas others specifying health-based regulation. The statutes can provide additional guidance that can influence the agency’s consideration of toxicological evidence, such as the 1996 amendments to the Safe Drinking Water Act, which direct EPA to use “the best available, peer-reviewed science and supporting studies conducted in accordance with sound and objective scientific practices”(SDWA, 1996).

Congress also controls regulatory programs through its budgetary authority. By directing how much money is directed to specific activities or programs, Congress can directly influence agency priorities and emphasis. If Congress is not pleased with an agency’s actions, it can adopt appropriation riders prohibiting the agency from spending money on controversial topics. In the past, for example, Congress has prevented EPA from working on climate change or radiofrequency issues by imposing annual riders for a time period.

In addition to these primary legislative controls, Congress has a number of other ways it can influence agency decision making. Congress can pressure agencies toward a preferred policy outcome through oversight hearings, letters to agency officials, and press releases. Congress has also adopted some oversight statutes to further control agency rulemaking. The Congressional Review Act requires federal regulatory agencies to file final rules with Congress before the rules can become effective. If the regulation meets the definition of a major rule ($100-million impact on the economy), Congress has 60 days to introduce a resolution of disapproval that, if adopted by both the House and Senate and signed by the President, can nullify the agency’s rule.

Congress also enacted section 515 of the Treasury and General Government Appropriations Act of 2001, known as the Information Quality Act (IQA), which required the OMB to promulgate guidance to agencies ensuring the quality, objectivity, utility, and integrity of information (including statistical information) disseminated by federal agencies (IQA, 2001). Federal agencies were also required by the IQA to publish their own agency-specific guidelines no later than one year after OMB’s guidelines, and also provide a mechanism for stakeholders to file petitions challenging agency actions that allegedly fail to comply with the data quality objectives. Private parties have challenged a number of scientific determinations by federal agencies under this provision, but the courts have determined that the agency decisions on these petitions are not judicially reviewable (Salt Institute v Leavitt, 2006).

Executive Oversight

The Executive branch also exerts overview of regulatory decisions, particularly through the OIRA within the White House OMB. OIRA reviews all draft significant regulations proposed by regulatory agencies, and although its review most often focuses on the economic aspects of the proposed rule, it also reviews the agency’s scientific assumptions and findings. Although OIRA cannot itself revise the proposed regulation, it can hold up its publication until it has reached agreement with the agency on acceptable revisions.

The Executive Branch has other mechanisms for oversight and influence of regulatory science determinations by agencies. The White House under various Presidents has put out a series of Executive Orders providing regulatory guidance that agencies must follow to the extent consistent with the governing regulatory statute. For example, Executive Order 13563 issued by President Obama in January 2011 requires that federal regulations “must be based on the best available science. It must allow for public participation and an open exchange of ideas. It must promote predictability and reduce uncertainty. It must identify and use the best, most innovative, and least burdensome tools for achieving regulatory ends. It must take into account benefits and costs, both quantitative and qualitative” (EO 13563, 2011).

The White House also issued a memorandum on Scientific Integrity in 2009 that required each regulatory agency to develop its own scientific integrity policy that implemented several principles of scientific integrity, including: “(a) The selection and retention of candidates for science and technology positions in the executive branch should be based on the candidate’s knowledge, credentials, experience, and integrity; (b) Each agency should have appropriate rules and procedures to ensure the integrity of the scientific process within the agency; (c) When scientific or technological information is considered in policy decisions, the information should be subject to well-established scientific processes, including peer review where appropriate, and each agency should appropriately and accurately reflect that information in complying with and applying relevant statutory standards.…” (White House, 2009).

This chapter has focused thus far on how toxicological data are used in federal regulatory programs. Yet, the relationship between regulation and toxicology is reciprocal, in that regulatory programs have also shaped and influenced the practice of toxicology. Modern toxicology has developed, in substantial part, in response to the information needs of contemporary regulation. But government regulation impinges on the discipline of toxicology in more direct ways as well. Regulatory agencies often prescribe the specific objectives and design of studies that are conducted to satisfy regulatory requirements. In addition, pressure to protect animals used in research has produced laws and regulations that govern toxicologists themselves.

