Commissioner
Food and Drug Administration
10903 New Hampshire Avenue
Silver Spring, MD 20093
Re:

Substances Generally Recognized as Safe (GRAS)
Docket No.: FDA-1997-N-0020

The Center for Science in the Public Interest (CSPI), Consumers Union, Environmental
Working Group (EWG), and Natural Resources Defense Council (NRDC) present this comment
regarding the Food and Drug Administration’s (FDA) 1997 and 2010 proposals to use a
voluntary notification process to designate a substance intended for use in food for humans or
animals as “generally recognized as safe” (GRAS). The 1997 proposal is an invalid
interpretation of the Food Additives Amendment of 1958 (FAA). To achieve its congressionally
intended purpose of protecting the public from unsafe chemical additives and come into
compliance with the FAA, FDA has an affirmative obligation to fix its proposal.
Contents
I.

Introduction: FDA’s Oversight of the Safety of Substances in the Food Supply is Broken ................. 3

II.

Legal Framework: The History of GRAS Reveals A Significant But Limited Exemption .................. 5
a.

The Food Additives Amendment of 1958......................................................................................... 5

b.

“Generally Recognized As Safe” ...................................................................................................... 6
i.

Early Implementation of the GRAS Exemption ........................................................................... 7

ii.

The 1997 Proposed Rule ............................................................................................................... 9

iii. Elements of the GRAS Standard................................................................................................. 13
iv. 2010 Reopening of Docket & Center for Food Safety Consent Decree ..................................... 14
III.

The 1997 Proposal Violates the FAA. ............................................................................................ 15

a. The 1997 Proposal violates the FAA and does not merit judicial deference under the Chevron
doctrine. .................................................................................................................................................. 16
i. The 1997 Proposed Rule is contrary to Congress’s “unambiguously expressed” intent under the
FAA..................................................................................................................................................... 17
ii. The 1997 Proposal is not based on a permissible construction of the FAA and therefore does not
merit Chevron deference. .................................................................................................................... 24
b. If FDA were to finalize the 1997 rule as proposed, it would be “arbitrary and capricious” under the
Administrative Procedure Act. ................................................................................................................ 39
IV.
Previous comments and studies in this docket demonstrate that the proposed notification system
does not achieve the purposes of the FAA. ................................................................................................. 41
a. A Key GAO Report found that FDA’s current GRAS system fails to ensure safety of substances
used in food as required by the FAA. ..................................................................................................... 42
b. The Pew Charitable Trusts’s studies found that FDA does not ensure the safety of new and
existing additives as required by statute. ................................................................................................ 43

 c. Previous comments on this docket agree that the GRAS notification system fails to ensure safety
of substances added to food as required by the FAA. ............................................................................. 47
V. Case Studies: Many supposed “GRAS” substances today are in fact unapproved food additives, in
clear violation of Congressional intent under the FAA. ............................................................................. 49
a. The GRAS status of caffeine illustrates how current system impedes informed and consistent
GRAS determinations. ............................................................................................................................ 50
b. Scores of withdrawn GRAS notifications illustrate how the GRAS notification system fails to
ensure the safety of substances used in the food supply. ........................................................................ 55
c. The current GRAS system permits novel “food additives” to bypass the statutory food additive
process. ................................................................................................................................................... 58
i.

Taste Modifiers ........................................................................................................................... 58

ii.

Nanomaterials ............................................................................................................................. 59

d. The lack of appropriate limits on “GRAS” designations allows substances whose safety is
questioned in the scientific community or unsupported by the scientific literature to impermissibly
bypass the food additive process............................................................................................................. 61
i.

Flavorings ................................................................................................................................... 61

ii.

Mycoprotein (“Quorn”)............................................................................................................... 63

VI.
In finalizing the rule, FDA must limit the categories of substances eligible for GRAS
determinations and set clearer requirements for the scientific standards. ................................................... 65
a.

FDA should limit the GRAS concept as Congress intended........................................................... 65
i. GRAS cannot apply to novel ingredients since a scientific consensus regarding their safety has
not yet developed. ............................................................................................................................... 66
ii. GRAS cannot apply to substances flagged as risks to human health by authoritative scientific
bodies or questioned in the scientific community. .............................................................................. 66
iii. FDA must provide a definition of what constitutes “harm” to human health. ............................ 68
iv. GRAS determinations cannot be based on unpublished research and must instead be based on
peer-reviewed, published science. ...................................................................................................... 68
v.

GRAS must be evaluated based on adequate science for the chemical being assessed. ............. 69

vi. GRAS determinations must be made by experts without conflicts of interest............................ 70
b. FDA must require public notice of GRAS determinations to assure that the burdens of general
recognition and ongoing assessment of safety are met. .......................................................................... 73
i.

To achieve Congress’s intent under the FAA, GRAS notifications must be mandatory. ........... 73

ii.

Mandatory notifications must include accurate and current exposure data. ............................... 75

iii. Mandatory GRAS notifications must be public. ......................................................................... 76
iv. FDA should require manufacturers to maintain proper documentation to support their GRAS
determinations. .................................................................................................................................... 77
c.

FDA’s Final Rule should outline a comprehensive safety program. .............................................. 78

VII.

Conclusion: FDA Can and Should Solve the GRAS Problem ........................................................ 80

2

 I. Introduction: FDA’s Oversight of the Safety of Substances in the Food Supply is
Broken
The FDA’s overarching obligation to oversee the safety of chemical additives derives
from the agency’s mission to “protect the public health by ensuring that – foods are safe,
wholesome, sanitary, and properly labeled” under the Food, Drug, and Cosmetic Act (FDCA).1
Its specific responsibilities with regard to food additives were expanded in the Food Additives
Amendment of 1958 (FAA), which was subtitled “[a]n Act to protect the public health by
amending the FDCA to prohibit the use in food of additives which have not been adequately
tested to establish their safety.”2 Given Congress’s description of the purpose of the Act, the
American public would likely find it both disturbing and surprising that thousands of chemicals
added to food today are not approved or even reviewed by FDA.
Instead, of the roughly 10,000 additives currently used in food, more than 3,000 have
never been substantively reviewed by FDA.3 For an estimated 1,000 of these substances, safety
decisions were made by the food industry without any notice at all to FDA.4 Rather than
enabling FDA to fulfill its mission to ensure that the substances added to foods are safe, the
current program leaves the agency in the dark. As Deputy FDA Commissioner for Foods
Michael Taylor remarked in August 2014, “We simply do not have the information to vouch for
the safety of many of these chemicals.”5
The story is one of regulatory decline, decay, and disrepair. After decades of insufficient
funding and internal will to accomplish its mission, punctuated by brief periods of greater
transparency and accountability, FDA essentially granted itself permission to forgo one of its
major jobs. In the 1997 GRAS notification proposal, the agency gave sanction to a twilight
world in which the food industry has been permitted to impose its own set of metrics. The result
of this irrational approach has been a regime in which the food industry has opted for the least
onerous, least costly, and least protective route for introducing new substances into the food
supply. Not only did the proposal create a toothless regulatory program, its voluntary nature
allows industry to bypass FDA’s oversight altogether. Troublingly, the substances designated as

1

21 U.S.C. § 393(2).
Food Additives Amendment of 1958, Pub. L. No. 85-929, 52 Stat. 1041 (1958).
3
Thomas G. Neltner, et al., Navigating the U.S. Food Additive Regulatory Program, 10 Comprehensive
Reviews in Food Science & Food Safety 291, 355 (2011).
4
The Pew Charitable Trusts, Fixing the Oversight of Chemicals Added to Our Food 5 (Nov. 2013),
available at http://www.pewtrusts.org/en/research-and-analysis/reports/2013/11/07/fixing-the-oversightof-chemicals-added-to-our-food. In addition, another 2,000 flavors have been approved as GRAS by the
industry trade group that approves flavors, the Flavor and Extract Manufacturers Association (FEMA).
FDA has received only the most basic information concerning the safety of these substances, but has not
conducted any systematic review or approval of them. Id.
5
Kimberly Kindy, Food additives on the rise as FDA scrutiny wanes, Washington Post, Aug. 17, 2014,
http://www.washingtonpost.com/national/food-additives-on-the-rise-as-fda-scrutinywanes/2014/08/17/828e9bf8-1cb2-11e4-ab7b-696c295ddfd1_story.html.
2

