Legal Aspects of the Food Additive Approval Process
[background paper commissioned by the Institute of Medicine of the National Academy of Sciences for a Food Forum workshop entitled Enhancing the Regulatory Decision-Making Approval Process for Direct Food Ingredient Technologies (1997)]
In the past few years, novel food substances have attracted significant public attention, most notably Procter & Gamble's fat substitute olestra and Calgene's bioengineered tomato. Some have criticized the U.S. Food and Drug Administration (FDA) for its lengthy delays in reviewing these and other substances added to food.1 At the same time, others have argued that the agency does not adequately ensure the safety of such substances. These latest controversies pose significant and timely questions about how best to regulate substances added to food. The problems are multifaceted, and policymakers will require input from both legal and scientific perspectives to address them.
This paper focuses primarily on a description of the FDA's regulation of substances intentionally added to food as it has evolved over the last several decades, and it concludes with a discussion of several possible avenues for reform. Debates about proposals to modify existing procedures must start with a proper appreciation of the difficulties encountered in the past. Moreover, the history of the food additive approval process illuminates significant and recurring challenges faced in the design of properly functioning regulatory programs. Studies of other federal agencies have uncovered valuable lessons about effective regulation that transcend the particular program under consideration.2
The federal government first asserted authority over the quality and safety of food products early this century. In 1906, reacting to widely publicized examples of industry abuses, Congress prohibited the introduction of adulterated or misbranded food and drugs into interstate commerce.3 The act provided that any food containing an ''added poisonous or other added deleterious ingredient which may render such article injurious to health'' would be deemed adulterated. 4 If they detected a safety problem, federal officials could initiate enforcement action to remove the product from the market, but the government would shoulder the burden of proving that the ingredient posed "a reasonable possibility of injury."5 Although Congress replaced this original statute in 1938 with the federal Food, Drug, and Cosmetic (FD&C) Act,6 the newer legislation retained the same basic system of after-the-fact policing for adulterants in food until the Food Additives Amendment of 19587 created a premarket review and approval system. These two enactments are discussed more fully below.
In the two decades that elapsed between the passage of the FD&C Act and the Food Additives Amendment, a number of developments rendered the original statutory design outdated. Indeed, the original legislation, based as it was on the 1906 Act, focused on the control of "adulterants" and did not fully anticipate the rapid progress in food processing technology and the growing utilization of intentional additives that would follow.8 Technological advances spurred by World War II allowed processors to offer more nutritious, palatable, and convenient foods, and consumers increasingly demanded such improved products. Meanwhile, progress in the biomedical sciences increased the understanding of human nutritional needs as well as chronic diseases. These pressures and advances have not yet abated, and, almost four decades later, one legitimately may ask whether Congress should consider once again updating (some would say overhauling) the statutory provisions governing food additives.9
The Food, Drug, and Cosmetic Act of 1938
Under the original FD&C Act, the FDA enjoyed broad responsibility but fairly weak regulatory authority over substances added to food. Section 402(a) provided that a food shall be deemed to be adulterated under the following circumstances:
If it bears or contains any poisonous or deleterious substance which may render it injurious to health; but in case the substance is not an added substance such food shall not be considered adulterated under this clause if the quantity of such substance in such food does not ordinarily render it injurious to health; or
if it bears or contains any added poisonous or added deleterious substance which is unsafe within the meaning of section 406 . . . 10
Section 406 of the Act provided in relevant part as follows:
Any poisonous or deleterious substance added to any food, except where such substance is required in the production thereof or cannot be avoided by good manufacturing practice[,] shall be deemed to be unsafe for purposes of the application of [the above-quoted] clause; but when such substance is so required or cannot be so avoided, the Secretary shall promulgate regulations limiting the quantity therein or thereon to such extent as he finds necessary for the protection of public health . . .11
The authority to promulgate food standards of identity under Section 401 provided the Agency with another, though cumbersome, regulatory mechanism for restricting the use of added substances by not authorizing their use as optional ingredients in standardized food.12
Under Section 402(a) of the FD&C Act, the FDA could initiate judicial proceedings to seize adulterated food or enjoin its continued marketing. 13 Unless the product exceeded one of the few tolerances established under Section 406, however, the agency would shoulder the burden of proving that the substance (1) was poisonous or deleterious and (2) may render the food injurious to health.14 As the Supreme Court held in construing the essentially identical language in the 1906 version of the statute,15 the government must show that the food containing a poisonous or deleterious substance creates a reasonable possibility of harm to consumers.16 Under the 1938 Act, the FDA shouldered an even greater burden of proof in the case of non-added substances (i.e., "ordinarily render it injurious to health").17 "Under either standard, the government must prove that the food itself probably will, or may, injure health, not merely that it contains a poisonous substance."18 Because of the time and effort required to undertake lifetime animal feeding studies of a substance, the FDA frequently could not satisfy this burden even if it had some legitimate basis for concern about safety.19
On the other hand, absent a tolerance established under Section 406, any added poisonous or deleterious substance would render a food adulterated,20 even if it was used at a level that was not "injurious" and served a useful purpose.21 In theory, Section 402(a)(2) reduced the FDA's burden of proof in taking enforcement action against a food product by eliminating any need for it to demonstrate that an added poisonous or deleterious substance might render the food injurious to health.22 The agency would still have to prove that an added substance was poisonous or deleterious. Once it made such a finding, however, the FDA would have difficulty authorizing the use of a safe and beneficial but technically "poisonous" substance in food.
Thus, the original statute created a safety standard for substances added to food that was both unnecessarily rigid and quite difficult for the FDA to administer. Because it shouldered the burden of proof, the agency experienced problems in regulating the introduction and marketing of foods containing potentially hazardous substances. Moreover, because the statute did not
require any advance notification of the introduction of a new additive for use in food, the agency might not even learn of its use unless safety problems subsequently came to light.23
Although the FDA initially lacked the authority to insist on testing, chemical manufacturers and food processors nonetheless had incentives to undertake limited safety evaluations. First, companies might fear eventual enforcement sanctions under the FDA's food adulteration provisions. Second, and more importantly, companies could not risk the adverse publicity that would accompany subsequently discovered hazards in their products. Finally, though less of an issue at the time, companies might face significant tort liability if consumers suffered injuries.24 In fact, most manufacturers did pretest new additives to some extent, but nothing prevented an unscrupulous firm from using an untested substance in food.25
The Food Additives Amendment of 1958
In 1950, Congress turned its attention to the growing use of chemical additives.26 After two years of hearings on the subject,27 a select committee of the House of Representatives chaired by James Delaney issued a report of its investigation. At the outset of its report, the committee noted the rapidly growing use of chemicals in the food supply: "There is hardly a food sold in the market place today which has not had some chemicals used on or in it at some stage in its production, processing, packaging, transportation, or storage."28 The committee recognized the substantial value of this growing use of additives, but it also expressed concerns about the accompanying potential health hazards.29 As noted in the report, the FDA estimated that approximately 700 chemicals were used in food at the time but that only 428 were known to be safe.30 The committee expressed particular concern over the lack of information about the possible chronic risks of existing food-use chemicals, and it urged Congress to amend the FD&C Act so that food additives would be governed by substantially the same safety and premarket review requirements applicable to drugs at that time.31
In the six-year period following the investigations by the Delaney Committee, Congress considered numerous bills and held further hearings on the subject.32 Unlike other legislative initiatives involving FDA-regulated products, which were prompted by real or perceived public health crises (most notably in the case of drugs), there was little apparent sense of urgency in crafting a proper response to the emerging concerns about new food additives. Instead, Congress acted with deliberation in designing an appropriate regulatory mechanism. 33 As an initial step, Congress amended the FD&C Act in 1954 to create a premarket approval system for pesticide residues in food, requiring the establishment of tolerances for any pesticide chemicals intended for use on a raw agricultural commodity.34
The food and chemical industries accepted the need for some system of premarket FDA review,35 but initially they preferred bills demanding little more than advance notification of an intent to market a new additive rather than bills requiring the issuance of a license by the agency.36 Administration officials objected that such an approach "would give no effect, except a delaying effect, to an adverse safety evaluation of the Secretary."37 Such a delay would give the FDA time to institute judicial proceedings to prevent marketing, where it only would have to establish that existing studies failed to demonstrate safety (rather than the more difficult burden of proving that an added substance was poisonous or deleterious and could render a food injurious to health); however, the agency would have to prove more than a reasonable possibility of harm in the abstract, under any condition of use; it would shoulder the more difficult burden of showing inadequate pretesting related to the safety of a substance for that particular intended use.
In contrast, the Department of Health, Education, and Welfare (HEW), of which the FDA was a part, sponsored a bill that would prohibit the use of a new food additive unless and until the FDA promulgated a regulation specifically authorizing its use.38 The agency did not conceive of this proposal as a product-specific licensing scheme authorizing a particular use by the applicant, as is the case with new drug approvals, but rather as a limitation on a proposed use uniformly applicable to all persons.39 Under such a generic approval regime, the FDA would issue a nonexclusive public regulation authorizing the use of an additive by any person wishing to do so, subject only to any patent protection remaining for the food additive petitioner. Without any recorded discussion of the issue, Congress ultimately opted for the FDA's proposed approach.
The bills also included various deadlines for FDA review of information about a new food additive. For instance, Congressman Delaney's original bill provided that a petition would be deemed approved if the Secretary failed to act on it within 60 days of filing.40 Under another bill, even if the Secretary concluded that an applicant had failed to demonstrate the safety of a new additive, a manufacturer or processor still could introduce the additive 30 days after notifying the agency of its intent to do so, unless the FDA first sought an injunction.41 The agency favored a premarket approval system but protested that it would need additional time to process submissions.42
In 1958, Congress enacted the Food Additives Amendment to the FD&C Act.43 By establishing a premarket review and approval system for food additives in new Section 409, Congress shifted the burden of proof on the safety issue from the FDA to the industry.44 Food additives now could not be used unless and until the agency deemed them safe. The legislation was not, however, motivated exclusively by safety concerns. Congress sought to promote continued innovation in food technology by giving the FDA greater flexibility to authorize limited uses of substances in food even if shown in animal tests to be poisonous at higher levels.45 The legislation also responded to proposals that had been introduced in a number of states to regulate new food additives. 46 The next three parts of this paper, focusing in turn on definitional, procedural, and substantive issues, set forth in greater detail the most important aspects of the Food Additives Amendment of 1958.
Definition of Food Additive
Definitional questions have assumed center stage in the application of the Food Additives Amendment over the last four decades. The statutory definition of the term "food additive" continues to pose significant interpretive difficulties. More importantly, as the food additive approval process has become increasingly cumbersome, industry has come to rely more heavily on certain exceptions enumerated in the definition. In fact, some of the recently suggested reforms focus on modifications of the existing definition in an effort to narrow the coverage of the Food Additives Amendment and, thereby, limit the scope of the FDA's premarket role.
Components of Food
The FD&C Act, as originally enacted, defined the term "food" as "articles used for food or drink .. [and] components of any such article."47 Putting aside the obvious tautology in the primary definition of the term, Congress clearly included "components'' of food within this provision.48 In addition, the operative provisions of the original statute—namely, the often vague prohibitions against adulteration and misbranding—refer to "substance[s]" which are added to food,49 as well as to "ingredients" from which a food has been fabricated. 50
Although Congress never directly modified any of these early definitions, the Food Additives Amendment effectively displaced many of the statute's broad prohibitions against adulteration insofar as they affected substances added to food.
The term "food additive" means any substance the intended use of which results or may reasonably be expected to result, directly or indirectly, in its becoming a component or otherwise affecting the characteristics of any food . . . , if such substance is not generally recognized . . . to be safe under the conditions of its intended use. . .51
By its terms, this definition does not include substances that are generally recognized as safe (GRAS), a potentially broad exception which, as discussed more fully in the next section, has become a central feature of the FDA's regulatory system. The definition enumerates a number of "indirect" additives included within its scope,52 but, as subsequently amended, it also specifically excludes several important substances that otherwise would be covered by the broad definition, including color additives, pesticide chemicals used on raw agricultural commodities, new animal drugs, and dietary supplement ingredients. 53 Other provisions of the statute now govern each of these latter substances.54
At its core, the definition of "food additive" applies whenever the manufacturer or food processor actually knew or should have known that a "substance" would become a "component or otherwise affect[ ] the characteristics of any food.'' The FDA routinely grapples with questions about the intended use of an item in order to determine its appropriate regulatory classification as, for instance, a food, drug, medical device, or cosmetic.55 Although Congress focused on "chemical" additives,56 items commonly available as a food might become a food additive (e.g., tomatoes in pasta sauce) unless otherwise excluded from the definition.57
The definition of "food additive" clearly covers more than just intentional, functional additives; Congress expressly included substances which may indirectly become a component or otherwise affect the characteristics of food by virtue of their use in production or processing.58 Such "indirect" or "incidental"59 additives may, for instance, include chemical substances used in food-contact packaging which unavoidably, and often at barely detectable levels, migrate into the food.60 Before the FDA may regulate a substance as an indirect food additive, however, it must have some evidence that the substance may reasonably be expected to migrate into food.61 The agency then may impose conditions on the use of a substance to assure that it will not migrate into food.
In the mid-1970s, the FDA grappled with evidence suggesting that acrylonitrile, a chemical used to fabricate plastic beverage bottles and other food-contact articles, might migrate into foods and might also cause cancer.62 After a formal hearing on the matter, the Commissioner decided that the acrylonitrile copolymer used to manufacture bottles qualified as a food additive whose safety remained unproven and, therefore, withdrew approval.63 The manufacturers immediately and successfully challenged this decision in court.
In Monsanto Co. v. Kennedy,64 the United States Court of Appeals for the District of Columbia Circuit reversed the Commissioner's decision. According to the court, the FDA had adduced no concrete evidence to support its conclusion that some residual (unpolymerized) acrylonitrile monomer would migrate from the interior surface of the plastic container into the beverage.65 The court remanded the decision so that the Commissioner could consider migration data from recently improved detection methods and also decide, even if migration was probable at an extremely low level, whether to exercise his discretion to disregard any such migration as
trivial.66 Relying on subsequently acquired migration data, the agency again concluded that acrylonitrile qualified as a food additive, but this time it also approved the additive as safe under conditions ensuring very little risk of migration.67
The FDA has approved innumerable indirect food additives over the years.68 In fact, the agency has devoted so much of its time and manpower to this task that it has, by comparison, seemingly neglected reviews of direct food additives.69 The FDA recently established an abbreviated exemption procedure for food-contact substances where the likelihood or extent of migration is so trivial—generally less than 0.5 parts per billion (ppb)-as not to require regulation.70 Except insofar as it affects the FDA's workload, the special questions posed by the regulation of indirect food additives are beyond the scope of this paper.
One troubling question concerns how to differentiate between a food additive and the food itself. Before a recent amendment to the statute, this problem arose most frequently with regard to dietary supplements, but the interpretive issue is broader. Over the years, the agency has attempted to classify nutritional supplements as food additives (or drugs) so as to shift the burden of proof on questions of safety (and effectiveness) from the government to the industry. In a pair of recent cases involving black currant oil (BCO), two federal appellate courts rejected the FDA's position that BCO qualified as a food additive when sold in capsule form.71 In both instances, the agency had seized shipments of the dietary supplement on the theory that the BCO, as a component in the dietary supplement capsules, was presumptively unsafe.72 Notwithstanding the rule of judicial deference to reasonable agency interpretations of ambiguous language in an enabling statute, 73 neither court found the FDA's reading of the "food additive" definition persuasive.74
According to the courts in these cases, only those components that somehow modify a food may be regulated as food additives. In the case of the BCO capsules, the courts held that the black currant oil was the sole active ingredient and, therefore, properly regarded as the entire food rather than an additive in the food.75 Indeed, in litigating the seizure actions, the FDA had conceded that BCO sold in bottles containing the pure liquid form would not qualify as a food additive; the courts decided that replacing the bottle with a gelatin capsule as the delivery vehicle for the BCO should not affect its regulatory status.76 In the capsule form, two inert substances (gelatin and glycerin) encased the BCO. The dispute turned on the interpretation of the phrase "becoming a component or otherwise affecting the characteristics of any food," which the agency wished to read in the disjunctive (i.e., either becoming a component or affecting the characteristics . . . ), while the processors favored an overlapping, unitary interpretation (i.e., affecting the characteristics by becoming a component or otherwise . . . ).77
Although hardly unambiguous,78 both courts accepted the latter construction of the statute, observing that BCO did not appear to constitute a component in the first place,79 and explaining that BCO certainly did not affect the characteristic of the gelatin and glycerin.80 If the FDA's interpretation were valid, then the agency could recharacterize every processed food as a food additive in order to shift the burden of proving the food's safety onto the processor, in effect reading the 1938 provisions governing food out of the statute even though Congress chose not to amend them in 1958. This judicial gloss on the definition may, however, excessively narrow the class of food additives insofar as it suggests that the "component" must meaningfully affect the characteristics of the food; it is not clear, for instance, whether the gelatin and glycerin (or an equivalent plasticizer) in the BCO capsules would fall within these courts' interpretation of the definition of the term.81
In earlier cases, courts have sustained efforts by the FDA to regulate similar products as food additives. Indeed, when a food contains more than one ingredient, each ingredient may be deemed a food additive, as the courts conceded in the BCO cases.82 For example, one federal district court accepted the FDA's contention that BCO was an additive when encapsulated with fish oil, vitamins, and minerals.83 Similarly, a federal appellate court recently decided that evening primrose oil was a food additive when encapsulated with Vitamin E.84
The difficult cases usually involved dietary supplements, in part because the FDA sought to find a meaningful regulatory mechanism for controlling these products. The agency also sometimes tried characterizing supplements as drugs rather than food.85 Indeed, the references to active and inactive ingredients make little sense in the context of conventional food products. To the extent that Congress has further restricted the agency's power to control the distribution of dietary supplements, these definitional issues may become less important in the future. Nonetheless, judicial interpretations of the statutory language will continue to limit the FDA's choice of regulatory responses to concerns about more conventional foods and food components.
