Existing Frameworks or Systems for Evaluating the Safety of Other Substances
The following text provides an overview of existing safety frameworks that have been developed by federal agencies, industries, or other organizations. The descriptions are based on presentations to the committee and information provided by the organizations themselves; the descriptions are not the committee evaluations of the frameworks.
PREMARKET APPROVAL OF FOOD ADDITIVES
“Food additives” include an array of substances that accomplish a variety of technical effects in food. Included are direct food additives (e.g., artificial sweeteners), food-processing aides (e.g., antimicrobials), food contact substances (e.g., food packaging), and, by legal definition, sources of food irradiation. Under the 1958 Food Additives Amendment (FAA, P.L. 85-929) to the Food, Drug, and Cosmetic Act (FDCA), the Food and Drug Administration (FDA) has responsibility for the premarket approval of food additives.1 The statute, as interpreted by FDA, establishes both the standard of data review (i.e., fair evaluation of the data of record), as well as the standard of safety (i.e., a reasonable certainty of no harm under the intended conditions of use). Notably, the statute exempts from premarket safety evaluation the use of substances in food that are “generally recog-
nized as safe” (GRAS) by qualified experts in light of scientific procedures or, for substances used prior to 1958, in light of scientific procedures or experience based on common use in food. As discussed in the next section, FDA has in place a process for assessing the worthiness of claims of GRAS status.
FDA conducts safety assessments of new food additives under the principle of establishing a reasonable certainty of no harm by applying a decision framework. This framework uses a risk-assessment approach that includes the compilation of available data and information and the application of toxicological and other types of decision elements.
FDA then reviews the available toxicology studies. It has developed guidelines (Toxicological Principles for the Safety of Food Ingredients, commonly known as The Redbook [OFAS, 2003]) for interested parties to use when assembling the required data for submission of a food additive petition. The Redbook outlines the types of toxicological testing FDA normally expects to be provided in support of the food additive’s safety, based initially on the additive’s chemical structure and probable human exposure. (The Redbook provides guidance; it does not include requirements.) Using the information submitted, FDA assigns additives to initial “concern levels” (or “minimum testing levels”) of I, II, or III. In its data review, FDA applies toxicological decision elements to further refine the scope of needed toxicological data.
From the animal studies, FDA determines the highest level of intake associated with no adverse toxicological effect in the most sensitive, longest duration, most relevant animal study. This “highest no-effect level” is then divided by an “uncertainty factor” (or “safety factor”), often a factor of 100, to account for both intra- and interspecies variability. The resulting value is the acceptable daily intake (ADI) for the additive. The ADI is compared to the estimated daily intake to determine whether the proposed use of the additive is consistent with a reasonable certainty of no harm.
For some substances, the traditional risk assessment approach is not applicable. For example, an additive may be so toxicologically inactive that not enough of the additive can be orally ingested by the test animals to elicit a toxic response without perturbing normal nutrition. In such cases it is difficult to determine an ADI. FDA may then employ other types of decision elements. In these cases, increased emphasis may be placed on, for example, chemical identity information and structure-activity relationships; data on absorption, distribution, metabolism, and excretion; and human tolerance studies (to look at physiological and nutritional responses).
Once all the information has been evaluated, FDA concludes whether the proposed use of a food additive is consistent with a reasonable certainty of no harm and can be safely marketed. After a new food additive is on the market, FDA may monitor the substance for safety through examination of
available clinical studies and postmarket surveillance (Personal communication, A. Rulis, FDA, January 25, 2002).
SELECT COMMITTEE ON GRAS SUBSTANCES
Based on the 1958 amendment for food additives to the FDCA, FDA developed specific processes to determine whether substances used in foods were safe for their intended use (see previous section). Food additives, as defined in the amendment, are subject to premarket approval by FDA unless they are GRAS or fall within another statutory exception (21 U.S.C. § 321(s)).
Following the passage of the FAA, FDA assumed a lenient approach to dealing with the GRAS exception. In the early 1970s, however, in response to public concern about the apparent carcinogenicity of cyclamate, which FDA had listed as GRAS, FDA adopted a more rigorous approach (Degnan, 2000). In 1972 FDA contracted with the Life Sciences Research Office (LSRO) of the Federation of American Societies for Experimental Biology for assistance with a comprehensive review of GRAS substances. LSRO established a Select Committee on GRAS Substances that examined monographs that provided all known data on physical and chemical properties of a substance, human exposure data, animal and human toxicity data, and reports of special studies on mutagenicity, carcinogenicity, and teratogenicity of each substance. The Select Committee reached one of five conclusions on each GRAS substance reviewed: (1) continue as GRAS, (2) continue as GRAS with limitations, (3) uncertainties exist—issue interim food additive order requiring further testing, but continue as GRAS until tests are evaluated, (4) evidence of adverse effects—establish conditions of safe use or remove GRAS status, or (5) inadequate data precludes evaluation—rescind GRAS status. The review was completed in 1982.