3

 GRAS without FDA approval are likely to be the most dangerous in the food supply, further
perverting the statute, and allowing the unknowns to swallow the law.6
The current secrecy around GRAS substances in foods also undermines FDA’s ability to
conduct meaningful scientific assessments of the safety of food additives or food contact
substances. 7 FDA cannot ascertain the cumulative effect of structurally-related or
pharmacologically-related substances, as required by law,8 given that both the agency and food
companies are in the dark concerning both the toxicology of and public exposure to such
substances. Nor can the public, or public health groups, assist in rationally ordering risks given
the dearth of information. As the evidence below makes clear, the current regime is a near-total
abrogation of FDA’s duty to protect public health and a violation of the statutory requirements
laid out in the FAA.
For these reasons, among others, FDA should not finalize the 1997 proposal9 as written.
To comply with the FAA, FDA should substantially revise and update its approach to GRAS in
two main ways.
First, the current breadth of the GRAS exemption far exceeds the intent of Congress in
drafting the FAA. FDA should take a fresh look at both the language of the statute and its
legislative history and thereby narrow the eligibility of substances for the GRAS exemption. In
its final rule, FDA should delineate and define the category of GRAS substances exempt from
the food additive petition process, by excluding novel substances (which include substances that
do not have a history of safe use in food) and any substances flagged as potentially unsafe by the
scientific community. Changes must also be made to the notification procedure to cure conflicts

6

The sometimes elaborate praise conferred on the agency's GRAS program by the food industry can be
viewed as an attempt to confer reality on a program that lacks conformity with the law, like courtiers
fawning over a naked Emperor about his beautiful clothes. See, e.g., Robert S. McQuate & Richard C.
Kraska, The Future of GRAS Regulations, Nutritional Outlook, March 4, 2015,
http://www.nutritionaloutlook.com/1503/GRAS (last accessed 4/9/15). In this article prepared by GRAS
consultants, the authors noted that the regulated industry has expressed support for the 1997 proposal,
calling it “a very thorough and comprehensive process,” and that the agency should not “disturb a system
that’s not broken” as the “[c]urrent GRAS system works well to adequately ensure the safety of our food
supply.” Id. Nevertheless, the central food trade association, the Grocery Manufacturers Association
(GMA), concedes the program has credibility issues. Leon Bruner, GMA’s chief science officer told The
Washington Post last August, “It’s the right time to take a step back....There are problems with
transparency. How can we be sure that the FDA is aware of ingredients?” Kindy, supra note 5. In
response, GMA has proposed a set of steps for industry to take to standardize GRAS assessments. See
McQuate & Kraska, supra. While the proposal by GMA amounts to an industry admission of the
program's profound deficiencies, a privatized, industry-dominated regulatory substitute could never fulfill
the Congressional assignment to FDA to assure the safety of the food supply.
7
21 U.S.C.A. § 348.
8
21 C.F.R. § 170.3(i) (safety determination requires consideration of the “cumulative effect of the
substance in the diet, taking into account any chemically or pharmacologically related substance or
substances in such diet”).
9
Substances Generally Recognized as Safe, 62 Fed. Reg. 18,938 (Apr. 17, 1997).

4

 of interest between industry and experts making GRAS determinations and eliminate data gaps
in the scientific support underlying such determinations.
Second, because FDA cannot assure public safety if GRAS notification remains
voluntary, FDA must require notification and publication of all GRAS determinations, along
with updated exposure data and reporting of adverse data and events on a periodic basis.
II. Legal Framework: The History of GRAS Reveals A Significant But Limited
Exemption
The FDA’s obligation to oversee the safety of chemical additives derives from the
agency’s mission to “protect the public health by ensuring that – foods are safe, wholesome,
sanitary, and properly labeled” under the FDCA.10 The original FDCA allowed the FDA to ban
poisonous or deleterious adulterants in food; however, it did not directly address food additives.
In 1958, in response to the rapid development of food technology and public concern that unsafe
and untested new additives were making their way into the food supply, Congress passed the
FAA to improve FDA oversight of additives.11
a. The Food Additives Amendment of 1958
The FAA established FDA’s obligation to regulate food additives,12 and required
premarket testing to “prohibit the use in food of additives which have not been adequately tested
to establish their safety.”13 Under the FAA, a “food additive”14 must undergo an FDA premarket approval process to determine whether the anticipated use and application of the new
additive is safe.15
All additives not exempt under the statute are presumed unsafe unless proven
otherwise.16 FDA approval requires a showing that the additive will be safe under the proposed
10

21 U.S.C. § 393(2).
62 Fed. Reg. at 18939 (stating that the amendment was passed “in response to public concern about the
increased use of chemicals in food and food processing”); Food Additives Amendment of 1958, Pub. L.
No. 85-929, 72 Stat. 1784 (1958) (codified at 21 U.S.C. § 348) (purpose of amendment is “to prohibit the
use in food of additives which have not been adequately tested to establish their safety”).
12
Food Additives Amendment of 1958, P.L. 85-929, 52 Stat. 1041 (1958).
13
Id. (codified at 21 U.S.C. § 348).
14
Food additives are defined broadly under the amendment to include “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,” including substances used in packaging, transport,
processing, and preparation. 21 U.S.C. § 321(s).
15
21 U.S.C. § 348.
16
In addition to the “generally recognized as safe” exemption, which is the subject of this comment, in
1995, FDA created the Threshold of Regulation rule, which exempts substances used in food contact
materials from regulation as food additives if the dietary concentration is below 0.5 parts per billion (ppb)
and if: (1) The chemical has not been shown to be a carcinogen; (2) There is no reason to suspect that it is
a carcinogen; and (3) There is no evidence that it presents other health or safety concerns. Food
Additives; Threshold of Regulation for Substances Used in Food-Contact Articles, 60 Fed. Reg. 36,595
(July 17, 1995) (codified at 21 U.S.C. § 170.39).
11

5

 conditions of use.17 The premarket review process begins when someone files a petition seeking
FDA’s approval for use of an additive.18 The petition must include the identity of the additive,
the proposed use, the intended technical effect, the method of analysis in food, and full reports of
all safety studies.19 A specific clause in the statute, known as the Delaney Clause, further
provides that carcinogenic additives may never be approved under the food additive petition
process.20 Safety is determined by considering the potential cumulative effect on consumers of
the substance in the diet, taking into account any chemically or pharmacologically related
substance or substances, the probable consumption of the substance in the diet, and safety
factors.21 If FDA finds the safety of an additive has been adequately established, the agency
proposes a regulation sanctioning the proposed use of the additive, subject to notice and
comment.
b. “Generally Recognized As Safe”
The bulk of the FAA outlines procedures for FDA’s safety evaluations of food additives
and the appeals process for agency determinations. But in a scant few words, the term “food
additive” specifically excludes substances that are “generally recognized as safe,” or GRAS.22
Under the statute, an item is GRAS – and therefore not a “food additive” – if its use is
generally recognized as safe by scientists knowledgeable about the safety of substances added to
food.23 General recognition of safety can be established through “scientific procedures,” or, for
substances used in food prior to 1958, through experience based on the same common use in
food prior to that date.24 Safety is defined, by later FDA regulations, to mean “a reasonable
17

21 U.S.C. § 348(b).
Id.
19
Id.
20
21 U.S.C.A. § 348 (c)(3)(A) (“[N]o 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[.]”).
21
U.S. Food & Drug Administration, Guidance for Industry and Other Stakeholders: Toxicological
Principles for the Safety Assessment of Food Ingredients (Redbook 2000), updated July 2007,
http://www.fda.gov/downloads/Food/GuidanceRegulation/UCM222779.pdf. The potential cumulative
18

effects of additives are assessed based on toxicity studies as directed in FDA’s “Toxicological
Principles for the Safety Assessment of Direct Food Additives and Color Additives Used in
Food” (referred to as the Redbook). Id.
22