In 1994, Congress amended the FD&C Act to address dietary supplements, 86 and it has been suggested that food processors might utilize these special provisions to avoid the rigors of food additive review.87 If a substance can be classified as a "dietary supplement" or as an "ingredient" in such a supplement, then it is excluded from the definition of the term "food additive."88 Unless the substance represents a "new dietary ingredient'' (because it was not marketed before October 15, 1994),89 a manufacturer or processor would only have to file a premarket notification with the FDA 75 days prior to commercial use.90
The definition of the term "dietary supplement," which was added to the statute in 1994, has several elements. First, it is a product:
intended to supplement the diet that bears or contains one or more of the following dietary ingredients: (A) a vitamin; (B) a mineral; (C) an herb or other botanical; (D) an amino acid; (E) a dietary substance for use by man to supplement the diet by increasing the total dietary intake; or (F) a concentrate, metabolite, constituent, extract, or combination of any [of these] ingredients . . . 91
Second, it "is intended for ingestion in tablet, capsule, powder, softgel, gelcap, or liquid form or, if not intended for ingestion in such a form, is not represented as conventional food and is not represented for use as a sole item of a meal or of the diet."92 Third, the product must be labeled as a "dietary supplement."93
It is possible that the manufacturer of a conventional food (such as a candy bar) containing a new food—use substance (such as an artificial sweetener or a fat—substitute developed in the 1980s) might assert that the product (1) is intended to supplement the diet, (2) contains one or more of the enumerated dietary ingredients (even if the new food-use substance itself did not fit within any of those categories), (3) is labeled as a "dietary supplement," and (4) is not otherwise represented as a conventional food. It seems unlikely, however, that a company would take such a gamble in trying to circumvent the FDA's food additive approval requirements for new food-use substances added to conventional foods, nor is it clear that such an approach provides much of an advantage over using the existing GRAS exception.94 Indeed, simply labeling the product as a "dietary supplement" will not by itself counteract other indicia that a product containing a new food-use substance is being represented as a conventional food, though the agency may be more hesitant to pursue enforcement efforts in such a case.
Generally Recognized as Safe
The statutory definition of "food additive" covers only a substance that "is not generally recognized . . . to be safe under the conditions of its intended use."95 Thus, in the peculiar meaning of the term as it is used in the statute, a substance that becomes a component of a food (even as an ingredient) would not be a "food additive" if it is generally recognized as safe (GRAS). Congress further defined GRAS as requiring that a substance used in food be
generally recognized, among experts qualified by scientific training and experience to evaluate its safety, as having been adequately shown through scientific procedures (or, in the case of a substance used in food prior to January 1, 1958, through either scientific procedures or experience based on common use in food) to be safe under the conditions of its intended use.96
Thus, GRAS status may exist if some level of scientific agreement about a substance's safety exists based either on appropriate testing or common use in food prior to 1958.
The GRAS exception obviously raises a number of serious interpretive difficulties. The
statute provides little further elaboration about the required degree of scientific agreement, the types of scientific procedures which could provide the necessary predicate for such agreement, or, in the case of substances in widespread use before 1958, the required nature and extent of such prior use. For instance, how could a substance be GRAS without experience based on common prior use—did Congress thereby intend to give manufacturers of new food-use substances the option of submitting test results to non-FDA scientists for their evaluation and possible stamp of approval? At least one witness at the 1957 congressional hearings apparently thought so, suggesting that a company could seek the advice of private or academic consultants on the question of whether there was general recognition of safety based on existing data.97 As discussed below, this has become a common practice.
Because the GRAS exception became a common feature of every one of the numerous bills on the subject considered by Congress,98 the legislative history sheds some additional light on these and other questions. Although not technically a "grandfather clause" (which would permanently exempt from coverage all substances used in food prior to the enactment date),99 the GRAS exception attempts to minimize the potentially significant and unnecessary burden that would otherwise be placed on both the industry and the FDA if the agency had to evaluate and formally approve common substances used in food.100 In addition, for substances that were not regarded as GRAS and therefore subject to regulation as food additives, Congress initially provided a transitional period of up to 30 months for compliance with the new premarket approval requirements,101 but it subsequently extended this phase-in period by almost five additional years. 102
The FD&C Act included similar GRAS language in defining the term "new drug,"103 as did the Pesticide Residues Amendment of 1954.104 Although conceding during the congressional hearings that the language was inherently ambiguous, the agency thought that it could apply this flexible GRAS exception to food additives in a sensible manner. 105 (Interestingly, just two years after enacting the Food Additives Amendment, Congress failed to include a GRAS exception in the Color Additive Amendments.106) As subsequently construed by reviewing courts, the exception applicable to drugs is quite narrow,107 in part because the statute requires that a drug be both GRAS and GRAE (generally recognized as effective). In 1973, the Supreme Court held that the exception in the definition of new drug required an "expert consensus" of both safety and effectiveness.108
FDA's implementing regulations, finally promulgated almost two decades after passage of the Food Additives Amendment, provide as follows:
General recognition of safety based upon scientific procedures shall require the same quantity and quality of scientific evidence as is required to obtain approval of a food additive regulation for the ingredient. General recognition of safety through scientific procedures shall ordinarily be based upon published studies which may be corroborated by unpublished studies and other data and information.111
Evidence of GRAS must relate to the conditions of intended use; general recognition of the safe use of a substance in a different product or at a different level would not suffice to escape the food additive definition.112 The exception turns not on safety itself so much as on recognition of safety by scientific experts. Testimony of an absence of any evidence of a health hazard would not suffice to establish GRAS status,113 at least not unless coupled with evidence of common prior use.114 If GRAS status is premised on common use prior to 1958, then such use must have been fairly extensive.115
Originally, the FDA categorically refused to recognize use outside of the United States.116 This policy did not, however, survive a subsequent judicial challenge.117 The revised regulations provide that prior foreign use may support GRAS status, but only if the information about such use is readily available and corroborated.118 In addition, GRAS status based on prior foreign use must satisfy domestic conceptions of safety.119 If GRAS status is based on prior foreign use, the FDA urges the manufacturer to seek its concurrence.120 Other sections of the regulations continue to define eligibility for GRAS status by reference to common use in the United States.121
The FDA's GRAS Lists. During the congressional hearings leading up to enactment of the Food Additives Amendment, the FDA submitted a "partial" list of what it would regard as GRAS substances including items such as butter, coffee, cream, gelatin, lard, lemon juice, margarine, molasses, mustard, olive oil, paprika, pepper, salt, sugar, vinegar, and wine.122 During the first several years after enactment of the Food Additives Amendment, the FDA listed in its regulations hundreds of ingredients as GRAS.123 The original GRAS lists included, for example, ascorbic acid, calcium chloride, caramel, and sodium phosphate.124
Because these inventories emerged without any detailed scientific assessment of the original safety data, much less of the data subsequently generated with constantly improving detection and safety assessment methods (as underscored by the discovery of evidence linking an artificial sweetener mixture containing cyclamate to cancer125), the FDA initiated a systematic review in 1969 in order to settle the GRAS or food additive status of a number of substances commonly added to food.126 The agency designated several categories of food ingredients for this review: substances of natural biological origin which were widely consumed as food before 1958 but subsequently were modified in certain respects by new production processes or selective breeding; distillates, isolates, extracts, and reaction products of GRAS substances; and substances not of natural biological origin or intended for consumption for other than their nutrient properties.127
The National Academy of Sciences (NAS) undertook ingredient usage surveys,128 and, in 1972, the Life Sciences Research Office (LSRO) of the Federation of American Societies for Experimental Biology (FASEB) established a Select Committee on GRAS Substances (SCOGS) to conduct reviews of the available scientific literature.129 Over a period of 10 years, SCOGS forwarded to the FDA detailed reports on 468 food substances (of which 422 were direct ingredients).130 The Select Committee first created an array of five standardized recommendations,131 and it concluded that 72 percent of the food substances under review
should remain GRAS and only one percent should immediately become subject to food additive requirements.132 Although the FDA planned to review each of these reports and pursue appropriate rulemaking, it has not completed its GRAS list review 15 years after receiving the last SCOGS report.133
A number of substances currently appear on the GRAS affirmation list that emerged from the FDA's comprehensive review. At present, almost 200 separate ingredients are included as GRAS for direct use in food. 134 The FDA concedes, however, that its GRAS lists are not exhaustive: "Because of the large number of substances the intended use of which results or may reasonably be expected to result, directly or indirectly, in their becoming a component or otherwise affecting the characteristics of food, it is impracticable to list all such substances that are GRAS."135 Thus, a substance "of natural biological origin that has been widely consumed for its nutrient properties in the United States prior to January 1, 1958, without known detrimental effects, which is subject only to conventional processing . . . will ordinarily be regarded as GRAS without specific inclusion" in one of the GRAS lists.136 More specifically, "by way of illustration, the Commissioner regards such common food ingredients as salt, pepper, vinegar, baking powder, and monosodium glutamate as safe for their intended use." 137
GRAS status does not free a substance of FDA controls. At a minimum, a GRAS substance must comply with any applicable food grade specifications appearing in the FOOD CHEMICALS CODEX,138 and it must perform an appropriate function (and be used at a level no higher than necessary to achieve its intended purpose) in the food or food-contact article in which it is used.139 In addition, a substance must comply with any specific usage limitations appearing in any GRAS affirmation regulation.140 If no specific limitations apply, GRAS status is lost only if the conditions of use differ significantly from those providing the basis for eligibility.141 Finally, "[n]ew information may at any time require reconsideration of the GRAS status of a food ingredient," and any revision of an existing GRAS regulation would be accomplished by the FDA through notice-and-comment rulemaking procedures.142 In contrast, any revision of a food additive regulation would require more cumbersome procedures.143
Private GRAS Determinations. Unless the FDA previously has decided otherwise, a person may take the position that a particular food-use substance is GRAS and, therefore, exempt from food additive approval requirements.144 In fact, there is no present requirement that the agency be advised of such private GRAS determinations. A few manufacturers have commissioned safety reviews by reputable scientific organizations, and FASEB has conducted a handful of private GRAS reviews during the last several years.145 For example, Procter & Gamble asked the Federation to review the safety of caprenin, a reduced calorie fat substitute; on the basis of FASEB's report, the company determined that this substance was GRAS, filed a GRAS affirmation petition with the FDA, and began selling it to food processors.146 Similarly, Nabisco Foods sought a FASEB review of salatrim, another fat substitute subsequently brought to market on the basis of a GRAS self-determination.147 Some have suggested that the National Center for Food Safety and Technology, an organization recently established in the Chicago area with private and government support, might play a similar role in the future.148
Whether undertaken for the FDA or a private entity, FASEB's LSRO assembles an ad hoc panel of experts from several different scientific disciplines to conduct the requested reviews. These experts usually are drawn from among FASEB's more than 40,000 members, and they are hired by LSRO to serve as independent consultants.149 The expert panels prepare study reports which are "peer-reviewed by an independent internal FASEB committee for clarity, objectivity, and scientific integrity . . . , [and] the reports of each study are published in scientific journals or are made available publicly."150
FASEB's review procedures depend to some extent on the terms of the particular agreement.151 The FDA often has contracted with the LSRO to conduct safety reviews of particular substances or to formulate recommendations on broad scientific questions. These contracts may set forth precise timetables for the solicitation of public input, the scheduling of meetings, and the completion of draft and final reports. As mentioned previously, the original GRAS list review required the creation of a special standing committee (SCOGS) charged with that enormous task, and FASEB managed to review almost 500 food-use substances in just 10 years. Its more recent safety reviews of BHA and MSG, again under contract with the FDA, took more than two years to complete and cost the agency approximately $300,000 each.152 Information about the speed and expense of FASEB's occasional private GRAS reviews is unavailable.
Similarly, a few industry associations have created their own expert panels to review the possible GRAS status of food ingredients, as the Flavor and Extract Manufacturers' Association (FEMA) has done for the last few decades.153 FEMA's project began in 1959, initially surveying the industry about the usage of different flavoring substances. The association then established a permanent panel—composed of six to eight recognized and independent experts from various disciplines including toxicology and biochemistry—to evaluate scientific literature reviews (SLRs) assembled for its consideration and then assess the GRAS status of those flavoring substances.154 Over the last 30 years, the expert panel's reports have been published periodically in the journal Food Technology.155 Furthermore, the SLRs underlying the Panel's GRAS determinations were made available to the public and forwarded to the FDA.156
In the case of a new flavoring substance, a company seeking an opinion about the flavor's potential GRAS status must submit an application form and literature search to FEMA's staff which, after a preliminary check for completeness, forwards the request and information to the expert panel for consideration at its next regularly scheduled meeting. 157 The available literature is evaluated against FEMA's published criteria, and a GRAS designation requires a unanimous vote by the panel; otherwise, the flavoring substance will be placed in a hold category for further study or be designated as not GRAS.158
Since the inception of this project, FEMA has considered more than 2000 thousand flavoring substances. The expert panel's initial set of reviews identified 1,118 substances as GRAS based on prior safe use and six more as GRAS based on the available scientific information. 159 The FDA incorporated only 277 of these flavors in its own GRAS list,160 but it also designated another 846 of these FEMA-reviewed substances as approved food additives on the strength of the existing safety data and without the need for filing separate petitions.161 In 1985, FEMA finished a comprehensive reevaluation aimed at updating its original GRAS determinations, dropping three flavoring substances from the list.162 In 1993, FEMA began a second reevaluation process, coupled with an effort to update and reformat all existing SLRs, which it hopes to complete in five years.163
The FDA has not challenged the marketing of flavors that FEMA has identified as GRAS, whether or not it has incorporated them into its own lists of GRAS substances or approved food additives.164 In extending the date by which persons would have to comply with its bulk flavoring requirements, the agency described the FEMA list as one of the ''reliable industry association GRAS lists."165 The FDA also occasionally refers to FEMA's GRAS listing of a flavor to support a GRAS affirmation proposal.166
Another potentially important source of food safety expertise resides in the Joint Expert Committees on Food Additives (JECFA), first organized in 1956 by the United Nations Food and Agriculture Organization (FAO) and the World Health Organization (WHO) and now associated with these organizations' Codex Alimentarius Commission.167 JECFA reports have influenced decisions by the FDA and other regulatory bodies, and its recommendations concerning
particular additives might be relied upon by companies in making GRAS self-determinations. Under the FDA's GRAS criteria, a report by JECFA or a comparable group certainly could qualify as "general recognition" of safety, even if a substance has never before been used in food.168
Although nothing prevents such GRAS self-determinations, the strategy carries obvious regulatory risks. On occasion, the agency has pursued enforcement proceedings, disagreeing with a company's belated claim that a substance is GRAS.169
Th[e] theoretical freedom of food processors to determine initially which ingredients require FDA licensure has little practical significance for widely used ingredients. Very few processors will purchase ingredients for which the supplier cannot provide documentation of FDA approval or acknowledgement as GRAS. Should a manufacturer independently conclude that an ingredient is GRAS it runs the risk that FDA may disagree and initiate regulatory action against its product.170
Some disagreement exists over who would shoulder the burden of proof on the question of GRAS status in the event that the agency challenges a company's self-determination; a few commentators have argued that the government would have to disprove that a substance is GRAS,171 but others have suggested that the burden would rest with the manufacturer to establish general recognition of safety once questioned by the FDA.172 (A similar question might arise when trying to determine whether a new substance falls within the scope of an existing GRAS or food additive regulation.173) As with the government's burden to prove adulteration under the original provisions of the FD&C Act, this is not a trivial issue.
In any case, the FDA allows any interested person to petition for an affirmation by the agency that a substance qualifies for GRAS status,174 and the mere filing of such a petition may protect a manufacturer from enforcement action during the pendency of the agency's review. 175 Because of what some commentators have described as a complete breakdown in the food additive approval process, the GRAS exception has come to serve an essential role in the existing regulatory scheme by allowing for the use of safe food substances without necessitating agency action.176 Whatever one thinks about the continued viability of the FDA's approval process, a matter taken up more fully later in this paper, the GRAS exception has proven to be a tremendously important feature of the current statutory design.
Prior Sanctions Exception. The statutory definition of "food additive" excludes "any substance used in accordance with a sanction or approval granted prior to the enactment of this paragraph pursuant to this Act, the Poultry Products Inspection Act . . . or the Meat Inspection Act."177 By virtue of the date of enactment of the Food Additives Amendment, such so-called "prior sanctions'' exist only if granted before September 6, 1958 (nine months later than the cutoff for prior use GRAS).178 A prior sanction may be either formal, as when a regulation establishing a food standard of identity includes a reference to an ingredient, or informal, usually in the form of correspondence from an appropriate agency official.179 The burden of demonstrating that a prior sanction exists rests with the person seeking to make use of this exception.