Testing Guidelines

An agency’s influence over the conduct of toxicology studies depends on its regulatory responsibilities. An agency such as the FDA or the EPA, which must confirm the safety of new substances before marketing, can dictate the kinds of tests that manufacturers must conduct to gain approval. By contrast, an agency or program that has no premarket approval function has less leverage.

Agencies such as the FDA and EPA that have been authorized to require premarket approval for certain categories of products have generally specified the types of tests they require before they will consider the safety of a compound (eg, tests for acute toxicity, subchronic effects, and chronic effects). Within each of these categories, an agency might set out more detailed requirements, essentially enumerating its “base set” data demands. In addition, the agencies usually adopt guidelines that describe preferred or acceptable methods for executing particular tests, such as a bioassay for carcinogenesis (EPA, 2011a; FDA, 2000). In addition, the EPA has established guidelines for several of the tests that it may mandate by rule or consent agreement for individual chemicals under TSCA § 4 (EPA, 2011a). Multinational bodies such as the Organization for Economic Cooperation and Development have sought to secure multilateral adherence to standardized test guidelines and minimum testing requirements for new chemicals (Page, 1982).

Both the FDA and the EPA have adopted another set of requirements that specify laboratory procedures for conducting tests required or submitted for regulatory consideration. These good laboratory practice (GLP) regulations prescribe essential but often mundane features of sound laboratory science, such as animal husbandry standards and record-keeping practices (EPA, 2011b, 2011c; FDA, 20011a). All of these requirements are intended to contribute to sound-regulatory decision making by ensuring the quality and integrity of toxicology data submitted to support agency decisions.

Food and Drug Administration

The FDA exercises premarketing approval authority over several classes of compounds, of which the most important, for present purposes, are new human drugs and direct additives to food.

Toxicology Testing Requirements for Human Drugs

In 1962, Congress expressly authorized the FDA to exempt investigational drugs from the premarket approval requirement so that they could be shipped for use in clinical testing, subject to conditions the agency believed appropriate to protect human subjects [FFD&C Act § 505(i)]. One condition that the FDA established was that an investigational drug first must have been evaluated in preclinical studies, which are then submitted to FDA as an IND application. This requirement appears in current regulations that amplify, in the text and in referenced guidelines, the types of tests that are to be performed and the design they should follow (FDA, 2011b). A drug’s sponsor is encouraged to consult agency personnel to get a precise understanding of what sorts of toxicology studies the agency expects prior to submitting the IND. Preclinical studies of substances that are candidates for use as human drugs must also meet the standards set by the FDA’s GLP regulations (FDA, 2011a). These regulations apply to all laboratories—university, independent, and manufacturer-owned—in which such studies are conducted. The work of the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) highlights the trend toward international agreement on test methods. Since its founding in 1990, the ICH—comprising the pharmaceutical regulatory agencies and industries from the European Union, Japan, and the United States—has issued over 50 harmonized guidelines on various aspects of pharmaceutical regulatory approval, including numerous guidelines on toxicology testing methods for human drugs.

Testing Requirements for Food Additives

The Food Additives Amendment and the Color Additive Amendments require premarket approval of new additives to human food. Both laws assume that laboratory studies in animals will provide the principal data for assessing safety. Thus, a petitioner must submit “full reports of investigations made with respect to the safety for use of such additives, including full information as to the methods and control used” [FFD&C Act § 409(c)]. The FDA has adopted regulations specifying the types of data that an applicant must submit (FDA, 2011c).