21 U.S.C. § 321(s).
Id.; Neltner et al., supra note 3, at 351.
24
21 U.S.C. § 321(s); 21 C.F.R. §§ 130.70(a), (c)(1). Importantly, the 1997 proposal states that it is the
use of the substance, not its mere presence in the food supply, which confers eligibility for the GRAS
exemption based on use. Thus, an “evaluation of whether an additional use of a substance that is GRAS
through experience based on common use in food is also GRAS requires a scientific procedures GRAS
determination when the use in question was not common prior to January 1, 1958.” 62 Fed. Reg. at
18950 (emphasis added). Until the 1980s, the agency refused to recognize use outside the United States;
in the wake of a legal challenge, prior foreign use may now support GRAS status, but only if the
information about such use is readily available and corroborated. Eligibility for Classification of Food
Substances as Generally Recognized as Safe, 53 Fed. Reg. 16,544 (May 10, 1988) (codified at 21 C.F.R.
§ 170.3(f)).
23

6

 certainty in the minds of competent scientists that the substance is not harmful under the
intended conditions of use.”25 However, FDA has not to date provided a definition of “harm” or
what may be considered “harmful.”26
i. Early Implementation of the GRAS Exemption
From the beginning, FDA’s authority to determine the GRAS status of a substance then
in the food supply has correctly been presumed by both the agency and industry. However,
FDA’s approach to assuring the safety of GRAS substances evolved significantly since the
passage of the FAA. Immediately after the law was passed, FDA published a list of GRAS
substances in the Code of Federal Regulations.27 Once listed, these substances could be used,
subject only to “good manufacturing practices” that address the quantity, quality, and grade of
the substance used in food.28
This list was updated over the following years.29 The GRAS list was not intended to be
“exhaustive,” and many substances considered GRAS by the food industry were not included on
FDA’s 1958 GRAS list.30 Although the FAA did not require premarket approval for GRAS
substances, manufacturers would generally seek informal review by the agency before marketing
substances by writing to FDA to request an informal “opinion letter” on the GRAS status of a
particular substance for a particular use.31
This practice continued until 1970, when President Nixon directed FDA to make a critical
evaluation of the safety of “food additives,” including listed GRAS substances, in response to
rising public concern about potentially unsafe GRAS chemicals. In response to this directive, the
Agency announced a comprehensive review of the safety of each substance on the GRAS list.32
The original 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.”33 Thus, the purpose of the Agency’s review was to evaluate
25

21 C.F.R. § 170.3(i).
Laurie J. Beyranevand, Generally Recognized As Safe?: Analyzing Flaws in the FDA's Approach to
GRAS Additives, 37 Vt. L. Rev. 887, 889 (2013).
27
The list was incorporated into the agency’s regulations as 21 CFR parts 182 and 582. 24 Fed. Reg. at
9,368 (Nov. 20, 1959).
28
62 Fed. Reg. at 18,939. GRAS status does not free a substance of FDA controls; use is still subject to
food grade specifications in the Food Chemicals Codex, and its use cannot differ significantly from those
providing the basis for eligibility. Institute of Medicine, Enhancing the Regulatory Decision-Making
Approval Process for Direct Food Ingredient Technologies 24 (1999), available at
http://www.nap.edu/catalog/9453/enhancing-the-regulatory-decision-making-approval-process-for-directfood-ingredient-technologies [hereinafter “IOM Workshop Summary”].
29
IOM Workshop Summary, supra note 28, at 23.
30
Id. at 24.
31
U.S. Food & Drug Administration, FDA’s Approach to the GRAS Provision: A History of Processes
(April 2006), available at http://www.fda.gov/Food/IngredientsPackagingLabeling/GRAS/ucm
094040.htm (last accessed Feb. 24, 2015) [hereinafter “FDA’s Approach to the GRAS Provision”]; Food
Additive Status Opinion Letters: Statement of Policy, 35 Fed. Reg. 5810, 5810 (Apr. 9, 1970).
32
FDA’s Approach to the GRAS Provision: A History of Processes, supra note 31.
33
IOM Workshop Summary, supra note 28 at 23.
26

7

 each substance using “contemporary standards” and “issue each item in a new (i.e., affirmed)
GRAS list, a food additive regulation, or in an interim food additive regulation pending
completion of additional studies.”34
FDA contracted with Federation of American Societies for Experimental Biology
(FASEB) to establish the Select Committee on GRAS Substances (SCOGS) to review and
evaluate the available scientific information on all GRAS substances.35 By 1982, SCOGS had
submitted opinions to FDA on the health aspects of more than 400 substances.36 Of these, the
Committee recommended revocation of GRAS status for 30 substances, and found that an
additional five substances raised safety concerns.37 As of 2009, the FDA had affirmed the
GRAS status of 17 of these substances and taken no action on the remaining ones.38 According
to the U.S. Government Accountability Office (GAO), FDA “could not readily explain why,
even though almost 30 years had passed since the committee completed its work[,] FDA has not
revoked the GRAS status of any of the 18 substances whose safety the committee questioned.”39
Also in 1970, FDA proposed revising the existing regulations regarding safety to state
that “[s]afe must be understood to connote that the Food and Drug Administration, after
reviewing all available evidence, can conclude there is no significant risk of harm from using the
substance as intended.”40 For substances not already on the GRAS list, manufacturers could seek
affirmation of GRAS status by submitting relevant information to the Agency. FDA also made
clear that new tests establishing harm could prompt removal from the GRAS list.41
In 1972, FDA promulgated a rule formalizing the opinion letter practice and providing
for regulatory approval of GRAS substances. Individuals could petition FDA to review the
GRAS status of a substance and issue a regulation confirming GRAS status.42 This process,
known as the GRAS affirmation petition (GAP) process, was subject to notice and comment.43
While such petitions were never mandatory, the GAP process was considered the “primary
mechanism for manufacturers to protect themselves from FDA enforcement actions.”44 Thus, in
the comparably few instances that industry made private GRAS determinations under the GAP
process, manufacturers would commission safety reviews by reputable scientific organizations in
order to address the “obvious regulatory risks” then posed by industry self-determinations.45

34

62 Fed. Reg. at 18,939.
FDA’s Approach to the GRAS Provision: A History of Processes, supra note 31.
36
Id.
37
U.S. Government Accountability Office (GAO), FDA Should Strengthen Its Oversight of Food
Ingredients Determined to Be Generally Recognized as Safe (GRAS), GAO-10-246 21 (Feb. 2010).
38
Id.
39
Id.
40
Eligibility of Substances for Classification as Generally Recognized As Safe in Food, 35 Fed. Reg.
18,623, 18,624 (Dec. 8, 1970) (codified at 21 C.F.R. § 121.3).
41
Id.
42
21 C.F.R. § 170.35(a).
43
Id.
44
IOM Workshop Summary, supra note 28, at 3-4.
45
Id. at 24-26. Specifically, in a few cases, industry commissioned private safety reviews by FASEB.
These reviews would involve published and peer-reviewed study reports. Id. at 24.
35

8

 In 1974, the agency issued a further rule specifying criteria for GRAS status, and
explained the difference between GRAS status and food additive petitions, and the procedures
used in each.46 FDA clarified that a “general recognition of safety” could be demonstrated by
either scientific procedures or by experience based on common use in food before 1958. The
agency emphasized that, in its view, “Congress intended the phrase ‘scientific procedures’ as
used in section 201(s) of the act to have the same dimensions as the full reports of investigations
required to prove the safety of a food additive under section 409 of the act.”47
Thus, the FDA concluded that GRAS determinations “require the same quantity and
quality of scientific evidence as is required to obtain approval of a food additive regulation for
the ingredient.”48 It based its interpretation on two prior Supreme Court decisions that required
the same standard of scientific evidence for GRAS drugs as for new drug applications.49 The
agency also clarified that scientific evidence for GRAS status must be “widely disseminated”
such that it becomes “common knowledge among such scientists.”50 The scheme established in
the 1970s was comprehensive; experts at the time believed that the new formal GRAS
requirements limited the GRAS category to substances used prior to 1958 and meant that the
“only mobility with respect to GRAS status is a loss of that status.”51
ii. The 1997 Proposed Rule
In 1997, FDA published a proposed rule that would create a new GRAS notification
procedure to replace the GRAS affirmation process and make several smaller changes to the
system. First, the 1997 proposal specifically emphasized that the safety standard for food
additives and GRAS substances is the same, but that GRAS also requires common knowledge of
the safety of a use or substance.52
The 1997 proposal also suggested revising the requirements for the “common
knowledge” element of GRAS. First, FDA proposed to broaden the description of common
knowledge by expanding the types of technical evidence of safety that form the basis for a
GRAS determination. Specifically, FDA proposed that “general recognition of safety through
scientific procedures be based upon generally available and accepted scientific data, information,
methods, or principles, which ordinarily are published.”53 This proposal is a subtle but
46