As with the exclusion for GRAS substances, a prior sanction only provides an exception for a particular food use of a substance.180 The agency has published regulations identifying a number of prior sanctions. Apart from nitrites, which were sanctioned by the Department of Agriculture as a color fixative and preservative in cured meats, 181 the prior sanctions listed by the FDA all cover substances used in food packaging materials.182 The agency recognizes that documentation of other sanctions may exist,183 and it invites interested persons to request publication of any such sanction, but the FDA also warns that a prior sanction will lapse if not
brought to the agency's attention once it undertakes consideration of a substance's food additive status.184
Unlike a GRAS determination, which the FDA always remains free to reconsider,185 a prior sanction exception to "food additive" status technically should be irrevocable,186 though the agency had once suggested otherwise.187 In any case, a prior sanction would not represent a permanent safe harbor for a substance. The sanction underlying the exception may, of course, be revoked.188 A prior sanction only provides an exception to the special statutory controls applicable to "food additives." If the FDA determines that a substance added to food "may render it injurious to health,''189 then the agency may restrict or prohibit its further use even if prior sanctioned. 190 A prior-sanctioned substance remains a "food'' subject to the FD&C Act's more general prohibitions against adulteration and misbranding. 191
The definitional issues discussed above are important only because they potentially affect the regulatory status of a substance.192 A food shall be deemed to be adulterated, thereby potentially triggering a variety of enforcement sanctions,193 "if it is, or it bears or contains, any food additive which is unsafe within the meaning of section 409" of the FD&C Act.194 Section 409 provides in relevant part that a food additive shall be deemed to be unsafe unless its particular use previously has been approved through the issuance of a regulation.195 In short, if a substance qualifies as a food additive, then it may not be used unless and until the agency specifically approves the additive.
This part elaborates on the procedures involved in seeking FDA food additive approval, reserving a discussion of the applicable safety standards for the next part. Because of the procedural and substantive complexities of the food additive petition process, food processors have begun to rely more heavily on a parallel mechanism that permits them to request that the FDA affirm their own judgment that a substance is GRAS. A comparison of the two review procedures reveals that these so-called "GRAS affirmation" petitions have provided no real shortcut for securing formal FDA authorization. There is an important difference, however, because nothing prevents the use of a substance before the FDA acts on such a petition. Although some have suggested that all direct additives undergo food additive review,196 this would surely overwhelm the agency. Others favor increased reliance on the GRAS exception to allow the FDA to allocate its resources more efficiently to review difficult food additive petitions.197 An evaluation of these and other suggested reforms must await a fuller description of different existing procedures and, in Parts V and VI, of the agency's actual practice over the years.
Food Additive Petitions
Normally, the issuance of a regulation authorizing the use of a food additive follows from a petition filed by the manufacturer of an additive or other interested person requesting such action.198 A food additive petition must contain information identifying the substance, its proposed use (as reflected in its labeling), all relevant data concerning its effect on food and residue detection methodologies, and full reports of safety studies.199 In addition, upon a request, the petitioner must provide the FDA with descriptions of production methods and facilities as well as samples of the additive for testing. 200 Certain confidential information and protected trade secrets provided to the agency are exempt from public disclosure.201
Pursuant to its implementing regulations, which reiterate and amplify the statutory procedures for the processing of food additive petitions, 202 the agency must notify the petitioner
within 15 days of receipt whether the petition is accepted for filing. 203 If a petition is not accepted for filing, the FDA will notify the petitioner in what respects the petition is incomplete.204 Within thirty days of accepting a petition for filing, the FDA must publish a notice of filing in the Federal Register containing the name of the petitioner and a brief description of the proposal.205 At any time before the agency issues an order, the petitioner may, upon its own initiative or at the FDA's suggestion, withdraw the food additive petition without prejudice to the right to refile it in the future.206
The statute and the FDA's regulations nowhere mention the availability of a public comment period immediately after the notice of filing and before issuance of an order.207 Although the judicial review provisions prompted substantial debate,208 the legislative history provides scant elaboration of the administrative procedures to be used with food additive petitions. HEW did, however, object to an industry suggestion that some sort of hearing procedure be available prior to the issuance of a final order.209
The agency's practice over the years confirms that no formal comment period was contemplated prior to the publication of an order.210 Of the hundreds of notices of filing published by the FDA in the Federal Register since 1965, none has invited public comment. Moreover, until fairly recently, the agency's decisions on food additive petitions have never alluded to or responded to comments submitted by third parties (the boilerplate language used in these orders simply refers to "data in the petition and other relevant information"), suggesting that such comments were rarely if ever filed in response to a petition.
In practice, the FDA now allows prepublication comment. When the agency decided in 1974 that the safety data contained in food additive and other petitions should be made available to the public, prepublication comment became increasingly feasible. In fact, the hope of soliciting public comment at this early stage appears to have motivated the change in the public information regulations.211 Although not explicitly required by the statute or the FDA's regulations governing food additive petitions, the agency now tacitly allows a post-filing opportunity for public comments on food additive petitions,212 and it explicitly invited such comments in the case of olestra.213
Within 90 days of filing, the FDA must issue a regulation or deny the petition, in either case notifying the petitioner of the reasons for such action, or, if the agency needs more time to study the petition, send written notification advising the petitioner of an extension of not more than 90 additional days.214 The agency's evaluation of a food additive petition requires input from a number of scientific and other FDA reviewers, as reflected in the accompanying diagram, and the statutory deadlines necessitate parallel rather than sequential reviews of a petition.215 Assuming that the agency has notified the petitioner of the need for an extension, a final order must be issued within 180 days after the filing of a food additive petition. In practice, these deadlines are rarely if ever met by the FDA.216
The agency's implementing regulations include provisions for "tolling" the statutory clock in certain instances where the petitioner submits further information.217 As explained by Dr. Alan Rulis, currently the Director of the Office of Premarket Approval at the FDA's Center for Food Safety and Applied Nutrition (CFSAN):
In practice, . . . the FDA allows this clock to run only when it has in its possession adequate information to review. Should it be determined that additional information must be submitted, a letter is issued to the petitioner and the clock stops. When a substantial supplement is received by the FDA, the clock is restarted again, but it may be reset to zero. Thus, the evaluation periods for petitions may extend beyond the nominal 180 days allotted by the statute.218
Although some delays result from the tardy submission of supplemental information by the petitioner, the agency generally does not meet the statutory deadlines even as modified by these
tolling rules. A final order granting or denying a food additive petition becomes effective immediately unless the FDA elects to stay the order pending any evidentiary hearing on objections filed after the publication of the order.219
The primary opportunity for external input on a food additive petition comes only after the FDA issues its final order in response to a petition. Within 30 days, "any person adversely affected by such an order may file objections thereto" and request a public hearing. 220
The Secretary shall, after due notice, as promptly as possible hold such public hearing for the purpose of receiving evidence relevant and material to the issues raised by such objections. As soon as practicable after completion of the hearing, the Secretary shall by order act upon such objections and make such order public. Such order shall be based upon a fair evaluation of the entire record at such hearing, and shall include a statement setting forth in detail the findings and conclusions upon which the order is based.221
This provision does not grant an automatic right to a hearing whenever a person files objections; within its discretion, the FDA may decide to forego a public hearing if the objections fail to raise any material issues.222 In fact, very few orders trigger any administrative or judicial challenge; in the FDA's history, only two orders involving direct food additives have provoked successful demands for a formal hearing.223 Apart from encouraging prompt action, the statute imposes no deadlines for responding to objections and hearing requests,224 and the agency sometimes has been slow to make a final decision.225
The FDA's order in response to objections may not take effect until 90 days after publication, in part to allow any person who will be adversely affected by such an order an opportunity to seek judicial review.226 Review is not available unless timely objections were filed to the FDA's initial order on a food additive petition.227 On judicial review, agency findings "with respect to questions of fact shall be sustained if based upon a fair evaluation of the entire record at such [public] hearing."228 If its order is not successfully challenged but significant new data subsequently become available, the FDA may, on its own initiative or in response to a citizen petition, issue a proposal to amend or repeal a food additive regulation.229
Sponsors of food additives spend approximately $20 million on average for research and development in pursuit of FDA approval.230 Unlike a new drug application, which gives the successful applicant an exclusive private license to sell a drug product, a food additive petition results in the promulgation by the FDA of a nonexclusive public regulation authorizing the use of the additive by any person wishing to do so. 231 In this respect, food additive regulations resemble FDA monographs for over-the-counter (OTC) drug products.232
The food additive petitioner may benefit from any remaining patent term.233 As delays at the agency have increased, however, this protection may dissipate soon after approval. Although limited patent term extensions may be granted by Congress, the statute does not extend to food additives the substitute market exclusivity periods now made available for approved new drugs.234 For this reason, ingredient manufacturers may no longer choose to pursue innovations that will necessitate the filing of a food additive petition.235
In order to generate the safety data required for approval, researchers must have access to the substance. An unapproved food additive may be used if exempted by the agency as "intended solely for investigational use by qualified experts when . . . consistent with the public health." 236 Unlike drugs and medical devices, which routinely make use of investigational exemptions to FDA approval requirements,237 food additives rarely are subject to an investigational use exemption. 238 Pursuant to implementing regulations, the exemption applies if an unapproved food additive is intended only for in vitro tests or investigational use in laboratory animals and is labeled as "Not for use in humans."239 If the investigational use involves food-producing animals, more stringent restrictions apply.240 In practice, the FDA has not objected to human
studies of new food substances so long as investigators comply with the agency's informed consent and other requirements for the protection of human subjects.
GRAS Affirmation Petitions
As explained above, even if a manufacturer or processor is confident about a substance's GRAS status, there is no assurance that the FDA (or a reviewing court) will agree with such a conclusion unless it previously has been GRAS listed or affirmed. Some private GRAS determinations represent a safer bet than others—the agency routinely defers to FEMA's decisions about flavoring substances, which are unlikely to pose a serious hazard in any event because of their extremely low level of use in food. In contrast, the FDA has challenged seemingly ad hoc and poorly supported claims of GRAS status asserted by a company in response to an agency charge that a food is adulterated for containing an unapproved additive.
In lieu of gambling on a private GRAS determination, a company might seek the FDA's concurrence on a substance's purported status as exempt from food additive approval requirements. Indeed, food processors may demand such concurrence before purchasing a substance from a manufacturer. After some experience responding informally to such requests,241 and in conjunction with its GRAS list review, the agency set forth specific procedures for seeking GRAS affirmation.242 Unlike food additive petitions, which are governed by fairly precise statutory procedures formulated by Congress, GRAS affirmation rests wholly on procedures developed by the FDA for this purpose.243
A person seeking affirmation of a substance's GRAS status must file a citizen petition containing information demonstrating that the substance satisfies the agency's GRAS criteria. In particular, such a petition must include a detailed description of the substance (including names, formulae, food grade material specifications, and descriptions of its quantitative composition and the manufacturing process244), a history of its use (including dates of use, data on past use, foods in which it was used, levels of use, and purposes), a description of available detection methods, and information concerning its safety and functionality.245 Unlike a food additive petition, the FDA will not recognize any claim for trade secret protection of information contained in a GRAS affirmation petition.246 The information concerning a substance's safety and functionality may include published scientific literature (or complete bibliographic references when copies of articles are not provided) or evidence that it is identical to a GRAS counterpart of natural biological origin, and the petition must include any adverse information or consumer complaints about the substance.247
Within 30 days of receiving a petition, the FDA must publish in the Federal Register a notice of filing and allow a 60 day period for public comment.248 After considering the petition and any comments received, the agency has two choices: if there is "convincing" evidence of GRAS status, the FDA will publish an order listing the substance in the regulations with other substances whose GRAS status it has affirmed; otherwise, it will publish a notice that the substance should be considered a food additive.249 Neither the statute nor the FDA's regulations establish any deadline for final action on a GRAS affirmation petition,250 and the agency has been very slow in acting on such submissions.251 In fact, petitions affirmed during the last decade have taken more than seven years on average.252 Thus, GRAS affirmation technically provides no shortcut over the food additive petition process,253 though, in practice, the concept remains attractive because manufacturers can and do market substances notwithstanding FDA inaction on a GRAS affirmation petition.254
Interim Food Additives
In 1972, the FDA designed a special transitional category for food-use substances whose safety has been called into question.255 The new procedural regulations emerged from the FDA's successful experience with issuing an interim order concerning brominated vegetable oil, a stabilizer in fruit-flavored beverages.256 As it explained in the preamble to the regulation:
The Commissioner recognizes that, with the vast increase in the quantity of scientific testing and in the sophistication of test methodology, there is virtually no[ ] natural or synthetic food substance that cannot be questioned on some technical ground. It would be impossible to require elimination from the food supply of every food substance for which such scientific questions have been or will be raised.257
The agency may promulgate an "interim" food additive regulation "when new information raises a substantial question about the safety or functionality of the substance but there is a reasonable certainty that the substance is not harmful and that no harm to the public health will result from the continued use of the substance for a limited period of time while the question raised is being resolved by further study."258 In promulgating an interim regulation, the FDA must abide by the same rulemaking procedures applicable to the issuance of a final food additive order,259 but the interim order is summarily revocable. The regulation may demand compliance "with whatever limitations the Commissioner deems to be appropriate under the circumstances.''260
In every case, a promise by one or more sponsors to undertake additional studies represents the primary condition for continued marketing. 261 Within 60 days after the effective date of an interim food additive regulation, an interested person must certify that "adequate and appropriate" studies have been undertaken to resolve the questions raised about the substance; otherwise, an order shall be published immediately to revoke the interim food additive regulation.262 In the preamble, the FDA emphasized that the purpose of an interim food additive regulation "is to assure that prompt action is taken to resolve the issue."263 Progress reports are due every six months; otherwise, the same revocation sanction may apply.264 "Promptly upon completion of the studies undertaken on the substance, the Commissioner will review all available data, will terminate the interim food additive regulation, and will either issue a food additive regulation or will require elimination of the substance from the food supply."265
During the past 20 or so years, the FDA has regulated the following substances as interim food additives: acrylonitrile copolymers, mannitol, brominated vegetable oil, and saccharin.266 In each case, the regulations called for additional toxicological studies, though with different levels of specificity.267 The original interim food additive regulation for saccharin contained a 17-month sunset provision,268 but the FDA subsequently revised the regulation to delete any expiration date. 269
With the exception of nitrites used in certain curing premixes,270 no previously listed interim food additive has been made subject to permanent restrictions or an outright prohibition. The FDA recently revised the mannitol regulation to authorize an additional method of manufacture without commenting on the fact that the "interim" regulation remained in place more than two decades after promulgation. 271 Apart from saccharin, which is the subject of a continuing regulatory moratorium imposed by Congress,272 the FDA's inaction on these interim food additives seems inexplicable.273 Courts have invalidated a similar "holding category" designed by the agency to authorize the continued use of an OTC drug product while awaiting the results from further studies of an active ingredient.274 Nevertheless, the FDA and the industry seem content to leave these particular additives in what has evolved into a semi-permanent regulatory limbo.
The discussion to this point has focused on definitional issues concerning which substances must undergo the food additive approval process and procedural issues about the processing of food additive and GRAS affirmation petitions. Ultimately, the FDA must evaluate such petitions in order to determine whether a substance is safe for a particular use in food. The criteria for this review raise fundamental questions of science and public policy. The definitional and procedural questions are not, however, simply a prelude for this discussion because choices about the substantive safety criteria cannot be disentangled entirely from decisions about their appropriate reach and mechanisms for their efficient application.
General Safety Standard
The FDA may not approve a food additive for a particular use unless the data presented to the agency establish that the additive is safe for that use.275
In determining, for purposes of this section, whether a proposed use of a food additive is safe, the Secretary shall consider among other relevant factors—
the probable consumption of the additive and of any substance formed in or on food because of the use of the additive;
the cumulative effect of such additive in the diet of man or animals, taking into account any chemically or pharmacologically related substance or substances in such diet; and
safety factors which in the opinion of experts qualified by scientific training and experience to evaluate the safety of food additives are generally recognized as appropriate for the use of animal experimentation data.276
Unlike most drugs, food additives are likely to be consumed by all segments of the population, including children and the elderly, potentially over the full course of their lifetimes. This difference in the aggregate exposure levels, coupled with the typically less direct benefits derived from particular food substances, helps explain the FDA's generally more conservative approach to safety assessment in this area.
The intended meaning of the above-quoted safety standard was discussed in the house report accompanying the bill that would become the Food Additives Amendment of 1958:
The concept of safety used in this legislation involves the question of whether a substance is hazardous to the health of man or animal. Safety requires proof of a reasonable certainty that no harm will result from the proposed use of an additive. It does not—and cannot—require proof beyond any possible doubt that no harm will result under any conceivable circumstance. This was emphasized particularly by the scientific panel which testified before the subcommittee. The scientists pointed out that it is impossible in the present state of scientific knowledge to establish with complete certainty the absolute harmlessness of any chemical substance.277
Thus, for food additive approvals, the statute demands neither confirmation of safety under all possible conditions of use,278 nor an absolute assurance of safety for every intended use. In addition, pursuant to its implementing regulations, the FDA has defined GRAS to require the same quantity and quality of safety data as required for food additive approval, unless the substance was in common use prior to 1958.279
As explained previously, courts have interpreted the safety standard contained in the general adulteration provisions as requiring proof by the FDA of a reasonable possibility of harm to consumers in support of a seizure or injunction of a substance in use.280 During hearings on the
food additive legislation, the administration objected to the inclusion in the statute of a definition of "safe" that explicitly used a "reasonably probable" standard for deciding whether to allow for the use of a substance in food.281 As the sponsor of the bill reiterated on the floor of the House, the legislation would not require proof of safety ''beyond any possible doubt" but rather a "practical certainty'' that no harm would result.282 Although "harm" is not separately defined, Congress evidently understood the term to mean a capacity to injure or otherwise cause disease.283 Presumably, this view would exclude undesirable side effects that posed no risk of adverse health consequences.
