FDA has made it abundantly clear on several occasions that there is no “blank check” to use a GRAS substance in any food, at any level of use or for any purpose. Regardless of the origin of a GRAS determination for a substance, the use of the substance has limitations — whether it is a substance from the original 180 GRAS substances proposed in December of 1958, from an FDA GRAS affirmation or from a “self-GRAS” in expectation of a “no objection” response to a GRAS Notification (with some possible gentle arm twisting by FDA at the pre-filing review or during the actual review period). Even those substances listed in the CFR, as “Multipurpose GRAS Food Substances” have at least some restrictions, if only limited to [current] good manufacturing practice (cGMP). However, cGMP also imposes limits, including, “The quantity of the substance added to food does not exceed the amount reasonably required to accomplish its intended physical, nutritive, or other technical effect in food.” The take home message is that there are always limitations and, when the substance is added to food outside of the approved limits, the use of the substance may no longer be GRAS and may be treated as an “unapproved food additive” subject to recall.
Following the passage of the 1958 Amendment to the Food Drug and Cosmetic Act (hereinafter referred to as the Act) that defined GRAS, substances with a history of use in food were first published in a Federal Register notice on December 9, 1958; this became the first of a series of publications collectively known as the GRAS list. The list was codified (i.e., entered into the Code of Federal Regulations) eleven months later and to which several substances were added in subsequent years. In 1969, two of these substances, calcium cyclohexyl sulfamate and sodium cyclohexyl sulfamate, the “cyclamates”, were reported as being carcinogenic and were removed from the GRAS list later that same year. To forestall public outcry that the GRAS system was faulty and consumers were exposed to unsafe ingredients, President Richard Nixon ordered a thorough review of GRAS and some “prior sanctioned”, substances, a task which was carried out over several years by the independent Select Committee on GRAS Substances (SCOGS)., By 1982, after 10 years of work, SCOGS had produced 151 detailed reports addressing over 400 substances.
To review substances, SCOGS gathered information from a variety of sources, including use levels and annual production data provided by manufacturers and published in the NAS/NRC GRAS Surveys of 1970 and 1972. The substances were grouped into like categories (largely based on chemical structure or natural origin) and contractors conducted searches of the scientific literature. On the basis of the information provided, the SCOGS generated conclusions for each individual substance or group in its final reports. Importantly, almost universal among the conclusions reached by SCOGS was at least one qualifying statement emphasizing the limits of approval: “The available information contains no evidence demonstrating that [the substance] constitutes a hazard to the public when used in the manner and quantity now practiced.” [emphasis added]
The SCOGS reports were submitted to FDA and most of the substances were eventually included into 21CFR§182 or §184, although other priorities at FDA precluded finishing the job in its entirety. Despite the variously worded conclusions of SCOGS, FDA undertook a Procrustean effort to allocate all SCOGS reviewed substances into one of five categories., In reality, SCOGS did not develop such terse categories and in fact, at least one report (caffeine) included a (dissenting) minority opinion.
When FDA attempted to codify (i.e., to publish the regulations in the CFR) its affirmations of the SCOGS reports and other GRAS affirmations FDA had performed (i.e., substances that were not the subject of a SCOGS report), the agency found it more convenient to list the ingredients in 21CFR§184 by number rather than by functionality as had been done in 21CFR§182, or by some other limitation, which are many in number. As can be seen by the accompanying table, the limitations of the various GRAS affirmations include the possibility of the following:
Source: The substance must be produced from a specific source, such as α-amylase only being produced from Bacillus stearothermophilus or alginic acid production from “certain brown algae”.
Manufacturing requirements: Production might be limited to a specific pathway, because a different extraction methodology or synthetic pathway might also co-elute or co-produce a substance with a different safety profile, therefore changing the acceptable daily intake of the final product.
Specifications: The specifications may be provided in the regulation itself or as given in a specific issue of the Food Chemicals Codex, otherwise a lesser quality product may be produced.
Functionality limitations: Functional (or “technical”) effects are described in 21CFR170.3(o). In general, there must be a rationale for adding a substance to food, although the substance may have one or more effect. That is, a substance may serve more than one technical effect in one food category versus another e.g., adipic acid as a flavoring agent, leavening agent or pH control agent. Also, functionality may be controlled by the amount of the substance added, as a higher level may impart a different functionality.
Food categories: The various food categories are described in 21CFR§170.3(n). As noted above, a substance may have one function in one food category and another function in a second or third food category. However, aggregate consumption of a substance is often the limiting factor, as the acceptable daily intake may surpassed by use in only a few high consumption categories, regardless of its functionalities.
Limits of addition: Many substances, such as flavoring ingredients, are self-limiting and addition of an excessive amount renders the food unappealing or inedible. Most substances are not self-limiting however and addition of too much can produce unapproved effects or impact safety. As an example of the former, lycopene provides an antioxidant effect at one level, but imparts color at a higher level of addition.