General Recognition of Safety and Prior Sanction for Food Ingredients, 39 Fed. Reg. 34,194, 34,194
(Sept. 23, 1974) (codified at 21 CFR § 121).
47
Id. at 34,194.
48
21 U.S.C. § 321(s); 21 C.F.R. § 170.30(b).
49
Weinberger v. Hynson, Westcott & Dunning, Inc., 412 U.S. 609 (1973); Weinberger v. Bentex
Pharmaceuticals, Inc., 412 U.S. 645 (1973).
50
39 Fed. Reg. at 34,194.
51
Fred H. Degnan, FDA’S Creative Application of the Law 25-26 (2d ed. 2006) (describing the 1974
proposal as a shift from industry GRAS standards to FDA’s standards for approval of food additive,
thereby limiting the applicability of the GRAS exemption and formalizing the GRAS affirmation
process).
52
62 Fed. Reg. at 18,942 (“[A] GRAS substance is neither more safe nor less safe than an approved food
additive.”).
53
Id. For consistency, FDA also proposed to revise the definition of “scientific procedures” to include
“scientific data, information, methods, or principles, which ordinarily are published.” Id.

9

 significant revision to the prior regulations, which had provided that scientific procedures relied
upon “shall ordinarily be based upon published studies which may be corroborated by
unpublished studies and other data and information.”54 The revision elevates the role of
unpublished studies, data, information, and methods from mere corroboration to primary support.
Second, FDA sought to clarify the role of publication in satisfying the general recognition
standard, explaining that publication is “usually necessary, but may not always be sufficient, to
satisfy the common knowledge element of the GRAS standard.”55 It also specified that the
common knowledge element precludes a GRAS determination if the data and information
evaluated by an expert panel are only available in files that are not publicly accessible.56 As this
document will demonstrate, this requirement for GRAS has largely been ignored in practice.
Finally, and most significantly, FDA proposed replacing the voluntary GRAS affirmation
petition process with a voluntary notification procedure.57 While the proposal was never
finalized, the notification procedure has been in force as an interim procedure since it was
proposed in 1997.58 FDA provided three main justifications for replacing the prior system. The
first rationale was its perception that the onerous GAP process was discouraging individuals
from requesting affirmation of self-determined GRAS status.59 The agency thus predicted that a
simpler process would provide an incentive for manufacturers to inform FDA of their GRAS
determinations, resulting in “increased agency awareness of the nation’s food supply and the
cumulative dietary exposure to GRAS substances.”60 Second, FDA predicted that the
“streamlined” notification process would allow the agency to redirect its resources to GRAS
substances “that are a priority with respect to public health protection” and to issuing industry
guidance on food safety issues.61 Third, FDA explained its belief that because it was replacing
one voluntary process (GAP) with another voluntary process, its proposal was “neutral” from a
legal and regulatory perspective.62

54

21 C.F.R. § 170.30(b).
62 Fed. Reg. at 18,943.
56
Id. at 18,943. In responses to GRAS submissions, FDA has indicated that it prefers studies to be
published and “available in the scientific literature for a ‘reasonable amount of time’ (at least several
months) before relying on them to make a GRAS conclusion” in order to provide “evidence that there is
agreement within the scientific community about the conclusions of the study that a substance is safe.”
FDA correspondence regarding GRN 362 for L-Carnitine (on file with CSPI).
57
62 Fed. Reg. at 18,941 (to be codified at 21 C.F.R. § 170.36).
58
Id. at 18,954. FDA included an “interim” notification process in the NPRM that would be used
“between the time of publication of this proposal and any final rule based on this proposal.” The
“interim” process simply directs manufacturers to use the notification process described in the proposed
rule. Id.
59
Id. at 18,941. As one law review article points out, the agency’s justification for its proposal is “ironic”
given that “the burdens cited by the Agency that would prevent an individual from filing a petition fell
largely on the Agency [as] the data and information submitted with the petition needed to be collected for
an independent determination of GRAS status regardless of whether the information was submitted to the
FDA.” Beyranevand, supra note 26, at 905.
60
62 Fed. Reg. at 18,941.
61
Id.
62
Id. at 18,951.
55

10

 In lieu of the GAP process, under the 1997 proposal, any person may notify FDA that a
proposed use of a substance is GRAS by submitting a “GRAS exemption claim.”63 A GRAS
exemption claim includes a “succinct” description of the substance, the applicable conditions of
its use, and the basis for the GRAS determination (i.e., through scientific testing or that it was
commonly used in that manner in food prior to 1958).64 It must include a statement that the
information supporting the GRAS determination is available for FDA review and will be sent to
FDA upon request; however, unlike the GAP process, the notification process does not require
notifiers to submit supporting information with a GRAS notice.65
Upon receiving a claim, FDA will inform the notifier in a letter of one of the following
responses: (1) FDA has “no questions” (and therefore, it is implied – but not stated – that the
substance may be used in food without further FDA review, though this is not a formal
affirmation of GRAS status);66 (2) the notice is insufficient; or (3) FDA has ceased to evaluate
the GRAS notice at the company’s request.67
FDA may question a notifier’s conclusion that a use of a substance is GRAS if the notice:
(1) does not adequately establish technical evidence of safety; (2) lacks evidence that the
knowledge of safety is generally available; (3) does not convince the agency that there is the
63

Id. at 18,941.
Id.
65
Id.
66
Id. at 18,938 (no questions letter “would not be equivalent to an agency affirmation of GRAS status
because FDA would neither receive nor review the detailed data and information that support the GRAS
determination”). For example, in a “no questions” response letter FDA issued regarding whey protein
isolate and dairy product solids, FDA wrote:
Based on the information provided by ADPI in GRP 1G0371, as well as other
information available to FDA, the agency has no questions at this time regarding ADPI's
conclusion that whey protein isolate and dairy product solids are GRAS under the
intended conditions of use. The agency has not, however, made its own determination
regarding the GRAS status of the subject use of whey protein isolate and dairy product
solids. As always, it is the continuing responsibility of ADPI to ensure that food
ingredients that the firm markets are safe, and are otherwise in compliance with all
applicable legal and regulatory requirements.
Agency Response Letter GRAS Notice No. GRN 000037 (Apr. 2000),
http://www.fda.gov/Food/IngredientsPackagingLabeling/GRAS/NoticeInventory/ucm154133.htm (last
accessed March 3, 2015). Despite this express disclaimer and the clear language in the regulation,
decision letters are still sometimes cited by companies as evidence of FDA clearance. Tom Neltner &
Maricel Maffini, NRDC, Generally Recognized as Secret: Chemicals Added to Food in the United States
4 (Apr. 2014), available at http://www.nrdc.org/food/files/safety-loophole-for-chemicals-in-foodreport.pdf; see also Erin Quinn & Chris Young, Why the FDA doesn’t really know what’s in your food,
Center for Public Integrity (Apr. 14, 2015), http://www.publicintegrity.org/2015/04/13/17112/why-fdadoesnt-really-know-whats-your-food?utm_source=feedburner&utm_medium=feed&utm_campaign=
Feed%3A+publici_rss+(The+Center+for+Public+Integrity+Latest+Stories) (quoting Neltner and Maffini
saying that they have seen ingredient manufacturers hawk their additives as “FDA approved” or
“approved by FDA procedures” even though “no questions” letters do not constitute actual approvals).
67
Memorandum from Linda S. Kahl, Center for Food Safety and Applied Nutrition, to Docket No. FDA1997-N-0020, Re: Substances that Are Generally Recognized as Safe (GRAS); Experience with GRAS
Notices 10 (Nov. 2010); see also Neltner & Maffini, supra note 66, at 9.
64