In its original implementing regulations, the FDA defined safe to demand "convincing evidence which establishes with reasonable certainty that no harm will result."284 In 1971, the agency revised this definition to require proof of "no significant risk of harm,"285 but it revised the definition again in 1976 to more closely track the legislative history quoted previously.286 In its current food additive regulations, the FDA defines "safe" as meaning "that there is a reasonable certainty in the minds of competent scientists that the substance is not harmful under the intended conditions of use."287 Unlike the current (though identical to the original) food additive definition, the color additive regulation demands "convincing evidence that establishes with reasonable certainty that no harm will result."288 Although one could argue that the narrower interpretation of the statutory safety standard contained in the original implementing regulations (demanding "convincing evidence") is entitled to greater deference than the current version, 289 courts have not directly addressed the validity of the current interpretation or suggested that it imposes any particularly significant hurdle to food additive approvals.290
In elaborating the food additive safety standard, the FDA's regulation tracks the factors enumerated by Congress.291 In general, the FDA has decided to follow "the principles and procedures for establishing the safety of food additives stated in current publications of the National Academy of Sciences-National Research Council."292 For instance, unless presented with evidence justifying a different margin of safety, the agency will apply a safety factor of 100 when using data from laboratory animals; "a food additive for use by man will not be granted a tolerance that will exceed 1/100th of the maximum amount demonstrated to be without harm to experimental animals."293 In other words, the estimated daily intake (EDI) of the additive, based on its intended use(s) and surveys of likely consumption patterns, may not exceed the allowable daily intake (ADI), which generally is set at 1/100th (as adjusted for body weight) of the no observable effect level (NOEL) derived from animal studies.294
Agency scientists from a number of different disciplines, including chemistry, toxicology, nutrition, and microbiology, might review a food additive petition.295 This safety assessment task calls for the exercise of professional judgment: the "FDA's approach is better characterized as an attempt, using the best, but admittedly imperfect, tools science makes available, to manage the uncertainty inherent in food safety evaluation."296 Agency reviewers have been criticized, however, for excessive conservatism and even paralysis in the face of such uncertainty.297 Whether or not such criticisms are well placed, the "recursive" nature of the evaluation may consume considerable time.298
The FDA also has endeavored to specify acceptable study designs for generating the safety data necessary to support the approval of a food additive but without imposing unduly rigid guidelines.299 In preparing food additive petitions, regulated firms benefit greatly from knowing in advance what types of toxicity studies the agency will deem acceptable. Although safety testing methods recommended by the NAS-NRC are presumptively acceptable, a petitioner may utilize alternative procedures if shown to be at least equally reliable.300 In addition, upon request the agency "will advise a person who wishes to establish the safety of a food additive whether [it] believes the experiments planned will yield data adequate for an evaluation of the safety of the additive."301
The agency once announced plans for a "cyclic review" of all approved food additives, in effect to update its original safety decisions, 302 but it eventually discontinued work on this gargantuan task.303 As part of the process, however, the FDA announced in 1982 the availability of its Toxicological Principles for the Safety Assessment of Direct Food Additives and Color Additives Used in Food (which is simply referred to as the "Redbook").304 In recently announcing the availability of a revised draft ("Redbook II''), the agency explained that this document "is intended to provide guidance to petitioners regarding criteria used by FDA for toxicological safety assessment of direct food additives and color additives, and in the agency's evaluation of the generally recognized as safe (GRAS) status of food ingredients,'' 305 arguably expanding its original scope beyond the review of food and color additive petitions.306 The FDA continues working to finalize Redbook II.
Under its existing procedural regulations, guidelines are binding only in the limited sense that the agency must accept submissions that have been prepared in conformity with such guidelines. These regulations expressly disavow any attempt to impose binding obligations on regulated parties,307 and the FDA has taken the position that the Redbook does not mandate the submission of particular toxicology information in food additive petitions.308 In practice, of course, petitioners will hesitate to deviate from agency guidelines for safety testing because of a fear that the failure to adhere to those guidelines might lead to delays or even disapproval.
The FDA recently proposed revising its general administrative procedures so that agency personnel would not be bound to adhere to formally announced opinions and guidelines.309 The utility of the new Redbook would be sharply reduced, however, if the agency ultimately makes such a revision to its generic procedural rules. It would substantially interfere with research and development efforts if sponsors could no longer conduct chronic animal studies with the assurance that these expensive and time-consuming studies would be acceptable to FDA reviewers when the studies were performed in accordance with the Redbook's guidelines.
Finally, Congress and the FDA have struggled to decide what role, if any, a consideration of benefits should play in applying the general safety standard. Earlier versions of the 1958 legislation would have premised approval on an agency finding of "functional value" in addition to safety.310 In response to strong industry opposition,311 the final legislation required only a consideration of functionality in the limited circumstances where considerations of safety necessitated a quantitative limitation on use and then without any judgment of social value.312 In promulgating its revised food additive and GRAS regulations in 1976, the FDA suggested that social utility is "inevitably a factor" in the safety determination,313 but a subsequent Commissioner expressly disavowed this statement a few years later during the course of withdrawing the new animal drug approval for DES.314
Carcinogenicity and the Delaney Clause
The general safety standard is modified by a proviso known as the Delaney clause which directs that "no additive shall be deemed to be safe if it is found to induce cancer when ingested by man or animal, or if it is found, after tests which are appropriate for the evaluation of the safety of food additives, to induce cancer in man or animal." 315 Professor Richard Merrill once remarked that "[t]he Delaney Clause is perhaps the most discussed, yet least used, provision of the [FD&C] Act."316 Indeed, lay press coverage incorrectly suggested that the recently passed Food Quality Protection Act repealed the Delaney clause,317 when Congress actually had only exempted pesticide residues remaining in processed foods and substituted a general safety standard for setting tolerances for such residues.318 To be sure, the recent legislation reflects congressional concerns about undue rigidities in the Delaney clause, but the proviso remains
fully applicable to food additives. Whatever its actual frequency of use, the Delaney clause has shaped much of the debate about the FDA's food additive approval process.
The Delaney clause appeared fairly late in the legislative process. The bill approved by the House Committee on Interstate and Foreign Commerce did not contain an anti-cancer clause, but the Committee, at Congressman Delaney's urging, added the proviso to the bill before bringing it to a vote on the House floor.319 As explained in the subsequent report of the Senate committee considering the bill:
We have no objections to that amendment whatsoever, but we would point out that in our opinion it is the intent and purpose of this bill, even without that amendment, to assure our people that nothing shall be added to the foods they eat which can reasonably be expected to produce any type of illness in humans or animals. We applaud Congressman Delaney for having taken this, as he has every other opportunity, to focus our attention on the cancer-producing potentialities of various substances, but we want the record to show that in our opinion the bill is aimed at preventing the addition to the food our people eat of any substances the ingestion of which reasonable people would expect to produce not just cancer but any disease or disability. In short, we believe the bill reads and means the same with or without the inclusion of the clause referred to. This is also the view of the [FDA].320
Although the drafters of the 1958 legislation claimed that the Delaney clause was no stricter than the general safety standard, its application has proven to be far more troublesome in practice.
The terms of the Delaney clause indicate that, if presented with data suggesting a carcinogenic effect, the FDA was to conduct a fact-based scientific inquiry. The language selected by Congress contemplated that the FDA will exercise scientific judgment in assessing the relationship between exposure to a chemical and a carcinogenic response.321 Technically, the Delaney clause does not apply to GRAS substances (or prior sanctioned ingredients), though the FDA initially took the position that evidence of carcinogenicity would undermine GRAS status.322
The FDA's approach to questions concerning the putative carcinogenicity of added substances conforms with Congress' expectations that it would employ scientific judgment. Indeed, the agency has played a leading role within the federal government to ensure that the best science is brought to bear in evaluating the carcinogenicity of chemicals, such as through the Interagency Regulatory Liaison Group (IRLG) and the Interagency Staff Group on Carcinogens of the Office of Science and Technology Policy (OSTP). FDA scientists were leading participants in these efforts, and the resulting reports, one released in 1979 by IRLG323 and the other released in 1985 by OSTP,324 provide strong support for the role of scientific judgment in evaluating evidence relating to carcinogenicity.325
The FDA's own practice over the years exemplifies the central principles embodied in these reports. The agency does not accept, or act upon, the proposition that a mere temporal association between administration of a chemical to test animals and an elevation in tumor incidence justifies, much less requires, a finding that the chemical "induces cancer." Study results suggesting such an association are instead subject to critical evaluation, and any conclusion is the product of an assessment of many different scientific elements. Evidence that a test substance is associated with a statistically significant increase in tumor incidence compared with concurrent controls is a necessary, but not sufficient, basis for concluding that it "induces cancer." For example, if the increase of tumors in test animals falls within the range commonly observed in historical controls, the significance of the test results is placed in question.
The agency frequently looks beyond quantitative analysis of tumor incidence to consider, based on qualitative scientific evidence, whether the effects observed in animal studies are biologically significant: "Determination that the incidence of neoplasms increases as the result of exposure to the test compound requires a full biological, pathological, and statistical
evaluation. Statistics assist in evaluating the biological conclusion, but a biological conclusion is not determined by the statistical results."326 The OSTP guidelines make the same point: "the final judgment regarding a compound's carcinogenicity is rarely based on a single finding of statistical significance. One must be assured that the effects observed are biologically significant as well."327 Thus, the FDA routinely considers relevant biological and biochemical data in determining whether a substance "induces cancer," including information on a variety of other biological functions such as dose response, tumor progression, and tumor latency, as well as the overall evidence from other studies.328
Even animal studies suggestive of significant carcinogenic activity are not relied upon by the agency without a close assessment of the overall weight of the evidence. In approving the color additive FD&C Blue No. 2, for example, the FDA stated that "the agency has consistently asserted that statistical factors must be analyzed in conjunction with biological factors in determining what, if any, conclusions can be drawn from a study."329 The FDA relied on a similar weight-of-the-evidence analysis when it approved D&C Green No. 5, identifying the following evidence as support for its conclusion that the color additive did not induce cancer:
(1) the fact that the incidence in the high dose group is within the normal range of background spontaneous incidence; (2) the absence of evidence of progressiveness that would be expected from a carcinogen; (3) the absence of observable non-neoplastic disease in the mouse liver; (4) the negative findings on testing D&C Green No. 5 for mutagenic activity; and (5) the atypical form of the apparent dose-response data.330
The agency concluded, on the basis of this weight-of-the-evidence analysis, that the color additive was non-carcinogenic and nontumorigenic, and it defended the decision a few months later when it rejected objections to the final regulation listing Green No. 5.331
More recently, in approving the sweetener acesulfame potassium, the FDA based its decision that this food additive is not a carcinogen on "the weight of all of the evidence; no single point provided complete proof in determining the question of carcinogenicity."332 In denying a petition for a hearing, the FDA defended its conclusion that the incidence of mammary gland tumors in female rats was not treatment-related by pointing out that these tumors are common in old-age rats of the strain used in the bioassays, that the incidence of these tumors fell within the range for historical controls, and that there was no evidence of a dose-response relationship or evidence of progressive tumor stages.333
The agency also once explained that exaggerated intakes of several common products may cause tumors through a secondary mechanism, but it concluded that "these foods and drugs are not, by reason of their capacity to induce liver damage when abused by being consumed at high levels, properly classified as carcinogenic because of their potential association with a higher rate of liver cancer."334 Over the last two decades, the FDA has continued to allow the use of a few substances that appear to cause tumors only at high-dose levels in animals, in part because it is well recognized that such substances are ubiquitous in the human diet. For instance, when the FDA approved a petition to allow selenium supplementation of livestock feed, it had to grapple with three studies that had found neoplastic lesions in the livers of rodents fed high doses of the substance. The agency ultimately concluded that there was no cancer risk to humans, discounting the observed tumors as linked to liver damage which occurred only at high doses.335 The application of the Delaney Clause to substances suspected of causing cancer by a secondary mechanism continues to provoke controversy, and the FDA's current policy on the subject remains unclear.336
Finally, if the agency finds that an additive has not induced cancer, then it applies the general safety standard rather than the Delaney clause in assessing any evidence that one of the additive's constituents is a suspected carcinogen. As the FDA explained in support of its policy:
[E]ach chemical in the complex mixture that constitutes a food additive could itself be considered to be a food additive. Each of these chemicals in some sense becomes a component of the food. . . . [T]he Delaney Clause does not apply to a carcinogenic chemical in a food additive absent a finding, after appropriate tests, that the additive as a whole induces cancer.337
Instead, the agency has applied the "reasonable certainty of no harm" interpretation of the general safety clause in a flexible manner that allows it to disregard truly negligible risks, and the courts have sustained the legality of the FDA's constituents policy.338 For instance, in discussing the proper application of the Delaney clause to methylene chloride as a food additive used to decaffeinate coffee,339 the FDA explained that it had used a one-in-one million lifetime risk of cancer standard "for determining whether the calculated upper bound risk of cancer posed by an impurity is low enough to be considered 'safe' within the meaning of the general safety clause." 340 With improvements in scientific methods for detecting substances and evaluating their potential for causing cancer, such issues will arise even more frequently in the future.341
Special Restrictions: Labeling
Although the Delaney clause may trigger disapproval of a carcinogenic food additive without regard to risk, the general safety standard applicable to food additives provides greater flexibility. A safety question unrelated to potential carcinogenicity would not automatically doom a food additive because the agency may be able to address such concerns by imposing restrictions on the use of the additive. The statute provides that, when the FDA issues a food additive regulation, it may prescribe the conditions for safe use:
including, but not limited to, specifications as to the particular food or classes of food in or on which such additive may be used, the maximum quantity which may be used or permitted to remain in or on such food, the manner in which such additive may be added to or used in or on such food, and any directions or other labeling or packaging requirements for such additive deemed necessary by [the Secretary] to assure the safety of such use.342
In issuing food additive regulations, the agency generally has imposed only one or more of these enumerated conditions where considered necessary, though it recently has become somewhat more creative in this regard.343 The agency may include similar limitations in GRAS affirmation regulations,344 but it has done so only infrequently.
Labeling requirements represent one of the significant conditions that may be imposed during the food additive approval process. The FDA relies to a great extent on labeling requirements as a mechanism for regulating the numerous products subject to its jurisdiction. 345 The agency has, however, mandated relatively few warnings for food products, though it has imposed special disclosure requirements in certain cases. "The agency's primary tool for handling a situation where population subgroups may be at increased risk from a food ingredient that is safe for most people is to use [ingredient] labeling to inform those persons who need or want to avoid the ingredient."346 Thus, food product labeling must disclose the presence of substances such as the color additive FD&C Yellow No. 5 (tartrazine)347 and sulfiting agents used as preservatives.348
In addition, bulk packages of certain additives or mixtures containing these additives, such as BHA and potassium iodide, must quantify the level of the additive,349 and curing mixes containing sodium nitrate or nitrite must include "Keep out of the reach of children" on the label.350 Any food containing aspartame must bear a statement advising persons who suffer from phenylketonuria (a rare metabolic disorder) that the product "contains phenylalanine."351
Finally, the FDA sometimes mandates the disclosure of information about possible non-serious side effects.352
Products that may trigger allergic reactions provide one of the best justifications for choosing a labeling strategy. In the case of Yellow No. 5, the FDA found evidence of a causal link between the color additive and serious allergic-type reactions in susceptible individuals. 353 The agency rejected comments urging it to ban Yellow No. 5, explaining that it selected the label declaration because this option minimized the societal impact while providing an adequate measure of protection for those sensitive to the color.354 In the case of sulfiting agents, the FDA did prohibit uses of these additives where labeling would be impractical (e.g., raw produce).355 For packaged foods, however, the agency opted for a simple ingredient declaration on the label, rejecting comments urging that a "Warning" be required.356 The FDA's recent and wideranging revisions of its food labeling regulations will provide even more detailed ingredient information of possible relevance for persons with allergies,357 but the agency again rejected suggestions that explicit allergy "Warnings" appear on product labels. 358
The FDA has taken the position that warnings on food products are appropriate only when based on sound scientific data with clear application to human health, stating that it "is unwilling to require a warning statement in the absence of clear evidence of a hazard."359 Substances found in foods sometimes have been associated with chronic hazards, but the FDA has never suggested that every food product containing a suspected carcinogen should bear a warning statement absent evidence of a risk under actual conditions of use.
[A] requirement for warnings on all foods that may contain an inherent carcinogenic ingredient or a carcinogenic contaminant .. would apply to many, perhaps most foods in a supermarket. Such warnings would be so numerous they would confuse the public, would not promote informed consumer decision-making, and would not advance the public health. 360
The FDA will prohibit use of the food or food additive (as it had proposed to do with saccharin361) if a chronic risk appears to be serious, rather than require a label warning,362 notwithstanding objections that such a policy unduly limits consumer freedom of choice. 363
The FDA has acknowledged the problems of information overload and dilution of warnings in a number of different contexts.364 Even if consumers do pay attention to warnings, they may well overreact to the information, particularly when the warning statement is intended to convey a subtle message about low probability risks. Requiring warning statements about unsubstantiated or insignificant risks in the labeling of useful products can distort consumer choices. For instance, many food products contain potentially carcinogenic nitrites, but at present these preservatives provide one of the best means available for protecting against botulism.365 In addition, the public should be more concerned with ensuring consumption of an adequate and healthy diet than with warnings about possible low-level carcinogens in food products.366 Absent consistent and differentiated cautionary statements allowing comparisons among the risks posed by different products, narrowly focused warning requirements may distort these trade-offs. To date, the FDA has only infrequently utilized its power to condition food additive approvals on special labeling statements.