The FDA’s regulations contain only general statements about the need for and features of toxicology studies. For many years, the agency maintained an advice-giving system in which it prescribed the type and design of tests to be performed. In 1982, the FDA first codified this “common law” in Toxicological Principles for the Safety Assessment of Direct Food Additives and Color Additives Used in Food, known thereafter as the “Red Book.” The Red Book, revised in 2000, describes the types of tests the FDA believes necessary to evaluate an additive’s safety (FDA, 2000). The agency’s requirements, which are in the form of guidelines rather than regulations, are calibrated to the purposes for which the additive will be used, to estimated levels of human exposure, and to the results of sequential studies. Tests of food color and additives must also comply with the FDA’s GLP regulations (FDA, 2011a).

Environmental Protection Agency

The EPA’s premarket approval authority over pesticides places it, like the FDA, in a position to dictate the design and conduct of studies on such compounds. The 1976 TSCA gave the EPA authority to mandate testing of other chemicals in use or scheduled for introduction and to specify, by regulation, test standards.

Toxicology Requirements for Pesticides

FIFRA requires the submission of toxicology studies, as well as other types of investigations, to support the EPA’s evaluation of a pesticide [FIFRA § 136(b)]. The statute also requires the EPA to publish guidelines specifying the kinds of information that will be required to support the registration.

The agency has issued regulations outlining the procedures for submission of registration petitions and their basic content (40 CFR part 158 and 161). The Office of Chemical Safety and Pollution Prevention (formerly the Office of Prevention, Pesticides, and Toxic Substances) has issued Harmonized Test Guidelines for all toxicology testing done under both FIFRA (pesticides) and TSCA (chemicals) (EPA, 2011a). For example, the Health Effects Test Guidelines (Series 870) include guidelines for approximately 50 different specific assays for acute toxicity, subchronic toxicity, chronic toxicity, genetic toxicity, neurotoxicity, and other toxicity testing. EPA has attempted to harmonize these guidelines whenever possible with applicable Organization for Economic Cooperation and Development (OECD) guidelines. Animal studies of pesticides must also comply with EPA’s own good laboratory practice regulations, which were inspired by the same investigations that led the FDA to promulgate its standards for testing laboratories and impose similar requirements.

Testing of Industrial Chemicals

The primary means by which the EPA may mandate health effects testing of new or existing industrial chemicals is Section 4(a) of the TSCA. That provision states that the administrator “shall by rule require that testing be conducted to develop data with respect to the health and environmental effects for which there is an insufficiency of data and experience” to permit assessment of whether a substance presents an unreasonable risk. This obligation to order testing is triggered by an administrative finding that a chemical presents a potential risk (based on the suspicion of toxicity) or that humans or the environment will be exposed to substantial quantities. The statute creates an Interagency Testing Committee (ITC) with members from EPA, CPSC, FDA, Department of the Interior (DOI), OSHA, Council on Environmental Quality (CEQ), National Institute for Occupational Safety and Health (NIOSH), National Institute of Environmental Health Sciences (NIEHS), National Cancer Institute (NCI), National Science Foundation (NSF), Agency for Toxic Substances and Disease Registry (ATSDR), National Library of Medicine (NLM), USDA, and the Department of Commerce to recommend a list of chemicals that should be tested and in what order of priority. Once the ITC has recommended a chemical substance for testing, the EPA is required by statute to either initiate testing or publish its reasons for not doing so within 12 months, although given the administrative burdens of such action the agency has frequently failed to meet this deadline.

This last requirement to take action on ITC recommendations within 12 months, coupled with the statute’s formal procedures for adopting test rules, led the EPA initially to rely on negotiations with chemical producers to secure voluntary agreements for the conduct of tests it thought appropriate for chemicals identified by the ITC (GAO, 1982). The practice was challenged by public interest organizations, which were excluded from the negotiations, and ultimately it was declared unlawful (NRDC v EPA, 1984). The agency amended its regulations to recognize two forms of mandates for testing: test rules and enforceable testing consent agreements.