11

 requisite expert consensus about the safety of the substance for its intended use; or (4) is so
poorly presented that the basis for a GRAS determination is not clear.68 FDA can also question
the claim if it is “aware of information that is not included in the notice but raises important
public health issues.”69
If FDA rejects a GRAS notice, it explains its safety concerns in a letter to the company
and publishes the letter on the agency’s Web site. However, a company may withdraw the notice
at any time and ask FDA to cease further review.70 In this case, the agency does not publish the
questions that prompted the agency’s concern and/or the company to withdraw its
notice. Regardless (and remarkably), evidence shows that those or other companies may
continue to market that substance for use in food, despite FDA’s questions about its safety.71
The current notification system represents an absolute nadir in FDA’s already-flawed
implementation of the FAA and is illegal under the statute. The proposal’s notification process
differs from the previous GAP process in that the notifier no longer must provide the detailed
data and information that support the GRAS determination as part of the submission, and that
FDA does not actually make a binding decision as to the GRAS status of the use of a substance
in its response.72 Thus, unlike the GAP process, FDA does not assess the underlying support or
issue a regulation affirming the GRAS conclusion made by the notifier. As a result, the
notification process does not offer the protection from FDA enforcement provided by the GAP
process.73
Yet, both systems suffer from the same overarching deficiency in that they are
completely voluntary. Both the GAP and notification process permit GRAS status to be
determined by industry without any oversight or input from FDA. In other words, they allow
industry to conduct what some have called a “secret GRAS” procedure as the sole prerequisite
for introduction of a new use of a substance in food.74 This lack of transparency and basic
information prevents the entire food safety system from working as intended by Congress.

68

62 Fed. Reg. at 18,950.
Id.
70
Id. at 18,951.
71
Neltner & Maffini, supra note 66, at 9. Companies ask FDA to cease evaluations of their GRAS
determinations with alarming frequency. A review of the list of GRAS notifications on FDA’s website
indicates that 84 notifications have been withdrawn out of a total 562 submissions – 15% of the total
number of notifications. GRAS Notices, http://www.accessdata.fda.gov/scripts/fdcc/?set=GRASNotices
(last accessed 2/25/2015). This issue is discussed in further detail below.
72
62 Fed. Reg. at 18,950.
73
Compare IOM Workshop Summary, supra note 28, at 3 (describing GAP process as the “primary
mechanism for manufacturers to protect themselves from FDA enforcement actions), with 62 Fed. Reg. at
18,950 (explaining that under proposed notification system, “FDA would not be in a position to affirm a
notifier’s conclusion that a use of a substance is GRAS”).
74
Neltner & Maffini, supra note 66, at 2-3 (noting that “Generally Recognized as SECRET” rather than
“Generally Recognized as SAFE” is a better name for the GRAS loophole). However, even though the
GAP process was voluntary, it appears that secret GRAS was much rarer under GAP than under the 1997
proposal. IOM Workshop Summary, supra note 28 at 26 (describing self-determination during the GAP
years as “carr[ying] obvious regulatory risks”).
69

12

 iii. Elements of the GRAS Standard
Through regulation, FDA has clarified that the safety standard for GRAS substances is
the same as for food additives. For a substance to be GRAS, there must be technical evidence of
safety (the “technical element”) along with the additional requirement that the safety of the
substance be generally known and recognized (the “common knowledge element”).75 To satisfy
the technical element, there must be information about the substance establishing that its
intended use is safe.76
Again, FDA has defined “safe” or “safety” to mean that “there is a reasonable certainty in
the minds of competent scientists that the substance is not harmful under the intended conditions
of use.”77 When assessing the safety of food additives, Congress directed that certain factors
must be considered:
(A)

the probable consumption of the additive and of any substance formed in
or on food because of the use of the additive;

(B)

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

(C)

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.78

As FDA has specifically directed that the safety standard for GRAS substances is the
same as for food additives, the same factors must be considered for GRAS substances.79 Thus,
the technical element of a scientific procedures GRAS determination “must consider the
probable consumption and cumulative effect of the substance in the diet.”80
In addition to technical evidence of safety, GRAS substances must satisfy a “common
knowledge” element. The “common knowledge” element includes two factors. First, the data
and information necessary to establish the scientific evidence must be “generally available.”81
Second, there must be a basis to conclude that there is consensus among qualified experts about
the safety of the substance for its intended use.82 This “consensus” does not require unanimity.
62 Fed. Reg. at 18,940 (“[A] GRAS substance is distinguished from a food additive on the basis of the
common knowledge about the safety of the substance for its intended use rather than on the basis of what
the substance is or the types of data and information that are necessary to establish its safety.”).
76
62 Fed. Reg. at 18940.
77
21 C.F.R. § 170.3(i).
78
21 U.S.C. § 348(c)(5).
79
21 C.F.R. § 170.3(i) (defining safety under the statute as involving these three considerations).
80
62 Fed. Reg. at 18,944.
81
Id. at 18,940.
82
Id. at 18,940.
75

13

 Generally, common knowledge is demonstrated through publication in a peer-reviewed scientific
journal; however, under the agency’s 1997 proposal, it may be established through other means,
including “secondary scientific literature,” an “expert panel” convened for this purpose, or
opinions from authoritative scientific bodies.83
iv. 2010 Reopening of Docket & Center for Food Safety Consent Decree
Although FDA never finalized the proposed rule, both the agency and the food industry
have been operating under the proposed procedure as an interim procedure since 1997. Between
February 1, 1998 and December 31, 2009, the Center for Food Safety and Applied Nutrition’s
(CFSAN) Office of Food Additive Safety (OFAS) received and filed approximately 26 GRAS
notices per year.84 It is unknown how many GRAS determinations were made by industry
during that time without notice to the agency. Of the notices received, approximately 79 percent
of GRAS notices came to closure with a “no questions letter,” 5 percent of GRAS notices ended
with an “insufficient basis letter,” and 16 percent of GRAS notices concluded with a “cease to
evaluate letter.”85
In 2010, FDA reopened the docket on the 1997 proposal in light of the length of time that
had passed since the 1997 proposal.86 In reopening the docket, the agency did not change the
1997 proposal, but specifically sought comment on certain aspects of the original proposal,
including the common knowledge element, proposed definition of scientific procedures, and
requirement of cumulative exposure estimates, among others.87
In early 2014, the Center for Food Safety (CFS), an advocacy group, filed a lawsuit
against FDA challenging the 1997 proposal and the agency’s failure to finalize the rule.88 In
October 2014, FDA entered into a consent decree with CFS that calls for FDA to submit a final
rule regarding GRAS for publication no later than August 31, 2016.89 We thus submit this
comment for FDA to consider as it prepares to finalize that final rule.90

83

Id. at 18,941.
Kahl Memo, supra note 67, at 3.
85
Id. at 10-11. For the “cease to evaluate” letters, the content of the notifier’s letter requesting that OFAS
cease to evaluate the notice and the content of OFAS’s response varied and depended on the
circumstances. Id. at 16.
86
Substances Generally Recognized as Safe; Reopening of the Comment Period, 75 Fed. Reg. 81536,
81536 (Dec. 28, 2010).
87
Id. at 81,536.
88
Press Release, Center for Food Safety, Center for Food Safety Sues FDA Over Food Additives (Feb.
21, 2014), available at http://www.centerforfoodsafety.org/press-releases/2924/center-for-food-safetysues-fda-over-food-additives#.
89
Center for Food Safety v. Burwell, Case No. 1:14-cv-267-RC, Consent Decree, ECF No. 14-1 (Oct. 20,
2014).
90
We urge FDA to consider these comments when the agency contemplates finalizing the rule in 2016,
particularly as the agency’s published deadline for finalizing the rule suggests ongoing consideration of
the final rule at FDA and there is considerably more than a year’s time between our submission of this
comment and FDA’s published calendar for a final rule. Agencies generally consider late-filed comments
“to the extent practicable.” See, e.g., 69 Fed. Reg. 16,054 (March 26, 2004) (Department of Health and
84