184.1(b)(2) – Restrictions on modification of a GRAS affirmation. According to a prior section of the regulation (i.e., 184.1(b)(1)), some leeway is granted in the use of certain substances, including additional uses by means of a GRAS determination. However, for those substances with a 184.1(b)(2) designation, additional uses must be via a food additive petition (regulation).
Prior sanction waived: Prior to September 6, 1958, FDA or the United States Department of Agriculture (USDA) could respond to a question by a manufacturer if the agency had an objection to the use of a substance; if the agency responded that it had “no objection” the substance was considered to have “prior sanction”. GRAS affirmation for the use of a particular substance often eliminated a “prior sanction” use.
Although not shown in the table, some substances with information indicating they should not be used in food, were allocated to 21CFR§189.1 “Substances prohibited for use in food,” a list that includes the flavor ingredients calamus, safrole and cinnamyl anthranilate, as well as certain products from cattle and lead solder (for sealing cans).
So then, what does the preceding information tell us about FDA’s willingness to use its authority over GRAS affirmations (and possible future FDA-initiated GRAS determinations for animal feed)? First, FDA has exercised the power to limit the use of some substances, despite the popular belief that a substance, once determined GRAS, is open to unrestricted use. The limitations for use should be a lesson for those that wish to apply flavor ingredients (such as extracts of rosemary or cloves) at newer, higher levels in food as an antioxidant (or preservative), when approval has only been for flavor use (conventionally at much lower use levels). Second, because the use of some substances (such as flavors) that were deemed by SCOGS as not constituting a hazard “when used in the manner and quantity now practiced,” may have outstripped the safety information supporting the original safety assessment by SCOGS (that is, at levels in common use in 1970-72 and on the basis of the state of the science 40 years ago). FDA may declare the substance is no longer GRAS based on new safety information as it has for partially hydrogenated oils (PHOs)., Further, FDA has had a long standing policy that the safety-in-use of a substance is the responsibility of the user, the lesson being that regulatory compliance does not constitute indemnification.
On the basis of what has been stated above, what predictions can be made about future FDA actions? First, on the basis of the critical reports on GRAS by the Pew Institute and the Office of Management and Budget, as well as the settlement achieved in the lawsuit from the Center for Food Safety (mandating finalizing of the GRAS Notification Rule and finalizing GRAS guidelines), we can expect FDA to require the updating of GRAS determinations on commonly used food ingredients (including GRAS substances and food additives), similar to the EFSA call for re-evaluation of food additives. Second, it is not inconceivable that a “no objection” for a GRAS notification may have a shelf-life, mandating re-submission or cyclic review. The loss of GRAS status of PHOs recognizes that the safe use of a substance is a dynamic process and requires updates to a GRAS determination with each new scientific publication impacting the safe use of the substance. Third, it is possible FDA may initiate a new survey on ingredient production and/or use levels in updated food categories and technical effects categories – but the sense of urgency and level of cooperation among manufacturers might not be as strong as it was in the early 1970’s.
Another likely consequence of FDA’s discretionary authority to restrict use of food ingredients, is the impending review of nearly 500 substances by the Center of Veterinary Medicine (CVM). These 500 substances are those listed as “ingredient definitions” in the Official Publication (OP) of the American Association of Feed Control Officials (AAFCO). Animal feed manufacturers have only recently been told that these 500 substances require federal (legal and regulatory) compliance (i.e., through a food additive petition to CVM or GRAS status). Lack of federal compliance means these ingredients face the possibility of being identified as unapproved food additives and foods containing these substances could not be added to animal food (including pet food) and could not even be shipped in interstate commerce.
On March 27, 2015, CVM published “The FDA Announces Strategy to Create Definitions and Standards for Animal Food Ingredients”, wherein CVM announced that approximately 250 ingredients will be reviewed for possible GRAS affirmation and the other 250 ingredients will probably require submission of a food additive petition. These ingredients were not previously regulated in the classical sense, but allowed to go to market through a system CVM identified as “regulatory (or enforcement) discretion.” These 500 ingredients represent a vital link in the chain of production of healthy diets for food producing animals and pets. The early AAFCO “ingredient definitions” were often very loosely defined, not designating a species or life stage (such as growing, reproductive phase or milk producing type of commercial livestock) or target animal type (e.g., for chickens, distinguishing between broilers and layers); as such, producers now use the ingredients with little restriction. However, CVM could easily set specifications, life stage or food type limitations, maximum limits, etc., much as the Center for Food Safety and Nutrition did for human foods in the 1970s and 1980s using the SCOGS reports. The animal feed market will be thrown into a turmoil, with many producers having to submit food additive petitions for ingredient approval or potentially be frozen out of the market, possibly for years. Some producers have taken the initiative and generated GRAS dossiers for their use(s) of a substance before CVM has an opportunity to set limits and put other ingredients off-limits by declaring that all changes to the mandated uses require a food additive petition (as in 21CFR§184.1(b)(2) for human food).