In 1972 FDA also established procedures for petitioning to affirm the GRAS status of a substance, which while still in use have rarely been used due to the significant amount of documentation required. Currently, in order to achieve GRAS status for a substance not used in food prior to 1958, four key criteria must be met: (1) general recognition of safety by qualified experts, (2) the experts must have the scientific training and experience necessary to evaluate the safety of the substance, (3) the experts must base their safety determination on scientific procedures, and (4) GRAS determination must fully consider the intended use of the substance (Hallagan and Hall, 1995).
In 1997 FDA proposed to replace the current GRAS affirmation petition scheme with one of notification, allowing any organization to notify FDA of a self-determination of GRAS. The interim policy allows this method to be used pending issuance of final regulations. The proposed rule would
also clarify the types of evidence needed to establish GRAS status (Degnan, 2000).
GRAS DETERMINATION FOR FLAVOR INGREDIENTS: FLAVOR AND EXTRACT MANUFACTURERS ASSOCIATION EXPERT PANEL
Because flavor ingredients are a type of food additive, the flavor industry has to adhere to the requirements laid out in the 1958 FAA. To determine GRAS status for flavoring substances, the Flavor and Extract Manufacturers Association (FEMA), the trade organization of the flavor ingredients industry, created its own independent expert panel. The FEMA Expert Panel, which has been reviewing flavoring substances since soon after the passage of the FAA in 1958, includes qualified experts in toxicology, pharmacology, biostatistics, pharmacokinetics, biochemistry, pathology, nutrition, organic chemistry, medicinal chemistry, and metabolism (Woods and Doull, 1991). The panel evaluates the available data on safety and use of flavoring ingredients and assesses whether the ingredients meet the criteria for GRAS status.
The FEMA Expert Panel has developed a safety assessment evaluation process for determining GRAS status. Once an application for GRAS status is submitted to the panel with a complete literature search, the first step is preliminary assessment of the data for adequacy by FEMA staff. These data are then evaluated by the panel using the following criteria: (1) exposure to the substance in specific foods, (2) natural occurrence in foods, (3) chemical identity and chemical structure, (4) metabolic and pharmacokinetic characteristics, and (5) animal toxicity (Woods and Doull, 1991). The panel examines toxicity and metabolic data on structurally similar compounds and considers the history of use of the substance (Hallagan and Hall, 1995).
Based on the weight of the evidence and expert judgment, the panel reaches one of three conclusions: (1) GRAS, (2) not GRAS, or (3) insufficient data to determine GRAS status. If data are insufficient, the panel will reexamine the substance after more data are available. The designation of GRAS status on a flavor ingredient must be based on a unanimous decision by the panel.
COSMETICS INGREDIENT REVIEW
As is the case for dietary supplements, there is no premarket regulatory system for cosmetic ingredients other than color additives that are regulated directly by FDA. The Cosmetics Ingredient Review (CIR) Program was established in 1976 by the Cosmetic, Toiletry, and Fragrance Association
(CTFA) to review and assess the safety of cosmetic ingredients in the marketplace.
The CIR Program is funded by industry, but its review process is independent and open to public and scientific scrutiny (Bergfeld and Andersen, 2000). Approximately 2,800 cosmetic ingredients were on the market in 1976 when the CIR Program was established. In response, the CIR Program developed a system to prioritize these ingredients before performing the safety review. First it excluded or deferred ingredients being reviewed by other groups, such as fragrances and ingredients being evaluated by FDA, including color additives and over-the-counter (OTC) drug ingredients. The CIR Program then grouped the remaining ingredients into chemically related families and prioritized based on the following factors: frequency of use, ingredient concentration in cosmetic products, area of human exposure, number of products containing the ingredient used by sensitive population subgroups (such as infants and the elderly), biological activity, frequency of consumer complaints, and skin penetration. Using a ranking methodology, ingredients were given a weighted score based on these factors and were then reviewed in priority order. Frequency of use and biological activity were given the most weight in the ranking. This priority listing and ranking methodology is updated periodically.
The safety review starts with a comprehensive literature search by CIR staff. The staff summarizes the available published data and publishes the summaries for public comment. During a 90-day period, interested parties may submit comments or additional data.
Following this comment period, a CIR Expert Panel begins its review of the collected data and determines whether more data are needed. The panel consists of seven scientists and physicians who serve as voting members and three nonvoting liaison members, representing CTFA, FDA, and the Consumer Federation of America. CIR emphasizes that voting members are careful to avoid any perceived or real conflicts of interest. Liaison members serve to keep consumer groups, FDA, and industry informed of the panel’s deliberations.