14

 III. The 1997 Proposal Violates the FAA.
The 1997 proposal is illegal under the FAA. Under it and the interim procedures now in
place, companies may avoid the required statutory food additive petition process, and substances
may be added to food without any real assurance of their safety.91
Rather than furthering the objective of the FAA, FDA’s 1997 proposal – and prior agency
actions – stymies it. Because FDA fails to effectively limit the substances that may fall within
the scope of GRAS, virtually all new chemicals are added to the food supply by manufacturers
using the GRAS exemption – even those that cannot meet the definition for general recognition
of safety. Furthermore, the GRAS exemption essentially replaces the formal food additive
petition process laid out in the FAA. This directly contradicts the statute, which designated the
food additive petition process as the primary means to assure the safety of new substances added
to the food supply.92
In addition, FDA incorrectly interprets the FAA as imposing no obligation on firms to
even inform the agency of any GRAS decisions.93 Because industry is not currently required to
notify FDA of new GRAS substances, FDA lacks a comprehensive catalog of such substances or
their dietary exposure.94 FDA therefore cannot police the border between food additives and
GRAS substances, and neither manufacturers nor FDA can base GRAS determinations (or food
additive petitions) on accurate exposure data and assess the cumulative effect of similar
Human Services will consider late-filed comments); 16 C.F.R. § 1021.12 (Environmental Protection
Agency will consider late-filed comments “to the extent practicable” on Environmental Impact
Statements); 79 Fed. Reg. 24,632 (May 1, 2014) (Department of Transportation); 78 Fed. Reg. at 70888
(Nov. 27, 2013) (Federal Railroad Administration); 79 Fed. Reg. 10760-01 (Feb. 26 2014) (Department
of Homeland Security); 79 Fed. Reg. 45,146 (Aug. 4, 2014) (Federal Highway Administration); 69 Fed.
Reg. 32,200 (June 8, 2004) (Department of Agriculture); 62 Fed. Reg. 24,614 (May 6, 1997) (Consumer
Product Safety Commission); 61 Fed. Reg. 1,412 (Jan. 19, 1996) (Office of Management and Budget).
CSPI, NRDC, and Consumers Union all previously commented during the open docket periods. These
comments supplement our original submissions to reflect relevant and significant information that has
emerged in the ensuing two decades since the rulemaking was proposed. Specifically, the Pew Charitable
Trusts and NRDC have published significant studies on the deficiencies of the GRAS program since the
close of the second comment period. This comment should also be considered in the interest of fairness –
the 19-year gap between the proposal and the final rule creates space for an empirical assessment of how
the proposal has or has not functioned – and much of our comment demonstrates that FDA’s proposal has
failed to meet the agency’s stated goals in the NPRM or correct pre-existing inadequacies in FDA’s
oversight of the safety of GRAS substances, and in practice undermines the statutory goals set out by
Congress.
91
To the contrary, this was the very assignment given to FDA by the law – the “Act to protect the public
health by amending the FDCA to prohibit the use in food of additives which have not been adequately
tested to establish their safety.” Food Additives Amendment of 1958, P.L. 85-929, 52 Stat. 1041 (1958).
92
Should industry complain that the food additive process is too burdensome, their problem is not with
FDA, but with the design of the law in the FAA. FDA cannot and should not use this as an argument to
circumvent clearly expressed Congressional intent.
93
Pew Charitable Trusts, Fixing the Oversight of Chemicals Added to Our Food (Nov. 2013), at 1.
94
GAO Report, supra note 37, at 33. Furthermore, FDA is not made aware of whether companies track
the evolving scientific information regarding substances determined to be GRAS, further limiting its
ability to make ongoing safety assessments. Id. at 25.

15

 substances as required by law. Furthermore, as will be demonstrated below, companies base
their GRAS determinations on stale, conflict-ridden, and often unpublished, non-peer-reviewed
science. In sum, the proposed rule establishes no real oversight over the safety of GRAS
substances, and thus achieves little regarding public health, in clear violation of the FAA.
As a result of these serious deficiencies, the current GRAS process as defined by FDA
“frustrate[s] the policy that Congress sought to implement” in passing the FAA and is an
unlawful abrogation of FDA’s responsibilities to protect the public health from both unsafe and
unproven additives.95 It is therefore contrary to the intent of Congress as “unambiguously
expressed” in the FAA. If FDA were to finalize the 1997 rule as proposed, it would thus be
unlawful under the statute.96 In the alternative, even if the statutory language were held to be
ambiguous, FDA’s interpretation is “inconsisten[t] with the design and structure of the statute as
a whole,” and does not merit deference under Chevron.97 Finally, if FDA were to finalize the
rule as proposed without revision, it would be “arbitrary and capricious” under the
Administrative Procedure Act for its failure to consider many manifest problems with the interim
system, the flaws and predictable errors in the original proposal, and the evidence presented to
the agency since 1997.98
The illegality of the proposed rule is amply demonstrated in the following section by
reference to the plain language and structure of the FAA and its legislative history. It is further
confirmed by the comments already on the docket and through several case studies illustrating
the practical flaws of the proposed system, which will be discussed in Sections IV and V. FDA
is fully aware of this record of failure99 and has an affirmative obligation to correct it in the final
rule on GRAS.
a. The 1997 Proposal violates the FAA and does not merit judicial deference
under the Chevron doctrine.
In light of the language, structure, and legislative history of the FAA, FDA’s
interpretation of the GRAS exemption is unlawful and deserves no deference by the courts.
Judicial review of FDA’s decisions is governed by Chevron’s two step analysis:
First, always, is the question whether Congress has directly spoken to the precise
question at issue. If the intent of Congress is clear, that is the end of the matter;
95

Fed. Election Comm'n v. Democratic Senatorial Campaign Comm., 454 U.S. 27, 32 (1981).
Chevron v. Natural Resources Defense Council, 467 U.S. 837, 842-43 (1984).
97
Util. Air Regulatory Grp. v. E.P.A., 134 S. Ct. 2427, 2442 (2014) (quoting University of Tex.
Southwestern Medical Center v. Nassar, 133 S.Ct. 2517, 2529 (2013)).
98
Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43-44 (1983).
99
Speaking to the New York Times in 2013, FDA’s Deputy Commissioner for Foods Michael Taylor
acknowledged a need for the agency to revisit the FAA: “From our standpoint, we do need to look at
whether this regime established by Congress almost 60 years ago gives us the information we need. It
would be desirable for F.D.A. to have more information on products being added to food.” Stephanie
Strom, Drink Ingredient Gets a Look, N.Y. Times, Dec. 12, 2012,
www.nytimes.com/2012/12/13/business/another-look-at-a-drink-ingredient-brominated-vegetableoil.html?_r=0.
96