The agency's statutory authority to impose these labeling requirements is hardly clear. The FDA does enjoy broad authority to regulate the labeling of food products generally, under the statute's prohibitions against misbranding, but Section 409 does not appear to authorize the imposition of special labeling requirements for the finished food product as a condition for food additive approval.367 Moreover, the FDA requirements for label disclosures of non-serious side effects do not represent conditions necessary to ensure safe use so much as to avoid consumer deception or product misbranding.368 The risk of deception represents a legitimate basis for
denying a food additive petition but technically not an appropriate ground under the statute for approving a petition subject to certain food labeling conditions.369
Although each of the examples discussed in this part is somewhat atypical (one might even say precedent-setting), these case studies should help illustrate a variety of the challenges encountered by the FDA in applying the definitional, procedural, and substantive aspects of the Food Additives Amendment. The agency experienced special difficulties in reviewing the artificial sweetener aspartame, Calgene's bioengineered tomato, and the fat substitute olestra. (Several less notable but still important additives have been discussed previously where pertinent.) As scientific advances increasingly create novel food-use substances, the FDA will continue to face difficult regulatory questions.370
Artificial Sweeteners: Aspartame
Aspartame, a chemically bound combination of phenylalanine, aspartic acid, and methanol,
is an artificial sweetener sold under the brand name NutraSweet™ and is the first sugar substitute approved in the wake of the controversies involving cyclamate and saccharin. G.D. Searle & Co. submitted a food additive petition for aspartame in 1973,371 and, more than one year later, the FDA approved this petition, authorizing certain "dry" uses of the additive.372 Because a rare genetic disorder prevents the metabolism of phenylalanine, which at very elevated levels poses a risk of mental retardation, the agency required that the statement "Phenylketonurics: Contains Phenylalanine" appear on the label of products containing aspartame.373
A few persons filed objections to the FDA's approval and requested a hearing. The agency eventually decided to stay the regulation's effective date pending a hearing on these objections because it discovered data discrepancies by a contract laboratory in the toxicological testing of other substances.374 Four years later, the FDA convened a three-member Public Board of Inquiry to conduct the hearing.375 The Board agreed with the FDA's initial determination that aspartame posed no risk of causing brain damage or neuroendocrine dysfunction, but it recommended against approval because of unresolved questions about brain tumors found during rodent studies of aspartame.376 The Commissioner rejected this recommendation and finally affirmed the original food additive regulation in 1981.377 He decided, however, to impose the following additional condition: "Searle is to monitor the actual use levels of aspartame and to provide such information on aspartame's use to the Bureau of Foods as the Bureau may deem necessary by an order, in the form of a letter, to Searle."378 One year later, Congress extended aspartame's patent term to restore some of the time lost during the lengthy delays in securing final FDA approval.379
In 1982, Searle submitted a food additive petition requesting an amendment to allow the use of aspartame in carbonated beverages,380 and the FDA approved this use the following year.381 Several persons filed objections to the regulation, requesting an immediate stay and a public hearing, but the agency rejected the objections without conducting a hearing.382 One consumer group brought a judicial challenge to the approval for use in carbonated beverages, but the court affirmed the FDA's decision.383 The court held that the objections did not raise material issues concerning the wet use of aspartame and that the more general concerns about the additive's safety had been resolved during the course of the original approval for dry use.384 In particular, the court found no material issues warranting a hearing on the objections concerning the risks of brain damage from changes in neurotransmitter production, nitrosamine formation,
carcinogenicity, embryotoxicity of degradation products, increased levels of consumption, or more rapid decomposition associated with its use in carbonated beverages.385
The FDA subsequently amended the food additive regulation on two dozen occasions to approve additional uses of aspartame,386 again rejecting renewed objections and denying hearing requests.387 Although inefficient, companies apparently have found this incremental approval approach to be somewhat beneficial—the FDA may approve the initial petition for a limited use more quickly and then accord supplemental uses some presumption for approval. In 1996, the agency approved a petition requesting that aspartame be permitted for use as a general purpose sweetener, thereby consolidating the 27 narrow sweetener uses that it had approved over the last couple of decades.388 Nonetheless, controversy about the FDA's approval of aspartame has persisted,389 prompting the agency to continue monitoring consumer complaints.390 Meanwhile, the attention of consumer activists has shifted to the FDA's review of newer artificial sweeteners and other direct additives. 391
Biotechnology: Calgene's Tomato
In 1994, Calgene began marketing the FLAVR SAVR™ tomato, a tomato which can remain on the vine longer, has a longer retail shelf life, and exhibits improved viscosity when used in processed foods. Calgene created the FLAVR SAVR™ through a recombinant DNA technique, specifically by isolating the polygalacturonase gene, which is responsible for producing the enzyme that breaks down pectin in the cell walls of the tomato during the ripening process, and reintroducing the gene into the plant in the reverse or "antisense" orientation.392 The antisense copy suppresses the production of the polygalacturonase enzyme. In this respect, Calgene's process simply rearranged the tomato's own genetic material (it was not replaced by material from a different organism), and it reduced the levels of an existing protein (it did not introduce a new chemical or increase the levels of other naturally occurring substances in the tomato).
In addition, the process required the insertion of a selectable "marker" gene, in this case the commonly used kanamycin resistance (kanr) gene derived from certain bacteria. Marker genes help identify those plant cells which have successfully taken up the transferred gene controlling a desired trait; in this case, cells that contain the kanr gene synthesize a protein, aminoglycoside 3'-phosphotransferase II (APH(3')II), which renders the cell resistant to the action of antibiotics.393 By adding kanamycin or a similar antibiotic to the laboratory growth medium, a researcher can screen out the many tomato plant cells with unsuccessful gene transfers. Although the marker gene serves no useful purpose after that point, it remains part of the genetic material in every cell of the growing transgenic plant.
After the completion of field testing authorized by the USDA,394 Calgene petitioned the FDA to issue advisory opinions on the regulatory status of both the kanr gene395 and the FLAVR SAVR™ tomato.396 One year after the FDA's 1992 publication of a policy statement on foods derived from new plant varieties,397 the company converted its initial request into a petition seeking food additive approval of the protein from the kanr gene as a processing aid.398
In 1994, shortly after holding a meeting of its Food Advisory Committee, the FDA responded to Calgene's petitions. First, it "concluded that FLAVR SAVR™ tomatoes have not been significantly altered when compared to varieties of tomatoes with a history of safe use."399 Although not stated in such terms, the FDA in effect found that this tomato was GRAS and, therefore, not a food additive when used in processed foods.400 The agency had once before approved a GRAS affirmation petition for bioengineered versions of the milk-clotting enzyme chymosin for use in making cheese and other products, as a substitute for the animal-derived version of the enzyme which previously had been affirmed as GRAS.401
Second, the FDA granted Calgene's food additive petition, approving the use of APH(3')II encoded by the kanr gene as a processing aid for developing certain new plant varieties.402 The agency concluded that the protein produced by the kanr gene is neither toxic nor allergenic and that estimated dietary exposure would be extremely low.403 Although it provides antibiotic resistance and could, therefore, interfere with the therapeutic use of human antibiotics, the FDA found that the protein would be inactivated and degraded by digestion or heating. 404 Finally, the agency concluded that the proposed additive created no risk of horizontal transfer of the kanr gene (and the spread of antibiotic resistance) into human cells lining the intestinal walls or into microorganisms found in the intestines or in the soil.405
Although the FLAVR SAVR™ tomato posed fairly limited complexities, the FDA and Calgene spent considerable time and effort reviewing its safety and deciding how best to regulate the product. In the future, biotechnology should become increasingly important in the development of food products, and some of the innovations will pose much more serious regulatory questions than did the Calgene tomato. In fact, DNA Plant Technology and Monsanto have developed tomatoes with even longer shelf lives by modifying the gene which controls ethylene production, and other more elaborate bioengineered food products are on the horizon.406
During its review of the Calgene petitions, the FDA began formulating its policies for regulating biotechnology food products. In 1992, the agency published a policy statement on foods derived from new plant varieties.407 A rudimentary framework previously had been described in the mid-1980s,408 but the FDA's 1992 policy statement represented the first careful elaboration of its intended regulatory approach with regard to the application of biotechnology in the development of new plant varieties for use in food products. Notwithstanding the comprehensive nature of the policy statement, the FDA continues to refine it, and subsequently published notices have requested further public comment on separate aspects of the subject, including appropriate safety testing,409 and labeling.410
At the outset of the policy statement, the FDA notes that revolutionary genetic modification techniques, such as recombinant DNA and cell fusion, will lead to the development of new varieties of food plants that would not have been possible using more traditional methods of selective breeding.411 It also recognizes that, because "these techniques are more precise, they increase the potential for safe, better-characterized, and more predictable foods."412 In discussing genetic modification techniques to achieve desirable traits, the agency distinguished between agronomic characteristics of the plant (e.g., yield and resistance to pests or disease) and quality characteristics of the food (e.g., preservation, nutrition, and flavor).413
The policy statement canvasses a variety of potential safety issues posed by genetic modification—some shared with traditional selective breeding and some unique to the new methods—including unexpected chromosomal effects, changes in the levels of naturally occurring toxins in plants, reductions in nutrient bioavailability, production of new substances, introduction of allergens, and increased antibiotic resistance from marker genes.414 The FDA decided, however, that it need not adopt any special regulatory approach for foods derived from genetically modified plants, preferring instead to "utiliz[e] an approach identical in principle to that applied to foods developed by traditional plant breeding."415 The policy statement embraces a functional rather than literal approach to the food additive issue: the agency will "require food additive petitions in cases where safety questions exist sufficient to warrant formal premarket review by FDA to ensure public health protection."416 The FDA did not, however, explain how it would become aware of bioengineered foods for which it might require the submission of food additive petitions.
In short, according to the policy statement, the agency plans to rely on its enforcement authority under the general adulteration provisions to ensure the safety of whole foods derived
from genetically modified plants, though it pointed out that the transferred genetic material and the intended expression products could be subject to regulation as food additives if not GRAS.417 Because nucleic acids appear in the genetic material of all foods and have posed no safety problems in the past, the "FDA does not expect that there will be any serious question about the GRAS status of transferred genetic material."418
The intended expression product also generally will not alter the GRAS status of any food derived from the new plant variety, unless it is "a protein, carbohydrate, fat or oil, or other substance that differs significantly in structure, function, or composition from substances currently found in food."419 Thus, if transferred genetic material introduces an unusual protein and/or alters a metabolic pathway to produce a new carbohydrate or other substance, then the FDA may call for the submission of a food additive petition. Furthermore, apart from the possibility that a new component in a whole food might be regulated as a food additive, the whole food itself could be regulated as a food additive when used as a component of a processed food. The agency notes, however, that it rarely has passed on the presumptive GRAS status of a whole food when used as a component in a processed food because most food plants "have been widely recognized and accepted as safe."420
The policy statement provides detailed guidance for industry in conducting safety evaluations of new plant varieties and in determining the need for a food additive petition.421 The "FDA believes that a scientific basis should exist to establish that new plant varieties do not exhibit unacceptable effects with respect to toxicants, nutritional value or allergens."422 It also invites companies unsure about a new product's regulatory status to consult with the FDA on an informal basis rather than having to submit a petition requesting an advisory opinion.423 In materials distributed at a 1994 Food Advisory Committee meeting, the agency announced its intention to propose a regulation which would require formal premarket notification for all bioengineered foods, including the submission of safety and nutritional information to the FDA.424 Although the agency has not yet published such a proposal, a number of companies have consulted with the FDA to discuss the status of bioengineered food products under development.425
Finally, the FDA rejected suggestions that all bioengineered foods disclose their origin in labeling, but it would require disclosure if a risk of allergenicity has been introduced by the insertion of genetic material from another source.426 The labeling question proved quite contentious in the wake of the agency's approval of the new animal drug recombinant bovine somatotropin (rBST).427 Because it found no difference between milk from cows administered rBST and other milk, the FDA did not require any special disclosure statement in labeling, but it also declined to prohibit truthful and nonmisleading claims that a dairy product was derived from cows that had not received rBST.428 In hopes of avoiding similar controversy over the FLAVR SAVR™ tomato, Calgene plans to provide point-of-sale information for consumers.429
Comments from consumer and environmental groups criticized the regulatory framework described in the 1992 policy statement, arguing that the FDA instead should use its food additive authority to require premarket notification, premarket safety testing, and labeling disclosures for all bioengineered foods.430 In a recent report, the GAO elaborated on the premarket notification suggestion as follows:
In principle, such a system would bridge the gap between the absence of a requirement for premarket approval of whole foods and of substances that manufacturers and others have determined are GRAS and the requirement for extensive premarket review of new food additives. Several premarket notification systems have been proposed that range in formality and complexity from relatively simple letters from food companies to notify FDA about the introduction of a new product to mandatory FDA premarket reviews of data submitted by manufacturers to support the safety and/or GRAS status of their food product.431
For the moment, the agency continues to rely on a flexible approach to determine in each case whether to impose any special requirements.
Although flexibility has obvious advantages in dealing with an emerging technology, some observers have called for a clearer regulatory response to bioengineered food products. The FDA's 1992 policy statement represents a first step in that direction, notwithstanding its emphasis on informal, ad hoc review. Reliance on the food additive approval process would not, however, provide the added certainty about the safety criteria needed to address the special questions posed by bioengineered substances. As the GAO concluded, "controversies over FDA's policy on biotechnology are reviving old questions about the adequacy of the existing regulatory framework for food and food ingredients."432 In the future, as food biotechnology becomes more sophisticated, the agency's policies will continue to evolve.
Novel Macroingredients: Olestra
On January 30, 1996, the FDA approved Procter & Gamble's food additive petition for the noncaloric fat substitute olestra for use in certain snack foods.433 Olestra's history is remarkable in a number of respects. Nearly 25 years elapsed between the company's initial contacts with FDA officials and final approval, though there was no food additive petition pending during most of that interval, and Procter & Gamble spent more than $200 million in the process.434 Olestra represents one of the first macroingredients approved as a food additive,435 and the final regulation conditions its use on special labeling, vitamin fortification, and the submission of postmarketing surveillance reports to allow for further agency review.436 No doubt the company will seek approval of supplemental petitions for additional uses in the future.