If additional data are required, an informal request is directed toward the cosmetic industry. If data are not forthcoming or are still inadequate for the safety assessment, a formal request is made. Once all the necessary data are received, the panel reviews them and produces a tentative report that is released for public comment. At the end of the comment period, comments are considered and the final report is written.
In determining safety for the final report, the panel looks at all the available data, considers structurally similar substances, and relies on panel members’ experience and expertise. The data needed for the safety assessment are dependent on the particular ingredient under review. However, the panel usually considers chemical and physical properties, impurities,
extent and type of use, concentration of use, subchronic or chronic toxicity, skin penetration, skin irritation, and skin sensitization.
In each final report, the CIR Expert Panel reaches one of four conclusions on the safety of a cosmetic ingredient: (1) safe as currently used, (2) safe with qualifications, (3) unsafe, or (4) insufficient data. If data are considered insufficient, the panel notes what data are lacking. In practice, this conclusion of insufficient data encourages manufacturers to undertake additional studies.
Unlike dietary supplements, premarket approval of new drugs places the burden of proof regarding safety on the manufacturer rather than on FDA. The evaluation of new drugs, new uses for approved drugs, and classification of OTC drugs is an intensive interactive process that evaluates both safety and efficacy. Manufacturers that want to develop and market a new drug must follow the FDA approval process that is modeled on a risk-benefit approach. Approval of a new drug requires extensive studies of the chemistry, manufacturing, and controls of the drug, toxicology and pharmacology of the compound in animals, and clinical trials of effectiveness and safety in humans. The timeframe and resources for this process are extensive (21 C.F.R. § 300 ).
A key initial step in the drug approval process is submission by the manufacturer of an Investigational New Drug (IND) application to FDA. The IND is a large collection of information that enables FDA to review the safety of the substance before clinical testing in humans is allowed to begin. The IND describes the ingredients, synthesis, manufacturing, purity, and microbiology of the drug product, as well as the stability, packaging, and labeling. Also included in the IND are data from rodent and nonrodent animal studies, such as pharmacokinetic and pharmacodynamic data from animal studies, genotoxicity studies, carcinogenicity studies, reproductive and teratogenic studies, and other toxicological data. When available, the application also includes published or unpublished human data. Because these data help FDA determine whether the human testing process will be allowed to proceed, the manufacturer also provides protocols outlining the Phase I, II, and III clinical studies it plans to conduct. After the IND is submitted, FDA has 30 days to review its content. If FDA does not contact the sponsor within that time, the proposed Phase I study may begin (21 C.F.R. § 312 ).
During Phase I studies, which focus on safety but not efficacy, human volunteers (who are usually healthy) are carefully monitored for tolerability, and pharmacokinetic data are often collected. The aim of Phase II is to evaluate the dose-response relationship and effectiveness of the drug in a
few hundred subjects who have the disorder the drug is intended to treat. These studies are usually double-blind and placebo-controlled to minimize investigator and subject bias. Phase III of the investigation consists of well-controlled trials to gather evidence on both effectiveness and safety of the drug and information needed for labeling. These are large trials of several hundred to several thousand subjects.
The data collected in all of the clinical studies enable FDA to approve or disapprove a drug based on a risk-benefit analysis. Once a drug is approved and marketed, additional safety information continues to be collected through mandatory submission of adverse event information from the manufacturer to FDA via MedWatch and other reporting mechanisms. FDA may also require the manufacturer to conduct postmarketing studies.
The process above describes the steps required for a new drug approval (NDA). In the years after proof of effectiveness was added to the NDA requirements, FDA wrestled with how to deal with the thousands of OTC drugs that were on the market though not covered by approved NDAs. Rather than make case-by-case challenges to such products, FDA decided to review them by therapeutic class with the assistance of expert advisory committees. The process that FDA established to accomplish this mission is known as the OTC Drug Review.
In 1972 FDA, with the help of 17 advisory panels, began its review of the more than 700 active ingredients with almost 1,500 uses in marketed OTC drug products. The aim of the review was to prepare monographs establishing the conditions under which OTC drugs would be considered generally recognized as safe and effective, and not misbranded, and thus exempt from the NDA process.
The OTC Drug Review consists of several phases. In the first phase, now complete, the advisory panels made recommendations regarding the categorization of products. Category I was for those drugs that the panel deemed to be generally recognized as safe and effective, and not misbranded if they satisfied specified conditions, including, among others, active ingredients and labeling indications. Category II was for products with active ingredients, labeling claims, or other conditions that resulted in them not being generally recognized as safe or effective or resulted in them being misbranded. Category III was for products with active ingredients, labeling claims, or other conditions for which the data were insufficient and for which further testing was thus required.