16

 for the court, as well as the agency, must give effect to the unambiguously
expressed intent of Congress. If . . . Congress has not directly addressed the
question at issue, the court does not simply impose its own construction on the
statute, as would be necessary in the absence of an administrative interpretation.
Rather, if the statute is silent or ambiguous with respect to the specific issue, the
question for the court is whether the agency’s answer is based on a permissible
construction of the statute.100
In this case, FDA’s interpretation of the statute does not merit deference at either step of
the Chevron analysis. First, Congress has directly spoken to the question at issue. The GRAS
category, as it is currently applied by FDA and the food industry, includes substances that that
are not “safe” or “generally recognized” as safe, and thus plainly contradicts Congress’s
“unambiguously expressed” intent under the FAA. Second, even if the statute were viewed as
ambiguous, FDA’s position should not receive deference as it is an impermissible construction of
the law in light of “the broader context of the statute as a whole.”101
i. The 1997 Proposed Rule is contrary to Congress’s “unambiguously
expressed” intent under the FAA.
Under the 1997 proposal, and prior decisions, FDA essentially cedes the regulation of
food additives to industry by applying a reading of the GRAS exemption that allows industry to
make secret GRAS determinations without meaningful limitations by the agency. As a result,
there are few restrictions, in practice, on what qualifies as “generally recognized as safe,” in
plain violation of the statutory language. Because anything can be GRAS, the exemption has
entirely overtaken the “food additive” petition process laid out in the FAA. This is plainly
contrary to the structure of the law, which makes clear that Congress was establishing a
premarket review process for substances added to foods. To ensure that the GRAS and “food
additive” definitions have meaning, GRAS determinations cannot be secret. Furthermore, safety
determinations under the FAA require exposure data, which cannot be supplied when industry
and FDA are both unaware of uses and exposures of substances in the food supply.
1. The current scope of the GRAS exemption exceeds the plain
language of the statute.
The text of the FAA shows that Congress intended to establish FDA oversight over new
food additives to ensure their use was safe before they were consumed by the general public,
exempting substances “generally recognized as safe.” The plain words of the GRAS exemption
indicate that it was intended to permit those uses of substances that were relatively well known,
uncontroversial in light of then-current scientific data, and long used for that purpose, while still
subjecting both unknown additives and potentially unsafe known additives to mandatory
premarket testing. Instead, FDA has expansively interpreted “generally recognized as safe,” and
in so doing has both undermined the statutory food additive petition process and failed to meet
its statutory obligations to ensure the safety of all substances added to the food supply.

100
101

Chevron, 467 U.S. at 842-43.
Robinson v. Shell Oil Co., 519 U.S. 337, 341 (1997).

17

 Today, almost all new substances in the food supply are self-determined GRAS, even
those that are lab-created chemicals, have no history of safe use in food, or are otherwise novel.
It is clear from the language and structure of the FAA that the current scope of the GRAS
category has expanded unrecognizably beyond the original intent of Congress. Both indicate that
GRAS was never intended to apply to unknown and unproven substances. In fact, these
substances were precisely the intended target of the “food additive” approval requirements under
the law.
The plain meaning of the words “general recognition” indicates that the exemption was
not intended to apply to novel or newly discovered substances known only to a few people.
“General recognition” of safety necessarily implies an awareness of the substance in the
scientific community. This obvious interpretation has been adopted by the courts and by FDA:
both have construed the language as requiring an “expert consensus” based on “substantial
evidence.”102 Such consensus is impossible to achieve if a chemical’s existence and use are
unknown to the general scientific community (or the relevant segment of it) and to FDA, as
“unawareness of a product among qualified experts precludes a finding of general
recognition.”103
Furthermore, courts have found that the substantial evidence required for a general
recognition of safety “consists of adequate and well-controlled studies that must be generally
available to the scientific community.”104 Thus, prior judicial analysis of the plain language of
the law indicates that the food additive GRAS exemption was not intended to apply to novel
substances whose properties or safety are not widely known. Yet, as demonstrated by the case
studies at section V infra, under the 1997 proposal, novel and unknown substances are
designated as GRAS and bypass the statutorily intended food additive petition process – a stark
divergence from the unambiguous language of the statute.
Moreover, because the 1997 proposal expands the technical evidence requirements for
GRAS to include “scientific data and scientific information,” thereby virtually eliminating the
requirement for peer-reviewed studies in deciding GRAS status, it broadens the description of
“common knowledge” so as to make the provision meaningless. Because industry’s GRAS
determinations are essentially unregulated, firms use their own employees, consultants, and
Weinberger v. Hynson, Westcott & Dunning, Inc., 412 U.S. 609, 632 (1973) (interpreting “generally
recognized” language in the drug context); United States v. An Article of Drug Consisting of 4,680 Pails,
More or Less, Each Pail Containing 60 Packets, Etc., 725 F.2d 976, 985 (5th Cir. 1984) (“‘[G]eneral
recognition’ requires a two-step showing: first, that there is general recognition in fact, i.e., that there is an
expert consensus that the product is effective; and second, that the expert consensus is based upon
“substantial evidence” as defined in the Act and in FDA regulations.”); 62 Fed. Reg. at 18,939 (“To
establish such recognition, the proponent must show that there is a consensus of expert opinion regarding
the safety of the use of the substance.”).
103
United States v. Articles of Drug . . . Hormonin, 498 F. Supp. 424, 431-32 (D.N.J. 1980) aff'd sub nom.
Appeal of Carnrick Labs., Inc., 672 F.2d 902 (3d Cir. 1981) and aff'd sub nom. United States v. Articles of
Drug, 672 F.2d 904 (3d Cir. 1981).
104
United States v. An Article of Drug Consisting of 4,680 Pails, 725 F.2d at 987 (emphasis added); see
also United States v. An Article of Drug . . . "Bentex Ulcerine", 469 F.2d 875, 880 (5th Cir. 1972)
(observing that if a substance is generally recognized as safe, one would expect to “find in the medical
literature over a period of years support for this premise from wide experimentation and study”).
102

18

 experts to make safety decisions, resulting in ubiquitous conflicts of interest.105 These
determinations are often based on stale and unpublished, non-peer-reviewed science.106 Such
backroom determinations cannot meet any plain definition of “general recognition” of safety in
the scientific community.
The statutory language also unambiguously designates as GRAS any substances whose
safety is seriously disputed among qualified experts.107 FDA itself has acknowledged that “it is
well settled that a mere showing that the use of a substance is ‘safe’ is not sufficient to exempt
the substance from the [FAA’s] definition of “food additive.”108 Instead, the substance must also
be shown to be “generally recognized” as safe.109 Again, both the agency and the courts have
found that “general recognition” requires a “consensus of expert opinion” regarding the safety of
the use of the substance.110
If a requirement for general consensus of safety among experts means anything at all, it
must exclude any substances that an authoritative body of scientists has linked to adverse
outcomes for human health, as it is self-evident that such substances cannot receive the benefit of
a “general consensus” of opinion among experts knowledgeable about their safety.111 Yet, under
the current system, as demonstrated below, substances flagged as a risk to human health by
authoritative bodies can be designated as GRAS, in violation of the unambiguous statutory
language.
FDA’s current system places no meaningful limits on the scope of the GRAS exemption.
As will be shown in the case studies, industry regularly designates novel and lab-created
substances as GRAS, as well as substances whose safety has not been proved those whose safety
105

See discussion of funding bias, infra at Section IV.b.
See id.
107
United States v. Undetermined Quantities of Various Articles of Drug . . . Equidantin Nitrofurantoin
Suspension, 675 F.2d 994, 1000 (8th Cir. 1982) (finding “a genuine dispute concerning the safety and
effectiveness of a drug product . . . precludes a finding of “general recognition”) (quoting United States v.
Articles of Drug (Hormonin), 498 F.Supp. at 431-32). This requires a comprehensive review of the
literature, as FDA regulations specify that an assessment of the safety of a food substance generally
involves an evaluation of information about its safety and functionality including all studies and tests of a
food additive on animals and humans and all studies and tests of a food additives for identity, stability,
purity, potency, performance and usefulness. 21 CFR Sec 171.1 (h(4)) (emphasis added).
108
62 Fed. Reg. at 18,939 (citing United States v. An Article of Food * * * Coco Rico, Inc., 752 F.2d 11,
15 n. 4 (1st Cir. 1985)).
109
Id. (citing Coco Rico, 752 F.2d at 15 n. 4); see also United States v. Articles of Food and Drug * * *
Coli-Trol 80, 518 F.2d 743, 745 (5th Cir. 1975).
110
Id.; see also United States v. Western Serum Co., Inc., 666 F.2d 335, 338 (9th Cir. 1982); United States
v. Articles of Drug * * * Promise Toothpaste, 624 F. Supp. 776, 778 (N.D. Ill. 1985), aff’d 826 F.2d 564
(7th Cir. 1987); United States v. Articles of Drug * * * Hormonin, 498 F. Supp. 2d 424, 435 (D.N.J.
1980). Courts further found that even published data may not be sufficient to establish GRAS status
unless they had been “collected” with regard to establishing the safety of a specific use and evaluated in
the context of a scientific “peer review.” Degnan, supra note 51, at 27.
111
Note that such substances would remain eligible for use in the food supply as additives approved at
levels below which they pose a risk to human health under the food additive approval process, should the
food industry be able to show they may safely used for some purposes and at some level in food.
106