TSCA test rules are subject to judicial challenge. Manufacturers challenged a 1988 test rule for cumene, arguing that the EPA had failed to support its finding that the substance enters the environment in substantial quantities with the potential for substantial human exposure. Although the court ultimately upheld the rule, it ordered the agency to articulate standards governing the definition of “substantial” (Chemical Manufacturers Association v EPA, 1990). Another court challenge to a test rule requiring neurotoxicity studies of 10 widely used and intensively marketed organic solvents resulted in a settlement with the EPA. The settlement required the EPA to enter into consent agreements with reduced testing requirements for seven chemicals, to eliminate testing requirements for two, and to postpone its decision on one. Both test rules and testing consent agreements specify what types of tests are to be done. Their design is governed by regulations promulgated at 40 CFR part 799, and can usually be satisfied by the EPA Harmonized Test Guidelines (EPA, 2011a). Compound-specific testing requirements are sometimes also written into individual TSCA test rules or consent agreements. All toxicology studies required by the EPA under the TSCA must also comply with its GLP regulations.

Validation of Alternative Test Methods

Regulatory agencies have also been instrumental in encouraging the development of new safety test methods. Federal agency validation of new toxicological test methods was formalized and harmonized by the Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM) Authorization Act of 2000, which permanently established the ICCVAM. Some 15 federal agencies participate in ICCVAM. Congress directed that one of the purposes of ICCVAM was to “ensure that new and revised test methods are validated to meet the needs of Federal agencies” [Pub L 106-545, §3(b)(4)]. In that regard, ICCVAM is directed to review and evaluate new or revised or alternative test methods and to provide test recommendations to federal agencies. Each federal agency administering a program that “requires or recommends acute or chronic toxicological testing” is required to adopt an alternative test method recommended by ICCVAM unless the agency determines that the alternative test method would not be consistent with its statutory goals.

The ICCVAM statute also requires that each federal agency “that requires or recommends acute or chronic toxicological testing shall ensure that any new or revised acute or chronic toxicity test method, including animal test methods and alternatives, is determined to be valid for its proposed use prior to requiring, recommending, or encouraging the application of such test method” [Pub L 106-545, §4(c)]. ICCVAM has established detailed criteria for the validation and regulatory acceptance of toxicological test methods (NIEHS, 1997). ICCVAM defines validation as “the process by which the reliability and relevance of a test method are evaluated for the purpose of supporting a specific use.” The ICCVAM criteria stipulate that test methods that have been validated according to the ICCVAM criteria are not automatically acceptable for regulatory agencies. The ICCVAM criteria thus specify a second set of criteria for regulatory acceptance, which address the existence of standard operating procedures for the test method, its cost-effectiveness, its utility for agency risk assessment, its potential for harmonization with similar testing requirements by other entities, and its potential for reducing the use of animals. The ICCVAM statute preserves the individual agency’s final authority to accept or reject a specific test method based on its statutory mandate and regulatory mission.

The ICCVAM process has now been used to validate a significant number of alternative test methods, including for acute oral toxicity, dermal irritation, ocular irritation, and allergic contact dermatitis. The ICCVAM Authorization Act also directs federal agencies to reduce, refine, and/or replace the use of animals in testing where feasible.

Computational Toxicology

Another area where regulatory demands and programs are pushing the development of new toxicology tests is the field of computational toxicology. Computational toxicology uses high-powered computers to discern patterns and findings from large data sets derived from quantitative structure–activity relationships, in vitro assays, and various ‘omic technologies (eg, toxicogenomics, proteomics, metabolomics). The regulatory driver for computational toxicology is the need to evaluate much larger numbers of chemicals much quicker and at lower costs, and perhaps eventually even with more robust and relevant results, than traditional animal studies. The shift toward computational toxicology was boosted by a 2007 Natural Research Council report entitled Toxicity Testing in the 21st Century: A Vision and a Strategy, which endorsed greater reliance on computational methods (NRC, 2007). EPA has launched an ambitious program to implement computational methods into its regulatory programs, which is called the ToxCast program, and EPA and the National Institutes of Health (NIH) signed a memorandum of agreement in 2008 to work together to develop these new toxicological methods for regulatory applications (Collins et al., 2008).