In 1991, anticipating that the FDA would not complete its regulatory review before the expiration of the 17-year term on the various patents covering olestra, Procter & Gamble petitioned Congress for special legislation to extend several of its patents well beyond 1994.437 The resulting hearings provided a detailed history of the company's R&D efforts.438
In 1969, company scientists synthesized olestra, a molecule with a sucrose core surrounded by six, seven, or eight medium- or long-chain fatty acids (classified as a sucrose polyester).439 Triglycerides, by contrast, have a glycerol core surrounded by three fatty acids. Olestra tastes, smells, and cooks like ordinary fats, and it can absorb fat-soluble substances. The human GI tract cannot, however, absorb olestra because it lacks enzymes capable of breaking it down. Because it passes through the intestines undigested, olestra provides no calories, but it also may interfere with the absorption of substances such as fat-soluble vitamins or certain lipophilic drugs,440 and it may cause unpleasant gastrointestinal side effects in some individuals. 441
After initially pursuing approval of the substance as a human drug, 442 the company filed its food additive petition in 1987.443 The precise explanation for the tardy decision to seek food additive approval remains unclear,444 but Procter & Gamble did encounter some delays while the FDA worked with it to define appropriate testing protocols for a substance intended for use at a fairly high level in the diet.445 In 1990, the FDA convinced Procter & Gamble to amend the petition to request approval of only a limited use in snack foods and also asked the company to undertake another long-term animal study.446
In reviewing the olestra petition, the FDA drew upon special expertise from other centers and federal agencies as well as external scientific consultants, and it assembled a six-member Regulatory Decision Team of senior agency managers to resolve issues that arose during the petition review process.447 In November 1995, after the agency convened a meeting of a special working group of the Food Advisory Committee to discuss olestra's safety,448 a majority of the full Food Advisory Committee concluded that there was a reasonable certainty that no harm
would result from the use of olestra in snack foods but also recommended the imposition of certain conditions on any approval.449 The FDA announced a final deadline for the submission of public comments in response to the food additive petition,450 which it then approved less than two months after the close of the comment period.451 The Center for Science in the Public Interest (CSPI) filed objections to the decision and requested a hearing but not a stay of the approval. 452 In the meantime, test marketing has begun in several cities.453
Shortly after finally approving olestra, the agency invited comments on the safety assessment of novel macroingredients,454 and scientists in and out of the FDA have begun to consider the subject more closely. 455 Because olestra is virtually not absorbed, traditional toxicology studies were less important than efforts to assess various digestive effects in humans.456 In addition, because of the relatively high consumption levels of macroingredients, it is extremely difficult to conduct animal studies at the exaggerated intake levels necessary in chronic feeding tests,457 leading some scientists to recommend greater reliance on postmarketing evaluations of such additives.458
In approving olestra, the FDA evaluated a wealth of scientific data. 459 In doing so, it emphasized the irrelevance in its safety determination of either concerns about unpleasant (but not otherwise harmful) side effects or the social benefits of reducing levels of dietary fat. 460 According to the GAO's chief investigator on olestra, the ''FDA feels that they really learned a lot by going through this process.'' 461 Ultimately, the FDA concluded that the use of olestra in certain snack foods was safe, but it imposed several unusual conditions. First, to compensate for the projected interference with the absorption of certain fat-soluble micronutrients, foods containing olestra must be fortified with precise quantities of vitamins A, D, E, and K.462
Second, the labels of snack foods containing olestra must bear the following boxed disclosure statement:
This Product Contains Olestra. Olestra may cause abdominal cramping and loose stools. Olestra inhibits the absorption of some vitamins and other nutrients. Vitamins A, D, E, and K have been added.463
Unlike most of the other labeling requirements appearing in food additive regulations, which call for warnings or ingredient disclosures to alert consumers who may be allergic to the substance, the FDA concedes that this statement does not represent a condition necessary to ensure the safe use of olestra.464 Instead, it argues that the statute's general misbranding prohibitions,465 as cross-referenced in Section 409 as a ground for denying a food additive petition,466 authorize the imposition of such a condition in this case.467 The agency did, however, invite further comments on whether it should revise what it characterized as only an "interim" labeling requirement. 468 Procter & Gamble filed comments urging a less explicit labeling requirement, while at the same time CSPI sought a more conspicuous warning.469
The requirement for postmarketing surveillance represents one of the most interesting features of the approval. The regulation itself does not mandate further testing by the petitioner; it only provides that the FDA "will review and evaluate all data and information bearing on the safety of olestra received by the agency."470 In the preamble, however, the agency explains that "as a condition of approval, Procter and Gamble is to conduct the studies that it has identified in its letter to FDA [dated January 24, 1996], consistent with the timetables identified in that letter,"471 and it warned that, "if Procter and Gamble does not conduct the identified studies and does not conduct them according to the articulated timetable, FDA will consider the approval . . . to be void ab initio and will institute appropriate proceedings."472
This threat is remarkable insofar as it treats the food additive approval as a private license rather than a public regulation available, subject only to patent limitations, to any firm wishing to
manufacture and sell the additive.473 The FDA did not purport to approve olestra as an interim food additive. The agency added that such a condition "is not without precedent," citing the more limited data collection requirement imposed on the manufacturer of aspartame, 474 and it also emphasized that the imposition of this condition "is not, and should not be interpreted as, an indication that FDA has somehow not determined that there is a reasonable certainty that no harm will result from the use of olestra in savory snacks."475 Six months after approval, Procter & Gamble reported fewer consumer complaints than expected.476 In the meantime, the FDA must review and eventually respond to CSPI's objections and hearing request.
The Composite Picture
Aggregate data provide a different and equally valuable perspective on the FDA's experience with the food additive approval process. Lengthy reviews of petitions for novel macroingredients may be unavoidable, but quantitative measures suggest habitual administrative delays for both simple and complex submissions. During the last decade, direct food additive petitions (original and supplemental) and GRAS affirmation petitions have languished at the agency for an average of more than five years. As more fully explained below, a number of factors may account for these delays, especially growing resource constraints faced by the FDA,477 and an associated variety of reforms have been proposed.
Track Records Compared
As explained previously, the FD&C Act establishes a maximum 180-day review clock for food additive petitions, but, pursuant to FDA regulations, this clock runs only intermittently after filing—the agency stops the clock while awaiting additional information from the petitioner and resets it at zero whenever substantial new information arrives. 478 No deadlines exist for action on GRAS affirmation petitions, and the FDA has been even slower in reviewing these submissions.479
In fact, the FDA has approved very few direct food additives during the last quarter of a century. In addition to aspartame and olestra, these include the antioxidant TBHQ,480 the multipurpose additive polydextrose,481 the sweetener acesulfame potassium,482 and the stabilizer and thickener gellan gum.483 The bulk of the agency's food additive review caseload has involved petitions for indirect food additives. 484 In addition, as happened most notably in the case of aspartame, the FDA has processed a number of supplemental petitions requesting amendments to authorize additional uses of previously approved additives. 485 Ironically, the agency occasionally takes more time to approve these supplemental petitions than the original. For instance, the original petition for polydextrose took two years to approve, but supplemental petitions for expanded uses required up to six years. 486
In the early 1990s, the FDA's Center for Food Safety and Applied Nutrition (CFSAN) undertook a series of internal management studies, including an evaluation of the food additive review process using data from 1979 through 1992.487 In the case of direct food additive petitions (both original and supplemental) approved during this period, the agency found a bimodal distribution, with review times ranging from one to three years and three to six years.488 Only three of the ninety petitions approved during this period required less than six months to review and six of the petitions required more than six years.489 Of the 42 substances affirmed as GRAS between 1979 and 1992, none took less than one year to review and 26 took more than four years.490
In 1995, Congress for the first time held oversight hearings concerning the FDA's food additive approval process.491 As summarized in the committee report issued several months after the completion of these hearings:
FDA has a backlog of 295 food additive petitions under review, some of which have been pending since the 1970s. Approximately 100 new food and color additive petitions are submitted to the FDA each year. . . . Direct additives comprise approximately 17% of pending petitions and may be under review for up to 10 years . . . . Seventy-five GRAS affirmation petitions are currently pending at FDA, the oldest of which was filed on August 31, 1972.492
Several witnesses provided graphical representations of the flow of the food additive approvals and GRAS affirmation petitions, and some of the more relevant graphs have been reproduced in the accompanying pages. For instance, over the course of a decade, review times for both food additive and GRAS affirmation petitions had doubled.493 FDA officials conceded that significant delays existed and announced various initiatives to speed reviews,494 but they also testified that the statute's 180-day deadline was totally unrealistic.495 Certain members of the subcommittee expressed dismay with the agency's position, recognizing that the deadline may be too short but disagreeing that the FDA could, therefore, simply choose to ignore it.496
In addition, agency officials explained that companies bore some of the blame for the delays in the process because of inadequacies in their submissions.497 Although it could reject such petitions without prejudice to a resubmission in the future, the FDA generally has worked to help petitioners correct any deficiencies in a pending petition.498 Companies apparently have tolerated lengthy delays as preferable to the stigma associated with an FDA rejection of a food additive petition.499 Apart from the unfavorable (and somewhat misleading) impact of this policy on average review times, one witness objected to what he regarded as the FDA's goal of eventually approving all additives unless patently unsafe.500 CFSAN officials subsequently advised industry that they plan to start rejecting incomplete petitions. 501
Some of the subcommittee members commented on the unwillingness of officials from individual companies to testify,502 and one witness suggested that their reluctance might reflect fear of retaliation for criticizing the FDA.503 Even so, representatives from several trade associations appeared at the hearings. They and others commented on what they perceived to be the FDA's significant risk aversion, complaining that reviewers evidently preferred to ask additional questions rather than exercise well-informed scientific judgment and recommend closure on a petition.504 Whatever the precise explanation, however, no one seriously disputes the fact that reviews of new food additives have become extremely slow.
Because so few petitions are approved in any single year,505 aggregate measures for the last decade are more valuable than year-to-year comparisons, especially reflecting the average review times for (1) first approvals of direct food additives, (2) approvals of supplemental petitions for additional uses of a direct additive (sometimes referred to as follow-on petitions), and (3) affirmations of GRAS petitions for direct ingredients.
Thus, for direct food additives approved during the last decade, an average of more than six years elapsed between filing and first approval. Not surprisingly, the FDA approved follow-on uses for previously approved additives more quickly (in just over three years), though still long past the three- and six-month deadlines imposed by the statute. Finally, GRAS affirmation petitions approved during the last decade required more than seven years on average, by far the slowest review process for substances added directly to food products.
It is difficult, of course, to measure these review periods with any real precision. For example, some significant but unreported portions of time are consumed while the FDA awaits submission of additional information by the petitioner.509 On the other hand, because of substantial delays in publishing notices of filing after the receipt of a petition, average review times may be understated if calculated from the notice of filing.510 Although FDA officials have pointed to trend data as evidence of recent improvements,511 the numbers for any particular year are too limited to justify such a conclusion and, as the backlogs of long-pending petitions are cleared, the average review times initially should worsen rather than improve.
After Congress subjected drugs to rigorous premarket FDA review for safety and effectiveness, some critics argued that the resulting delays in approval kept valuable therapeutic agents out of the hands of needy patients.512 More than 30y years later, complaints about a "drug lag" continue.513 Although similar delays in approvals of direct food additives may seem less consequential,514 access to innovations in food technology does have some public health dimensions.515 Wholly apart from delays in reviewing particular applications, the inevitability of substantial delays and confusion about applicable regulatory controls may inhibit research and development efforts.516 The lengthy delays in FDA reviews of GRAS affirmation petitions pose fewer such concerns because, for the most part, petitioners proceed to market once the agency initially accepts their submissions and seem to care little about how much time may elapse before formal affirmation.
A Catalogue of Proposed Solutions
A number of factors may account for the delays observed in the agency's review of food additives, and a corresponding range of reforms has been suggested. Some of the solutions would require fairly modest revisions in the FDA's internal procedures, but other proposals call for radical overhauls of the current system. This section summarizes the most salient flaws in the system and suggested reforms.
Internal Management Initiatives
The FDA recently reorganized CFSAN, in part as an attempt to streamline its food additive review activities,517 and it announced renewed efforts at publishing guidelines and tracking work flow among reviewers. 518 To be sure, the lack of coordination and frequent turnover among
reviewers has created significant inefficiencies in the past.519 Moreover, clearer guidelines and other outreach efforts such as pre-filing meetings with petitioners may help minimize reported problems with the submission of incomplete food additive petitions.
Such initiatives promise only modest improvements in efficiency,520 however, and there are no guarantees that the FDA will make good on its assurances of streamlining the process.521 Although they would represent dramatic improvements over the existing time frames, none of the announced performance goals—to be phased in gradually only after the agency begins reducing the current backlog—would meet the statutory deadlines.522 One recent bill would have codified performance goals for the FDA,523 but Congress failed to enact the proposed reforms. In any event, the agency will be free to ignore such goals unless some action-forcing mechanism accompanies any new deadlines.
Some critics have suggested the establishment of a statutory "hammer" to force the FDA to act by a specified deadline.524 A type of hammer appeared in some of the early food additive bills,525 and this mechanism is currently in vogue to prompt the timely promulgation of broad regulations.526 But such a tool seems ill-suited to the task of forcing expeditious action on food additive petitions.
Instead of tolling and ignoring the statutory clock as it does now, the FDA could circumvent a statutory hammer simply by issuing repeated denials of a petition until it was satisfied about the safety of a food additive.527 Such a response might be desirable insofar as it encourages greater initial care in compiling petitions. The sole benefit for industry would be a much earlier opportunity to seek judicial review, but this opportunity would provide only an illusory advantage over the present system—the petitioner still would shoulder the burden of proof, and courts accord significant deference to the FDA's scientific judgments.528 If it could not circumvent a hammer, the agency would in theory be forced to approve food additives about which it may harbor some lingering safety concerns. The FDA might then routinely insist on the submission of postmarketing reports in such instances, as a backstop against a rushed approval decision. Even if no serious problems materialize, future backlogs may develop as FDA scientists have to spend their time reviewing these periodic reports. Statutory hammers might have some place in a revised food additive approval procedure, but, as a freestanding mechanism for reform, they suffer from significant shortcomings.
Resource constraints provide one obvious explanation for the delays in FDA reviews of food additive and GRAS affirmation petitions. CFSAN staffing levels have remained constant even while food additive petitions have become more complicated and its other regulatory responsibilities have burgeoned.529 The agency could improve utilization of these resources by better prioritizing its review activities. For instance, reducing the intensity of reviews for substances likely to pose little or no public health risk, such as indirect food additives or additional uses of direct additives, would allow reviewers to focus on the more significant petitions involving novel ingredients.
The FDA recently exempted some indirect food additives from full review if estimated human exposure falls below a "threshold of regulation," 530 thereby allowing CFSAN to allocate more of its resources to direct additives.531 Pursuant to this new regulation, the FDA may issue a letter exempting a food-contact substance from food additive petition requirements if a request
for an exemption demonstrates, through migration data and any existing toxicological studies, that the likelihood or extent of migration to food is so trivial as not to require further regulation.532 Although it recognized the need for timely reviews of exemption requests, the FDA declined to impose any deadlines on itself.533 The resulting potential for internal delays may limit the policy's value as a mechanism for redirecting agency resources to nontrivial additives.534
By requiring FDA action on exemption requests,535 the threshold of regulation policy is more like an abbreviated premarket review procedure than a premarket notification system. Congress recently enacted a simplified premarket notification procedure for "new dietary ingredients," 536 though as part of an effort to deregulate the dietary supplement industry rather than to free up resources for the FDA's direct food additive reviews. Over the last 20 years, the premarket notification system for medical devices—designed by Congress as a temporary procedure for exempting "substantially equivalent" devices from full premarket approval requirements—has evolved to become the FDA's primary review mechanism and, as a result, also more cumbersome than the simple notification procedure originally envisioned by Congress.537 Unless reoriented, a similar fate may befall the new threshold of regulation policy or similar premarket notification reforms.
In addition to exempting inconsequential food-contact substances, the agency might use an abbreviated review process for supplemental petitions seeking approval for additional uses of a previously approved additive.538 The FDA also might assign higher priority to the review of novel food additives which promise important public health benefits. At present, the agency reviews submissions in the order that they are received, and existing backlogs may mean that scientific reviews of a newly filed petition do not even begin for a lengthy period of time.539 Although the statute apparently prohibits any consideration of benefits in making safety determinations,540 the FDA could prioritize its reviews on the basis of such criteria.541 Similar prioritization systems exist for new drug reviews, though the statute expressly authorizes the consideration of benefits in the drug approval process. 542 Instead, however, the FDA's announced performance goals—varying for three tiers of petitions based on their length, complexity, or novelty of the issues—promise relatively faster reviews of petitions for the least complex and novel additives.543
Imposing User Fees
Some have recommended legislative authorization of user fees to address the resource problem,544 as Congress recently did in the prescription drug area.545 Indeed, the FDA specifically requested user fees during congressional hearings which eventually led to the enactment of the Food Additives Amendment.546 Such a provision had been included in the Pesticide Residues Amendment of 1954,547 and it also appeared in Congressman Delaney's food additive bills.548 The 1958 legislation did not, however, provide the FDA with the authority to impose such fees.
By definition, user fees are intended to reimburse agencies for the particular benefits they provide to regulated entities. More than two decades ago, the Supreme Court held that the imposition of user fees on regulated companies for the benefit of the general public is nothing more than another tax on industry.549 As explained previously, a food additive petition directly benefits competitors as well as the petitioner. Unlike a new drug application, which gives the successful applicant an exclusive private license to sell a drug product, a food additive petition results in the publication by the FDA of a regulation authorizing use of the additive by any person wishing to do so, subject only to any remaining patent term.550
If only the petitioner were required to pay for an FDA service that inures to the benefit of the entire food industry, developers of innovative additives would be unfairly penalized.551 Congress could avoid this concern by converting food additive approvals into more of a private licensing system like that used for new drugs. In fact, in addition to attacking the causes of existing delays (whether by imposing user fees or otherwise), some have emphasized the need to soften the negative competitive effects of such delays by ensuring a period of market exclusivity to the petitioner in the event that patents are unavailable or about to expire.552 Even then, however, the level of user fees would have to be far lower than currently charged for new drugs in recognition of the generally smaller R&D investment and limited commercial potentional of most new food additives. Otherwise, the imposition of user fees would further discourage companies from pursuing FDA approval.
Frustration with delays in the FDA's drug approval process has triggered calls for extramural reviews by scientific organizations.553 A similar approach has been suggested as a way of expediting food additive reviews. The FDA frequently involves scientific organizations in an advisory capacity, for instance, when it sought NAS and FASEB input during the GRAS list review,554 and, in 1995, the agency announced plans to contract out reviews of routine toxicity studies in currently pending food additive petitions.555
Many academic scientists and industry representatives favor greater reliance on such outside organizations.556Under one suggested model, a petitioner willing to pay for it could ask the FDA to contract out the review of its safety data, and both the FDA and the third-party organization would have a limited opportunity to request that the petitioner first collect additional information. A number of issues would have to be resolved, including how to comply with or modify the Federal Advisory Committee Act (FACA),557 whether fees should be channeled through the agency,558 and what force to give the third party's recommendation. Some have suggested that the recommendations of an outside organization should be deemed adopted by the FDA unless it could demonstrate (within a limited period of time after the issuance of a report) that the petitioner had failed to establish the safety of a substance.559
In a related vein, some have suggested that the agency place greater reliance on the GRAS exception, though not necessarily through the formalized GRAS affirmation process that exists at present.560 Perhaps this could be accomplished by contracting out the review of GRAS petitions, much as the FDA did in the 1970s when it contracted out data reviews for substances appearing on the initial GRAS lists,561 or by giving presumptive weight to the conclusions of certain organizations when commissioned by the petitioner to review the GRAS status of a food substance.562 The agency recently proposed a simplified GRAS notification procedure to replace the affirmation petition procedure in hopes of redirecting its resources to higher priorities.563
Efforts at constructive revisions of the statute and regulations should continue, utilizing one or more of the reforms discussed above. In the meantime, because of the breadth of the GRAS exception, industry already has the unusual opportunity to reform the food additive approval process by itself and substantially reduce existing delays. After all, GRAS is a definitional exception (or exclusion) rather than an exemption procedure, and self-determinations premised on published extramural reviews by reputable scientific organizations could almost totally
supplant the industry's existing dependence on the FDA. For instance, companies can and have commissioned extramural reviews by FASEB and then submitted a GRAS affirmation petition primarily to advise the FDA of their self-determination.564 This approach could become routine, requiring not a change in the law but in the industry's attitude.