The take home messages are clear regarding the power of FDA. First, FDA has used its power to control the use of a food ingredient in regard to naming, sources, manufacturing technique, food categories to which it may added and levels of addition. Second, because it has been too often assumed that a GRAS substance may be used in any food, at any level, for any purpose, the GRAS status of an ingredient may well be challenged by FDA. Third, the agency has now even used its power to determine that a substance once accepted as safe may no longer be GRAS. The stage is set for a test of FDA power in the near future with FDA’s finalizing of the GRAS notification procedure and CVM’s evaluation of non-federally compliant animal feed ingredients.
 GRAS = Generally Recognized As Safe
 Getting to the heart of the matter, FDA consistently notes it is the use or application of a substance that is permitted via a GRAS determination or food additive petition, and not the substance per se. However, as we shall see later, even some manufacturing techniques are prohibited, presumably because unsafe substances are co-produced.
 23 Fed. Reg. 9512 (Dec. 9, 1958) (proposed 21CFR§121.7 (1958))
 Abandoned as a practice in 1997 with the publication of: “FDA proposed simplified GRAS notification system” 62 Fed. Reg. 18937 (April 17, 1997).
 CFR = Code of Federal Regulations
 21CFR§172.5 (see regulation for other requirements).
 “In the past, it has been too often assumed that a GRAS substance may be used in any food, at any level, for any purpose. As a result, the use of some GRAS food ingredients have proliferated to the point where their GRAS status has been brought into serious question.” 39 Fed. Reg.34194 (Sept 23, 1974).
 Prior to passage of the Act, FDA or the U.S. Department of Agriculture had granted explicit approval of specific uses of substances in food prior to September 6, 1958 through regulations specifying their use, through directives or formal opinions, or by letters offering no objection to such uses. Some of these so-called “prior sanctioned” food substances are listed in Part 181 of the CFR, but some prior-sanctioned substances remain unlisted. (http://www.fda.gov/Food/IngredientsPackagingLabeling/GRAS/SCOGS/ucm084142.htm)
 Even prior sanction substances have restrictions on use – “A prior sanction shall exist only for a specific use(s) of a substance in food, i.e., the level(s), condition(s), product(s), etc., for which there was explicit approval by the Food and Drug Administration or the United states Department Agriculture prior to September 6, 1958.” 39 Fed. Reg.34194 (Sept 23, 1974)
 Burdock, G.A. and I.G. Carabin (2004). Generally recognized as safe (GRAS): history and description. Toxicology Letters 150:3-18.
 Some substances from the original GRAS lists remain un-reviewed, but are still considered GRAS and are recorded in 21CFR§182 et seq. According to the regulation, “When the status of a substance [i.e., recorded in 21CFR§182] has been reevaluated, it will be deleted from this part, and will be issued as a new regulation under the appropriate part, e.g., “affirmed as GRAS” under part 184 or 186 of this chapter; “food additive regulation” under parts 170 through 180 of this chapter; “interim food additive regulation” under part 180 of this chapter; or “prohibited from use in food” under part 189 of this chapter.”
 (http://www.fda.gov/Food/IngredientsPackagingLabeling/GRAS/SCOGS/ucm084142.htm) Other references cite 115 reports by SCOGS (http://www.fda.gov/food/ingredientspackaginglabeling/gras/scogs/default.htm).
 NAS/NRC = National Academy of Science/National Research Council
From Carob Bean Gum, PB221952 (National Technical Information Service). (https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml?searchQuery=PB221952)
The fact that FDA actions were not always perfectly congruent with SCOGS conclusions was challenged Federation of Homemakers, Inc. v. Harris. (Hutt et al., 2007. Food and Drug Law Cases and Materials 3rd edition, Foundation Press, NY, p.417), wherein it was decided that SCOGS report conclusions were not binding on FDA.
 It is worthy of note that ”[m]any of the substances currently listed in 21 CFR Part 184 (e.g., menhaden oil, 21 FR 184.1472) were affirmed as GRAS as a result of FDA’s review of a GRAS affirmation petition rather than as a result of FDA’s agency-initiated GRAS review.” http://www.fda.gov/Food/IngredientsPackagingLabeling/GRAS/SCOGS/ucm084142.htm
 21CFR184.1(b)(2) “If the ingredient is affirmed as GRAS with specific limitation(s), it shall be used in food only within such limitation(s), including the category of food(s), the functional use(s) of the ingredient, and the level(s) of use. Any use of such an ingredient not in full compliance with each such established limitation shall require a food additive regulation.”
 Burdock, G. (2015). The Emerging Regulatory Crisis for Pet Food Ingredients. (http://www.petfood2.com/digital-issues/2015/07/the-science-of-food-safety.aspx)
 Federal Register 39:34194-5, 1974.
 Federal Register 78:67169-67175, 2013 and 80:34650-34668, 2015
 Had PHOs been a food additive, getting it off the market would have been much easier under §409 of the Food Drug and Cosmetic Act.
 For a more complete description of the AAFCO ingredient definition, see “The Emerging Regulatory Crisis for Pet Food Ingredients” by George A. Burdock and available at http://www.petfood2.com/digital-issues/2015/07/the-science-of-food-safety.aspxTags: animal feed, CVM, FDA, GRAS