Animal Welfare Requirements

Researchers who conduct studies funded by federal agencies must comply with the Animal Welfare Act (AWA), and some are also subject to restrictions imposed by the Public Health Service (PHS). Recipients of grants from the Department of Education, the Department of Agriculture, or the EPA are subject only to the AWA. Those funded by the Department of Energy or by the PHS (which includes all research agencies within the Department of Health and Human Services, including the NIH) must also comply with PHS policies. Restrictions on animal use also appear in the GLP regulations adopted by the FDA and the EPA (Reagan, 1986).

Animal Welfare Act

The AWA was initially adopted in 1966 and has been subsequently amended several times (1970, 1976, 1985, 1990, 2002, 2007, 2008). It is administered by the Animal and Plant Health Inspection Service (APHIS), a part of the USDA. The AWA, which protects only warm-blooded animals and excludes birds, rats, and mice bred for research, requires all covered research facilities to register with APHIS and agree to comply with applicable AWA standards for humane treatment of animals. Each facility must file an annual report signed by a responsible official that shows that “professionally acceptable standards governing the care, treatment, and use of animals” were followed for the year in question (AWA, 1998).

Pursuant to the AWA, APHIS has established specific requirements for the humane handling, care, and transportation of dogs and cats, guinea pigs and hamsters, rabbits, nonhuman primates, marine mammals, and other warm-blooded animals. The regulations governing facilities address living space, heating, lighting, ventilation, and drainage. The health and husbandry provisions address feeding, watering, sanitation, veterinary care, grouping of animals, and the number and qualifications of caretakers (APHIS, 2011).

The AWA requires each research facility to establish an Institutional Animal Care and Use Committee (IACUC), composed of three or more members, one of whom must be a veterinarian and one of whom must represent community interests and who may not be affiliated with the institution. In 1989, the APHIS expanded the responsibilities of the IACUC and increased the minimum size to five members, and must include a veterinarian, an animal research scientist, a nonscientist, and an individual not affiliated with the institution. At least one member of the IACUC must now review and approve the animal care and use components of all proposed research activities. Prerequisites to approval include the avoidance or minimization of discomfort, distress, and pain; the use of pain-relieving drugs where appropriate; the consideration of pain-free alternatives; and euthanization when an animal would otherwise experience severe or chronic pain or distress that cannot be relieved.

The IACUC is also responsible for conducting semiannual inspections of the facility itself and of the program for humane care and use of animals. Committee reports are filed with the APHIS and with any federal agency funding the research.

Public Health Service Policy

The PHS Policy on Humane Care and Use of Laboratory Animals applies to research funded by the NIH and other federal agencies that uses any vertebrate animals, and thus has a broader reach than the AWA. The PHS policy requires each facility to submit an annual report, called an “Assurance,” which is evaluated by the NIH Office of Laboratory Animal Welfare (OLAW) to determine the sufficiency of animal care.

The PHS policy imposes two primary obligations on researchers as a condition for receiving federal funding: each institution must adopt an Institutional Program for Animal Care and Use, and it must establish an IACUC. The IACUC must review all applications for research funding and review the institution’s programs to ensure compliance with NIH standards. The Health Research Extension Act of 1985 requires that PHS-funded institutions provide training on methods to reduce animal suffering similar to that mandated by the AWA for its personnel. It also requires that researchers’ grant applications justify any proposed use of animals (NRC, 1988).

Research facilities subject to either the AWA or PHS may wish to consult a National Academy of Sciences’ report that details suggestions for developing institutional compliance programs (NRC, 1991). Scientists working with no federal funding who expect their research to be submitted to the FDA or the EPA are not subject to the AWA or PHS policies, but they must comply with the animal protection provisions of those agencies’ GLP regulations. These regulations prescribe adequate living conditions, detail requirements for veterinary treatment, and impose specific record-keeping requirements (EPA, 2011b, 2011c; FDA, 2011a).

Cases

Secondary Sources

Statutory and Regulatory Materials