Although food processors traditionally have insisted on some sort of FDA seal of approval, chemical manufacturers need to better explain to their customers that such approvals are neither necessary, rapidly acquired, nor terribly meaningful in some cases.565 Why should processors feel more comfortable using an interim food additive such as mannitol (which theoretically could be withdrawn by the FDA with little or no advance warning566) than a substance that FASEB or FEMA found to be safe under the agency's GRAS criteria? No doubt the food processors fear the possibility of a seizure or injunction by the FDA and a resulting loss of public confidence. But the agency's past actions against unapproved food-use substances all have involved in-house and poorly documented GRAS self-determinations rather than those based on published extramural reviews.
In the highly unlikely event that the FDA initiated enforcement proceedings or sent a warning letter to a food processor, the additive's manufacturer could respond and perhaps even file a declaratory judgment action in federal court to settle the GRAS question. The FDA also could take a less adversarial approach and initiate—perhaps at the behest of a public interest organization—a rulemaking proceeding that might eventually require a food additive petition for the substance, but the agency also might do that with a substance previously affirmed as GRAS. If an additive manufacturer has assembled good safety data and chooses to present these data to a qualified and independent panel of experts (or regulatory authorities in another country), the GRAS exception currently allows the manufacturer to do so instead of filing a food additive petition with the FDA. If food processors would come to understand the limited practical value of an FDA affirmation or approval, a market for private certifications could emerge without any changes to the FD&C Act or the agency's regulations.567
Unless the agency chooses to obstruct such a change because it fears a loss of power or control, it also could reap substantial benefits from the development.568 In fact, the FDA could facilitate this process by issuing guidelines for ensuring the independence of third-party organizations.569 Of course, if FASEB or another organization concludes that a substance is not GRAS for a particular use on the basis of the evidence presented, then a company could attempt to "appeal" this conclusion by filing a food additive petition with the FDA, perhaps after first undertaking additional studies. Moreover, Congress could decide to encourage manufacturers of innovative but unpatentable substances to seek food additive approval by providing them with some form of market exclusivity.
Public interest groups surely will object to any loss of FDA control over the food additive approval process.570 In some respects, privatization would bring us full circle, back to the time when food companies assured themselves of the safety of any substance used in food, and the FDA shouldered the burden of proving otherwise.571 But now another expert body would review and publish the industry's judgments, applying more uniform and stringent safety assessment criteria than had existed prior to 1958.572
The industry frequently calls for the repeal of the Delaney clause. Even if otherwise sensible, such a reform probably would have little impact on the pace of reviews—petitioners would still have to conduct chronic toxicity studies in appropriate cases, and the FDA would apply the general safety standard conservatively even without the added strictures of the Delaney clause. Although some critics have complained of excessive risk aversion by the agency, no one seriously suggests that Congress should water down the general safety standard. Reforms in the food additive approval process will come from changes in definitions and procedures (and attitudes), not from modifications in the substantive approval criteria.
Combatting Sham Petitioning
Competitive manipulation of the FDA's procedures provides another possible explanation for delays, at least in a few cases, and deserves separate consideration. The food additive approval procedures may be vulnerable to "sham" petitioning, defined as the improper use of administrative procedures by firms seeking to delay or prevent entry into the market by a would be competitor.573 As I recently have concluded elsewhere, "real opportunities exist for sham petitioning in administrative proceedings, especially when market entry requires some sort of agency licensing as in the pharmaceutical, transportation, communications, and energy industries."574 The same may be said of the food additive industry.
Upon the acceptance or refusal of a food additive petition for filing, virtually the entire petition becomes available for public disclosure. 575 The petition thus becomes an easy target for comments from interested persons, both from well-intentioned scientists and public interest groups and from competitors motivated chiefly by a desire to delay approval of the petitioner's product. In recent cases, taking advantage of the FDA's de facto comment period,576 parties have made anonymous submissions late in the agency's internal review process apparently to create unwarranted delays in the issuance of food additive regulations.
For example, the FDA published a notice of filing for the sucralose petition almost 10 years ago.577 Three years later, it sent the company a draft of the Federal Register order approving the petition, but, later that same year, a law firm submitted comments on behalf of an unnamed client raising concerns about the safety of sucralose. After a lengthy interruption in the FDA's internal processing of the final regulation, a revised draft of the final regulation reached the Office of General Counsel for approval, but the same law firm submitted a letter to the agency reiterating its original concerns, and thereafter it again wrote to the FDA purporting to raise new questions about sucralose. More than a year later, CSPI sent the FDA unsolicited information it had received from an anonymous source concerning an early draft of one of the company's studies.578 Each of these submissions apparently interrupted the approval process. In the meantime, regulatory authorities in Canada and Australia have approved sucralose.579
The bizarre chronology of events surrounding the sucralose petition may reflect bad faith intervention by third parties, possibly for purposes of retaining a competitive advantage in the artificial sweeteners market. Extremely tardy comments were submitted with exquisite timing, apparently calculated to yield the greatest disruption of the FDA's internal processing of the food additive regulation and to delay eventual approval.580 Even if this interpretation of events is unwarranted, 581 the experience with sucralose suggests that the food additive petition process may be susceptible to such abuse, and FDA officials have conceded that this may represent a problem.582 Comments submitted several years after a notice of filing is published, and just as the agency appears ready to approve a food additive petition, can only serve to needlessly delay approval of important and legitimate new products.
Unfettered opportunities for prepublication comment may cause excessive delays which are contrary to the purpose of the act.583 Courts will order agency action when delays become unreasonable.584 Congress imposed a strict deadline for agency action in response to food additive petitions, and FDA delays in responding to food additive petitions have been subject to judicial scrutiny in at least two instances. In a challenge to the FDA's failure to act promptly on a food additive petition for cyclamate, the court allowed discovery to proceed because it found sufficient merit in "Plaintiffs claim that its petition is being studied into oblivion."585 In another case, involving the additive gentian violet used in premixed medicated feed for livestock, a court found that the FDA's "failure and refusal to act upon the [food additive] petition . . . constitute[d] a clear denial of the rights of Plaintiffs [which] effectively takes their property without due process."586
The FDA's practice of accepting and responding to comments whenever submitted before publication of a food additive order contravenes an explicit statutory deadline requiring that the agency issue a final order within 90 days of filing, up to a maximum of 180 days if necessary to fully study and investigate the petition.587 In practice, this deadline may be impossible for the agency to comply with, particularly when submissions include large amounts of scientific data. At a minimum, however, the deadline prescribed by Congress requires that the agency not entertain or respond to comments received after the statutory 90-day deadline has passed.588 The FDA itself has recognized the importance of placing some strict limits on the time for filing comments and other responses. For example, in setting an explicit comment deadline with respect to olestra, the agency explained that, "[a]bsent such boundaries, it will be difficult for FDA to reach a decision because the underlying data set could be shifting continuously."589 In addition, the FDA's revised procedural regulations for the review of OTC drug products provide strict time limits for comments submitted in response to a proposed monograph.590
If the FDA plans to accept and respond to comments before publication, it should place a strict limit on the time for filing comments. Otherwise it could never comply with the 90-(or 180-)day deadline contained in the statute; if the FDA were to adopt a regulation allowing a comment period on food additive petitions exceeding 180 days, the agency's action would be patently unlawful.591 A 60-day prepublication comment period for food additive petitions would provide a reasonable time frame. For instance, the FDA's Center for Veterinary Medicine, acting pursuant to identical statutory and regulatory authority, often includes an explicit 60-day comment period in the notice of filing.592
Late filed comments addressing a food additive petition would not be ignored by the agency. In the unlikely event that tardy comments happen to suggest innovative interpretations of existing safety data that escaped the FDA's attention prior to approval, these issues can be raised in objections to the final regulation.593 This is the procedure used by the agency in its OTC Drug Review. Any comments received after the 90-day comment period following publication of a proposed monograph are deferred by the agency until after a tentative final monograph (TFM) is published and interested persons are given an opportunity to file objections and request a public hearing.594 New data and information submitted after this period but before the final monograph is published would be treated as a petition to amend the monograph.595 New information discovered after a food additive regulation becomes final is handled in essentially this fashion.596
The proposed 60-day comment period would not prevent the FDA from considering important new safety data submitted any time prior to the publication of the food additive order. The agency could create conditions for the prepublication consideration of such late-filed information to ensure against submissions that contain no new safety data but are interposed for the purpose of delay. Certification requirements could ensure that the food additive petition process is not subject to abuse and unreasonable delay while at the same time retaining sufficient flexibility to take legitimate new information into account no matter when it is submitted.597 The FDA already employs certification requirements in connection with other submissions such as citizen petitions.598
Almost four decades after the enactment of the Food Additives Amendment and 20 years after the political controversy over saccharin, public attention again has turned to the regulation of food additives in the United States. Food additive petitioners have become frustrated with lengthy delays in the review process, while other interested parties engage in obstructionist
tactics and then vocally criticize approved additives as unsafe. The FDA is caught between these competing factions, struggling to do more with fewer resources. Out of necessity, the agency has been forced to improvise, sometimes ignoring its unrealistic statutory directives.
The GRAS exception has functioned tolerably well and created an important route to market, and the food industry could make even broader use of this definitional exclusion. For its part, the FDA could facilitate GRAS self-determinations, continue efforts at streamlining the review process, exempt certain types of substances altogether to allow a reallocation of its scarce resources, and revise its procedures to minimize the risk of sham petitioning. Finally, Congress might require the agency to implement one or more of these steps, authorize the imposition of user fees in exchange for some period of market exclusivity, and codify procedures for the extramural reviews of GRAS affirmation and food additive petitions.
Although controversy inevitably will surround the choice among these and other reforms, the FDA cannot simply continue ''muddling along'' in its current fashion. Unless someone fixes the food additive approval process, existing regulatory hurdles will inhibit future scientific advances in food technology. Generally, the standards for safety assessments should remain in place, but the time may have come for an overhaul of the mechanisms used to ensure the safety of substances added to food.
Appendix: International Comparisons
Different countries' approaches to food additive regulation often are characterized as either "positive" or "negative" listing systems. In a positive list system, only those additives identified as permissible by the designated regulatory authority may be used in food. In a negative list system, any additive may be used so long as it is not specifically restricted by the designated regulatory authority, which normally shoulders the burden to prove that a substance is unsafe. The Food Additives Amendment represented a shift in the United States from a negative to a positive list system. Actually, however, because of its special definition of the term food additive, the 1958 amendment established a "mixed" system, imposing a positive listing requirement for food additives while subjecting GRAS food substances to something that resembles a negative listing approach.599 Many other nations also have hybrid systems, though their particular mixes differ depending in part on how the term food additive is defined.600
The following sections briefly sketch out the regulation of food additives in several industrialized countries, identifying relevant differences in approach. Although the FDA participates in international harmonization efforts, notably with regard to acceptable designs for the testing of pharmaceuticals,601 reciprocal recognition of foreign approvals remains unlikely. Some argue that the United States encounters the difficult regulatory problems before other countries do.602 Others have suggested that excessive FDA delays drive companies to seek initial regulatory approvals in foreign countries.603 Unfortunately, comparative data on review times are not readily available,604 though anecdotal evidence discussed herein suggests that food-use substances often enter foreign markets more rapidly.
Under the Canadian Food and Drugs Act,605 the Health Protection Branch (HPB) of the department of Health controls the safety of food products sold in the country.606 The statute prohibits the sale of food that "has in or on it any poisonous or harmful substance," and it delegates broad authority to the Department to issue regulations governing "the use of any substance as an ingredient in any food."607
Pursuant to the HPB's implementing regulations, originally promulgated in 1964, the term "food additive" means "any substance the use of which results, or may reasonably be expected to result, in it or its by-products becoming a part of or affecting the characteristics of a food."608 Like the Food Additives Amendment of 1958, the definition excludes agricultural chemicals (such as pesticides) and animal drugs, but it also excludes the following: various flavors; vitamins, mineral nutrients, and amino acids; and food packaging materials.609 Most of the substances excluded from the definition are governed by separate sections of the regulations. Although indirect additives in food packaging are not subject to a premarket review requirement in Canada, an informal (and expeditious) system has developed to respond to industry requests for letters of opinion.610
Although the HPB regulations do not include a GRAS or prior sanction exception in the definition of food additive, they do exclude "any nutritive material that is used, recognized or commonly sold as an article or ingredient of food."611 The department recently proposed regulations to require that a company file a premarket notification along with safety data 90 days before introducing for sale a "novel" food or a food produced through a novel process, with novelty defined as a lack of significant previous use in Canada.612
Food products sold in Canada may contain only those food additives appearing on a list of approved substances.613 In order to add a substance to the list, a company would have to submit an application to the Department of Health which includes, inter alia, "data establishing that the food additive will have the intended physical or other technical effect," and "detailed reports of tests made to establish the safety of the food additive under the conditions of use recommended."614 Within 90 days after the filing of such a submission, the department shall advise the applicant whether, and under what terms, it intends to recommend listing.615 The HPB approves an additive if it is safe, does not lead to deception, and offers some benefit to consumers. 616
The United Kingdom
In the past, some have criticized the regulation of food additives in the United Kingdom as weak.617 The relevant statute prohibits adding "any article or substance to the food, [or] using any article or substance as an ingredient in the preparation of the food," so as to render the food "injurious to health," and it authorizes the issuance of regulations governing the use of substances in food.618 Two agencies, the Ministry of Agriculture, Fisheries and Food (MAFF) and the Department of Health and Social Security (DHSS), share responsibility for implementation of the act.
The MAFF has issued lists of permitted additives, subdivided into roughly two dozen functional categories, but certain other categories (including flavorings, starches, and enzymes) currently are not subject to positive lists.619 The MAFF recently modified its food additive lists—for instance, by authorizing the use of cyclamate—as required under the latest directives issued by the Council of the European Union (EU) and its predecessor, the European Economic Community (EEC). 620
In order to use an additive falling within a regulated category but not appearing on an existing list, a firm would have to seek permission by submitting safety and utility information to the MAFF. Evidence of utility must be provided because, "no matter how safe an additive might be, it is policy to prohibit compounds unless they are, in some sense, needed."621 The ministry refers any request and accompanying information to its Food Advisory Committee and the DHSS's Committee on Toxicity, and both committees report back a recommendation to the
MAFF, which, after inviting public comment on the reports, may then publish a draft regulation for further comment and eventual finalization. 622
The European Union
EU policy on foodstuffs directly affects domestic food additive regulation in the United Kingdom and much of the rest of Europe.623 The free movement of goods among member states is promoted both through the principle of "mutual recognition" and efforts at the "approximation" (harmonization) of national laws. Because of difficulties and delays in harmonizing product standards, the EEC's largely self-executing mutual recognition principles initially provided an important mechanism for increased trade.624 As explicated by the European Court of Justice in a series of cases involving food products, a member state may not prohibit the sale of food containing an unapproved additive which is authorized in another state unless: (1) an importer is given an opportunity to seek approval of the additive through an authorization procedure that will be completed in a reasonable time and subject to judicial review, and (2) the failure to approve the additive is justified on public health grounds or lack of utility.625
The harmonization of food additive regulations in the EU recently has accelerated, making the need for mutual recognition among inconsistent domestic regulations less relevant. A framework directive adopted in 1989 defines the term "food additive" as follows:
[A]ny substance not normally consumed as a food in itself and not normally used as a characteristic ingredient of food whether or not it has nutritive value, the intentional addition of which to food for a technological purpose in the manufacture, processing, preparation, treatment, packaging, transport or storage of such food results, or may be reasonably expected to result, in it or its by-products becoming directly or indirectly a component of such foods.626
The definition specifically excludes processing aids, agricultural chemicals, flavoring substances (which are the subject of a separate framework directive), and added nutrients such as vitamins and minerals. 627 The definition lacks a GRAS exception, though it does exclude "characteristic ingredient[s]" of food and, as in Canada, the recently proposed regulations on novel food and food ingredients incorporate such a concept.628
The 1989 framework directive sets forth general criteria for the use of food additives. The commission will approve a food additive only under the following circumstances:
[T]here can be demonstrated a reasonable technological need and the purpose cannot be achieved by other means which are economically and technologically practicable; they present no hazard to the health of the consumer at the level of use proposed, so far as can be judged by the scientific evidence available; [and] they do not mislead the consumer.629
Unlike the limitations on the FDA's ability to consider utility, the EU demands evidence that "the proposed use of the additive would have demonstrable advantages of benefit to the consumer."630 Only then will questions of safety be considered, for which an additive "must be subjected to appropriate toxicological testing and evaluation." 631
After the establishment of a list for a particular category of food additives, only those substances included on the list may be used in food.632 In 1962, the EEC issued its first positive list to regulate additives throughout the European Community, reducing the number of color additives for use in food previously permitted by member states.633 Over the course of the next decade, the EEC issued similar directives applicable to such ingredients as preservatives, antioxidants, and emulsifiers.634 These lists subsequently were amended on several occasions, most recently in 1994 through the issuance of comprehensive new lists covering colors and
sweeteners,635 and a 1995 consolidated list of miscellaneous other additives (including over 250 individual substances for preservative and numerous other uses) which are not otherwise the subject of a final or pending list.636 In order to become operative, these positive lists must be transposed into the national laws of each member state, as accomplished in the United Kingdom through the MAFF's regulatory amendments late in 1995.637
The EU continues to work on directives for the regulation of food packaging638 and flavoring substances.639 In addition, it appears ready to issue regulations requiring premarket approval of "novel" foods and ingredients, defined as substances (other than additives subject to other council directives) not previously consumed to a significant degree in the EU which contain or are derived from genetically modified organisms (GMOs) or are otherwise the result of significant new processing methods.640 The proposed regulations call for the submission of a technical dossier to demonstrate the safety of novel foods for review by national authorities and, in certain cases, the EU's Standing Committee on Foodstuffs.641 Labeling disclosures would be mandatory in most cases.642
The regulation of food additives in Japan is somewhat difficult to discern from the available literature. Indeed, lack of information about or transparency in the standard-setting process has complicated market penetration by foreign firms in the past.643 The United States has included food additives as one of the many sectors over which it continues to have trade disputes with Japan.644
Like most other countries, Japan utilizes a positive or mixed list system. The Ministry of Health and Welfare, with input from the Subcommittee on Food Additives and Toxicology of the Ministry's Food Sanitation Investigation Council, regulates substances used in food.645 The ministry has approved ("designated") approximately 350 substances as food additives,646 and it will now begin regulating "natural" food additives.647 Moreover, the ministry recently proposed revisions to streamline its food additive approval procedures—in addition to more clearly specifying the array of safety studies required in an application, the revision calls for a standard period of one year for processing a food additive application.648
Japanese officials have been active in developing appropriate regulatory strategies for applications of biotechnology in the food and other industries.649 Guidelines issued by the Ministry of Health and Welfare call for manufacturers to file applications comparing their bioengineered substances to existing additives.650 For instance, after spending six months reviewing applications for recombinant chymosin preparations, the ministry affirmed that the substance satisfied its biotechnology guidelines.651 By comparison, the FDA took two years longer to affirm the GRAS status of this substance.652
|This page in the original is blank.|
See id. at 26,648 (citing 21 C.F.R. § 170.30(f)(2), the provision in the regulations which provides that substances of natural biological origin widely consumed prior to 1958 but then significantly altered through selective breeding will be reviewed for possible GRAS affirmation). Of course, as previously explained in Part II, the tomato itself would not be subject to regulation as a "food additive" when sold as produce, even if not GRAS.