In the second phase of the review, FDA published the panels’ recommendations as Advanced Notices of Proposed Rulemaking (ANPRs). These ANPRs included proposed monographs establishing the conditions under
which OTC drugs in specific therapeutic classes would be generally recognized as safe and effective, and not misbranded (Category I). In the third phase, after considering the public comments received in response to ANPRs, the agency issued proposed rules designated Tentative Final Monographs. In the final step of the process, the agency, after receiving further comments, publishes final monographs. As of March 1, 2001, most, but not all, of these final monographs had been published (CDER, 2001). Final monographs set forth the mandatory conditions for an OTC drug to be considered generally recognized as safe and effective, and not misbranded, including active ingredients, dosages, permitted combinations of ingredients, warnings, and labeling requirements.
NEW CHEMICALS PROGRAM
Under the New Chemicals Program, the Environmental Protection Agency (EPA) is given the authority to regulate the entry and use of new chemicals into the U.S. marketplace. This program, mandated by Section 5 of the Toxic Substances Control Act (TSCA) in 1976, seeks to manage the potential risk from new chemicals both to humans and to the environment. Manufacturers or importers of new chemicals are required under TSCA to notify EPA through a premanufacturer notice (PMN) that must be submitted at least 90 days prior to manufacture or import of the new chemical. New chemicals are defined as those that are not listed on EPA’s TSCA Chemical Substance Inventory of existing chemicals. The burden of proof for identifying risk rests with EPA.
EPA receives petitions for approximately 2,000 new chemicals from manufacturers each year (Personal communication, L. Scarano, EPA, October 11, 2001). At submission, the manufacturer provides the PMN, which includes information on chemical and physical identity and properties, product uses, proposed production or importation volume, by-products, human exposure, disposal practices, environmental releases, pollution prevention efforts, and available information on health or environmental effects. A multidisciplinary team of experts is responsible for reviewing the safety information provided in the PMN. The first step is to determine whether the substance is already on the TSCA inventory. If it is not, the team then evaluates chemical structure, how the chemical is synthesized, the intended use of the chemical, and the physical and chemical properties of the chemical. They also check for analogs in an EPA analog database. About 30 percent of the applications are not reviewed after this stage; these substances consist of polymers, which because of their molecular weight and other properties are considered unlikely to present significant hazard potential.
The next step of the process is to estimate the potential environmental
and health hazards using analog analysis, quantitative structure activity relationship models, and expert judgment. The structure activity team has identified 54 structural alert categories that may indicate a potential concern for chemicals that fall into these categories (Personal communication, L. Scarano, EPA, October 11, 2001).
The third step is to prioritize the results of the safety evaluations and to decide if further review is warranted. If it is required, the next step is a more detailed standard review. In this step, a risk assessment is conducted, human health hazard information is evaluated, and the chemical is assigned a qualitative determination of the hazard concern level. Evidence of adverse effects in human populations and conclusive evidence of severe effects in animal studies constitute a high hazard concern level. A moderate level of concern results from suggestive animal studies and analog data and knowledge that the chemical class has produced toxicity. The low concern level is for those chemicals for which no concern was identified. At this point, depending on the hazard concern level and considering the estimated exposures and releases, EPA will inform the manufacturer that the chemical presents potential risk issues and that more testing is needed. If EPA does not act to regulate the chemical, the manufacturer may commence production or importation.
Bergfeld WF, Andersen FA. 2000. The cosmetic ingredient review. In: Estrin NF, Akerson JM, eds. Cosmetic Regulation in a Competitive Environment. New York: Marcel Dekker. Pp. 195–216.
CDER (Center for Drug Evaluation and Research). 2001. Milestone Status of OTC Drug Review Documents as of March 1, 2001. Online. Food and Drug Administration. Available at http://www.fda.gov/cder/otc/milestone/pdf. Accessed March 28, 2002.
Degnan FH. 2000. FDA’s Creative Application of the Law. Washington, DC: Food, Drug, and Law Institute. Pp. 19–26.
Hallagan JB, Hall RL. 1995. FEMA GRAS—A GRAS assessment program for flavor ingredients. Flavor and Extract Manufacturers Association. Regul Toxicol Pharmacol 21:422–430.
OFAS (Office of Food Additive Safety). 2003. Redbook 2000. Toxicological Principles for the Safety of Food Ingredients. Online. Available at http://www.cfsan.fda.gov/~redbook/red-toca.html. Accessed March 28, 2002.
Woods LA, Doull J. 1991. GRAS evaluation of flavoring substances by the Expert Panel of FEMA. Regul Toxicol Pharmacol 14:48–58.