See 55 Fed. Reg. 10,932, 10,935 (1990) (codified at 21 C.F.R. § 184.1685 (1996)); see also 53 Fed. Reg. 16,191 (1988) (notice of filing of GRAS affirmation petition for rDNA alpha-amylase enzyme); 51 Fed. Reg. 10,571 (1986) (same); GAO, NEW FOOD TECHNOLOGIES, supra note 9, at 42 (noting that the FDA's review of the GRAS affirmation petition for chymosin took 2.5 years and represented "a learning experience for the agency"); Degnan, supra note 98, at 580-81 (urging greater reliance on the GRAS exception for bioengineered foods); cf. 60 Fed. Reg. 32,904 (1995) (affirming the GRAS status of several enzyme preparations from animal and plant sources).
See 59 Fed. Reg. 26,700, 26,711 (codified at 21 C.F.R. §§ 173.170, 573.130 (1996)) (approving APH(3')II encoded by the kanr gene for use in the development of genetically modified cotton, oilseed rape and tomatoes). In the preamble, the FDA emphasized that "[o]nly the translation product of the kanr gene, APH(3')II, and not the gene itself, is being regulated as a food additive." Id. at 26,701 ("[T]he DNA that makes up the kanr gene does not differ from any other DNA and does not itself pose a safety concern as a component of food.").
See id. at 26,702-03.
See id. at 26,703-04.
See id. at 26,704-06.
See, e.g., CFSAN, Bioengineered Foods Derived from New Plant Varieties (last modified Aug. 1, 1996) <http://vm.cfsan.fda.gov/~lrd/biocon.txt>(listing 22 final consultations conducted with various companies since the 1994 Calgene approval); BIOTECHNOLOGY AND FOOD INGREDIENTS (ISRAEL GOLDBERG & RICHARD WILLIAMS eds. 1991); Alan Goldhammer, The Regulation of Agricultural Biotechnology: An Industrial Perspective, 48 FOOD & DRUG L.J. 501, 50910 (1993); Rhonda L. Rundle, Bright Future Is Predicted For Pest-Resistant Seeds, WALL ST. J., Aug. 31, 1995, at B4.
See 57 Fed. Reg. 22,984 (1992); see also GAO, NEW FOOD TECHNOLOGIES, supra note 9, at 9-10, 31-56 (describing the FDA's approach to the regulation of biotechnology as applied to food, and framing the terms of the debate over the proper regulatory approach); Robert A. Bohrer, Food Products Affected by Biotechnology, 55 U. Prrr. L. REV. 653, 659-66 (1994) (summarizing the FDA's 1992 policy statement).
See 51 Fed. Reg. 23,309, 23,312-13 (1986) (discussing the GRAS or food additive status of bioengineered food substances); 49 Fed. Reg. 50,878, 50,878 (1984) (same); see also 57 Fed. Reg. 6753 (1992) (announcing supplemental OSTP principles); 53 Fed. Reg. 33,182 (1988) (announcing that FASEB would undertake "a study of which scientific concepts and considerations are most appropriately used to determine the regulatory status of foods and food ingredients that are produced by new technologies"); Jeffrey N. Gibbs & Jonathan S. Kahan, Federal Regulation of Food and Food Additive Biotechnology, 38 ADMIN. L. REV. 1, 16-26 (1986); Stephen H. McNamara, FDA Regulation of Food Substances Produced by New Techniques of Biotechnology, 42 FOOD DRUG COSM. L.J. 50 (1987); Michele J. Brace, Comment, Regulation of Genetically Engineered Foods Under the Federal Food, Drug, and Cosmetic Act, 33 AM. U. L. REV. 899 (1984).
See 61 Fed. Reg. 8291 (1996) (announcing that FASEB will undertake a comprehensive review of the appropriate scientific criteria and principles for the assessment of the safety of food ingredients, noting that recent innovations such as bioengineering may "present new situations for which an alternative approach to safety assessment may be needed"). The 1992 policy statement suggested that "many of the scientific considerations for evaluating the safety and nutritional aspects of food from new plant varieties . . . are similar regardless of the [traditional or new] technology used." 57 Fed. Reg. at 22,991; see also International Food Biotechnology Council, Biotechnologies and Food: Assuring the Safety of Foods Produced by Genetic Modification, 12 REG. TOXICOL. & PHARMACOL. S1 (1990).
See 58 Fed. Reg. 25,837 (1993); 57 Fed. Reg. at 22,991. Furthermore, the policy statement lacks any real binding effect. See supra note 309 and accompanying text.
57 Fed. Reg. at 22,984. The Agency recognizes that "[m]ost, if not all, cultivated food crops have been genetically modified" by traditional breeding techniques. Id. at 22,984 n.3; see also id . at 22,985 ("The genetic modification techniques used to develop new plant varieties constitute a continuum."). The policy statement does not address food uses of biotechnology other than in the development of new plant varieties. See id. at 22,985; cf. Steven W. Frank, Food Additive Models for the Regulation of Recombinant DNA Technology Under the Federal Food, Drug, and Cosmetic Act, 45 FOOD DRUG COSM. L.J. 169, 173-79 (1990) (describing the range of potential agricultural applications).
57 Fed. Reg. at 22,986 (describing various genetic modification techniques); see also Charles S. Gasser & Robert T. Fraley, Transgenic Crops, SCI. AM., June 1992, at 62.
57 Fed. Reg. at 22,985.
See id. at 22,986-88 (adding that some of these concerns are magnified with regard to plants used in animal feed because of animals' different consumption patterns); see also Ron Winslow, Allergen Is Inadvertently Transferred to Soybean in Bioengineering Test, WALL ST. J., Mar. 14, 1996, at B6. In its earlier statements, the FDA called for a food additive petition in every case where a substance added to animal feed was produced with recombinant DNA technology. See 51 Fed. Reg. 23,302, 23,311 (1986).
57 Fed. Reg. at 22,984. "The method by which food is produced or developed may in some cases help to understand the safety or nutritional characteristics of the finished food. However, the key factors in reviewing safety concerns should be the characteristics of the food product, rather than the fact that the new methods are used." Id. at 22,984-85; see also OECD, SAFETY EVALUATION OF FOODS DERIVED BY MODERN BIOTECHNOLOGY (1993); David A. Kessler et al., The Safety of Foods Developed by Biotechnology, 256 SCIENCE 1747 (1992).
57 Fed. Reg. at 22,990.
Id. ("For example, if a food derived from a new plant variety contains a novel protein sweetener as a result of the genetic modification of the plant, that sweetener would likely require submission of a food additive petition. . . ."). The policy does not further define what might constitute a significant modification, though similar language appears in the GRAS regulations. See 21 C.F.R. § 170.30(f)(2) (1996); see also Gibbs & Kahan, supra note 408, at 16 n.80 (noting that, "[a]t one point, FDA had proposed that a 'significant alteration' would be a 10% or greater increase in a toxicant or a 20% or greater reduction in a nutrient . . . [but] [t]hese numerical criteria were not adopted"); Frank, supra note 411, at 193. The flow charts make reference to "macroconstituents'' as necessitating possible review, but again the Agency fails to define the term. See 57 Fed. Reg. at 22,998-23,004.
Id. at 22,990.
See id. at 22,991-23,004 (including several flow charts but emphasizing that they do not impose any new regulatory requirements).
Id. at 22,991. "This guidance section provides a basis for determining whether new plant varieties are as safe and nutritious as their parental varieties." Id. at 22,992. This sounds like the test of "substantial equivalence" for medical devices under Section 510(k) of the FD&C Act to determine the need for premarket review. See 21 U.S.C. §§ 360(k), 360c(i)(1) (1994); see also Lars Noah, Amplification of Federal Preemption in Medical Device Cases, 49 FOOD & DRUG L.J. 183, 207-09 (1994). Indeed, the Agency cites international organizations that have proposed a "substantial equivalence" test for new foods. 57 Fed. Reg. at 22,992.
See id. at 22,985. As a consequence, a petitioner may avoid the need to submit an environmental assessment, a potentially significant matter for biotechnology products even though an assessment may have been submitted at an earlier stage to the USDA or EPA. Food additive and GRAS affirmation petitions must either include an environmental assessment or claim a categorical exclusion. See 21 C.F.R. §§ 25.22(a)(10)&(12), 25.24(b) (1996); see also Foundation on Economic Trends v. Heckler, 756 F.2d 143, 153-55 (D.C. Cir. 1985) (enjoining deliberate release experiment of genetically engineered organism until the National Institutes of Health completes an environmental assessment). The FDA recently proposed expanding the list of categorical exclusions. See 61 Fed. Reg. 14,922, 14,927-29 (1996).
See Background Material, FDA Joint CFSAN/CVM Advisory Committee Meeting, Procedures for Industry-FDA Interaction Prior to Commercial Distribution of Foods Derived from New Plant Varieties Developed Using Recombinant DNA Techniques (last modified Nov. 1994) <http://vm.cfsan.fdagov/-lrd/biopro.txt>. Previously, FDA officials had expressed concerns that it lacked the statutory authority to create such a premarket notification system. See GAO, NEW FOOD TECHNOLOGIES, supra note 9, at 12.
See, e.g., CFSAN, Bioengineered Foods Derived from New Plant Varieties (last modified Aug. 1, 1996) <http://vm.cfsan.fdagov/-lrd/biocon.txt> (listing 22 final consultations conducted with various companies since the 1994 Calgene approval); 1995 Hearings, supra note 1, at 13 (statement of Linda A. Suydam, Interim Deputy Commissioner of Operations, FDA); FDA Sets New Procedures for Consulting with Biotech Firms on Safety Issues, FOOD CHEM. NEWS. July 15, 1996, at 3.
torpor to jeopardize food research and innovation."); id. at 35 (statement of Sanford A. Miller, University of Texas Health Science Center) ("The agency's inability to reach decisions on their pending petitions has resulted in nearly a complete stasis in innovation in the food industry."); id. at 54 (statement of Stephen A. Ziller, Vice President, Grocery Manufacturers of America) ("The inherent delays resulting from this inability to make final decisions is beginning to turn off major food companies from investing in new additives and technology."); id. at 69 (statement of Rhona S. Applebaum, Executive Vice President, National Food Processors Association).
See Douglas L. Archer & Catherine M. DeRoever, CFSAN's Program Offices, 48 FOOD & DRUG L.J. 487, 489 (1993) (describing the new Office of PreMarket Approval). The FDA also has proposed to coordinate with the USDA any reviews of substances intended for use in meat and poultry products. See 60 Fed. Reg. 67,490, 67,492 (1995). In addition, as part of the President's "Reinventing Government" initiative, the FDA published an advance notice of proposed rulemaking for its food additive regulations, though all of the suggested changes represent non-substantive revisions intended to eliminate redundancies or ambiguities in the rules governing specific additives. See 61 Fed. Reg. 29,711, 29,712-14 (1996).
See 1995 Hearings, supra note 1, at 10-14 (statement of Linda A. Suydam, Interim Deputy Commissioner for Operations, FDA) (describing initiatives such as performance goals and outreach efforts to better educate petitioners about data requirements, and proposing to reduce environmental assessment requirements for routine petitions); see also Rulis, supra note 218, at 538-39; Administrative Reforms Could Significantly Reduce Petition Backlog, FOOD CHEM. NEWS, Feb. 26, 1996, at 3; FDA Food Ingredient Review Reform is Ongoing, With More Planned for the Future, FOOD CHEM. NEWS, Dec. 25, 1995, at 33; FDA Staff Manual Guide, Policies, Authority, and Procedures for Food and Color Additive Petitions and GRAS Affirmation Petitions (Apr. 1993). Many of these guidelines are available through CFSAN's home page, http://vm.cfsan.fda.gov/-dms/opa-toc.html.
See 1995 Hearings, supra note 1, at 78 (statement of Robert C. Gelardi, Executive Vice President, Calorie Control Council).
See Rulis & Tarantino, supra note 215, at 141 (conceding that most "improvements in the efficiency of the process will take place at the margin, that is, improvements by factors of 10% to 20%"); OPA Contract Looks at Petition Review While Headway Seen on Backlog, FOOD CHEM. NEWS, Oct. 28, 1996, at 17 (reporting that the inventory of active petitions had been reduced from 295 in June 1995 to 240); see also 1995 Hearings, supra note 1, at 10-14 (statement of Linda A. Suydam, Interim Deputy Commissioner for Operations, FDA) (cautioning that "the zeal to speed up the process not be allowed to override the credibility and integrity of that process"); Hutt, supra note 1, at 128 (doubting that the FDA will "reform itself from within, and restore the efficiency and effective regulatory approach taken in the early 1960s").
See 1995 Hearings, supra note 1, at 72 (statement of Rhona S. Applebaum, Executive Vice President, National Food Processors Association) ("We have no reason to expect that new resolve by FDA to mend its ways, in response to current public concerns about its performance, will be sustained over the long term that is necessary for achievement of real action."); id. at 49 (testimony of Sanford A. Miller, University of Texas Health Science Center) ("I don't believe they are going to meet even their performance standards."); see also id. at 2 (statement of Hon. Christopher Shays, Subcommittee Chairman) (noting that, in the week preceding the hearings, the FDA finally acted on several GRAS affirmation petitions that had been pending since 1972).
For petitions classified in Tiers I and II, the goal "is to issue an Agency response—that the submission is adequate, or provide a complete description of why it is not adequate—within" 90 and 180 days respectively. See 1995 Hearings, supra note 1, at 13 (statement of Linda A. Suydam, Interim Deputy Commissioner for Operations, FDA). Moreover, for most Tier III petitions, the announced goal is to provide such an initial response within 360 days, id., double the maximum statutory time frame for final action. See id. at 25-27 (statement of Hon. Christopher Shays, Subcommittee Chairman); see also Protecting the U.S. Consumer from Food-Borne Illnesses: Hearings Before the Subcomm. on Human Resources and Intergovernmental Relations of the House Comm. on Government Reform and Oversight, 104th Cong., 2d Sess. (1996) (testimony of Michael Friedman, Deputy Commissioner for Operations, FDA) (announcing hopes to complete reviews on 50% of new direct food additive petitions within 12 months during the first year after the existing backlog is cleared).
See S. 1477, 104th Cong., 2d Sess. § 103 (1996) (calling for a reduction in backlogs within two years and requiring the publication of annual reports of progress in satisfying the existing statutory deadlines).
See H.R. REP. NO. 436, 104th Cong., 1st Sess. 14 (1995) (additional views of Hon. David M. McIntosh and Hon. Mark E. Souder) (concluding that the "inertia" within the FDA "cannot be eliminated without a significant enforcement hammer," and suggesting that the Agency be given a short period of time after a more realistic review period to either approve the additive or deny the petition if it "can demonstrate that an additive has not been found to be safe"); S. 1477,