19

 has been questioned by authoritative bodies. FDA’s system thus permits GRAS determinations
for substances that cannot meet any possible interpretation of the words “general recognition of
safety.” As a result, FDA’s interpretation of the GRAS exemption is legally unacceptable:
agencies do not have the power to “rewrite[e] unambiguous statutory terms.”112
2. FDA’s expansive interpretation of GRAS nullifies the statutory
food additive petition process.
The effectively unlimited expansion of GRAS also violates the statute because it allows –
and in fact encourages113 – unapproved food additives to masquerade as “GRAS” substances.
The statutory design of the FAA makes clear that the food additive petition process was the
primary process Congress intended to assure the safety of new and untested substances added to
the food supply.114 Yet, because FDA has failed to sensibly limit the scope of the GRAS
exemption, far more new substances added to food today bypass the formal food additive
petition process than go through it, rendering the FAA’s food additive petition process virtually
irrelevant.115

112

Util. Air Regulatory Grp. v. E.P.A., 134 S. Ct. at 2446.
The GRAS exemption – whether FDA is notified or not – is a far less labor-intensive process than the
food additive petition process, which can sometimes take years to reach approval. Thus, it would be
reasonable to conclude that it would be more attractive to industry – and the extreme decline in food
additive petitions compared to GRAS notifications since 1997 (displayed in Washington Post graphic)
indicates that this is in fact the case.
114
Should industry complain that the food additive process is too burdensome, their problem is not with
FDA, but with the design of the law in the FAA. FDA cannot and should not use this as an argument to
circumvent the clearly expressed Congressional intent.
115
See Comments by Lisa Lefferts, Center for Science in the Public Interest, quoting Tom Neltner,
“Conflicts of Interest in Approvals of GRAS Additives: Out of Balance,” slides available at
http://www.pewtrusts.org/~/media/legacy/uploadedfiles/phg/content_level_pages/issue_briefs/PotentialC
OIinGRASWorkshoppdf.pdf (“Over the past ten years, GRAS Notifications outnumber Food Additive
Petitions by more than 14 to 1 for direct additives.”); see also IOM Workshop Summary, supra note 28,
at 1 (“In the absence of timely approval, industry is likely to rely more heavily on the statutory exception
for Generally Recognized As Safe (GRAS) substances to facilitate the marketing of new ingredients.”).
113

20

 Source: Graphic by Kimberly Kindy and Cristina Rivero for the Washington Post, August 17, 2014.
Data Sources: Natural Resources Defense Council, Food and Drug Administration, Center for Science in the Public
Interest, Pew Charitable Trusts.

Thus, FDA’s expansive interpretation of the GRAS exemption allows substances that
should be subject to the food additive petition process to bypass agency oversight, a stark
nullification of the unambiguous statutory procedure laid out by Congress in the FAA. This is
further exacerbated by the voluntary nature of the GRAS notification process, as “secret GRAS”
prevents FDA from ensuring that “food additives” under the law are subject to the statutorily
required preapproval process. For the “food additive” definition to have any meaning in
practice, FDA must police the GRAS exemption to ensure that potentially unsafe additives are
pretested for safety under the food additive petition (FAP) process as required by the statute –
which it cannot do under a voluntary notification process.
The “secret GRAS” process also creates a perverse and undermining incentive: because
industry does not ask for or receive explicit FDA approval, FDA may not be notified of
substances that pose the most substantial risks to public health. As CSPI noted in its original
comments on the proposed rule, “there will always be some companies who will never notify
FDA of their GRAS self-determinations [and t]hose companies are the ones most likely to have
self-declared substances as GRAS that may not actually be safe.”116 The voluntary notification
system thus allows unapproved food additives that may in fact be the most dangerous to evade
the regulatory process.
As a result of these deficiencies, in violation of any possible construction of the FAA, the
GRAS exemption has completely overtaken the definition of “food additive.” The language and
structure of the statute make clear that this interpretation could not have possibly been the
intention of Congress in passing the FAA. Indeed, if Congress had intended the GRAS
116

CSPI Comment Re: Substances Generally Recognized as Safe; Proposed Rule Docket No. 97N-0103,
July 16, 1997.

21

 exemption to overtake the statutory food additive petition process, it would have said so, as
Congress “does not . . . hide elephants in mouseholes.”117 Instead, FDA’s expansive
interpretation of GRAS turns an exception into the dominant rule, in plain violation of
Congress’s unambiguous intent.
3.

“Secret GRAS” eviscerates the safety standard under the
statute.

Under the “secret GRAS” system sanctioned by FDA under the current regulations, many
“food additives” avoid the statutorily mandated food additive petition process, in plain violation
of the statute. Moreover, because the current notification system is voluntary, FDA has limited
oversight over these additives and is unaware of public exposures to and the use of an estimated
1,000 substances – so-called “secret GRAS.”118 This leaves the FDA, industry, and public
unaware about uses of substances in foods and their exposure levels – yet analysis of such data is
essential to making a determination of safety under the plain language of the FAA.
The FAA specifically dictates that FDA consider exposure and cumulative effects when
determining the safety of food additives. 119 To meet these requirements, FDA must be able to
estimate what type of foods contain the substance and how much of the substance is likely to be
in each type of food. If someone other than FDA has made a GRAS safety determination for the
substance, the agency is obligated by subparagraph (A) to know about the exposure likely to
result from that decision. If FDA is not notified of the decision (as is often the case under the
current voluntary notification system), then neither FDA nor other manufacturers can estimate
probable consumption as required by Congress. Instead, under the current system, multiple
companies may privately conclude that an ingredient is safe at a specific level, while their
combined uses actually result in unsafe exposure levels.120 In other words, the express
117

Whitman v. Am. Trucking Associations, 531 U.S. 457, 468 (2001).
See Neltner, supra note 3, at 342; Neltner & Maffini, supra note 66, at 2.
119
In determining, for the purposes of this section, whether a proposed use of a food additive is safe, the
Secretary shall consider among other relevant factors-(A)
the probable consumption of the additive and of any substance formed in or on food
because of the use of the additive;
(B)
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
(C)
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.
21 U.S.C. § 348(c)(5). This safety standard also applies to GRAS substances, which by regulation must
meet the same standard for safety as that for food additives. 21 C.F.R. § 170.30(b). These requirements
were also reiterated by Congress when it passed the Food and Drug Administration Modernization Act
(FDAMA) in 1997, which specifically required consideration of criteria such as the probable consumption
of such food contact substance and potential toxicity of the food contact substance in food contact
substance notifications. Pub. L. 105–115, §309, 111 Stat. 2354 (1997) (codified at 21 U.S.C. § 348(h)).
120
The Center for Public Integrity provided an effective illustration of this problem:
Consider how this could play out at a family’s breakfast table. A company might
privately conclude that an ingredient is safe at a specific level in cereal. Another might
do the same for its use in muffins. Ditto for a third company adding it to juices. A
118

22

 requirements of the FAA for food additive petitions – and for GRAS under applicable
interpretations upheld in the courts – cannot be met when notifications are voluntary.
In addition, to estimate the cumulative effect, Congress explicitly requires FDA to take
into account “any chemically or pharmacologically related substance.”121 To meet this statutory
requirement, the agency must know the chemical structure of all of the components of the
additive in order to identify chemically-related substances. It must also know how the additive
acts upon and affects the body, for example, the mechanism of action and physiological effect of
the substance, in order to identify pharmacologically-related substances. Then, to assess the
cumulative effects, FDA must know the amount of exposure, taking into account all chemically
and pharmacologically-related substances. Only by knowing all of this could FDA – or another
industry notifier for that matter – begin to consider the cumulative effect of the substance and
chemically and pharmacologically-related substances and satisfy the express requirements of the
statute.122
The language and structure of the statute definitively show that the system Congress
created in the FAA cannot function when such determinations are secret.123 Thus, the voluntary
nature of the notification system – “secret GRAS” – is illegal under the FAA. FDA itself has
acknowledged that “it would be informative for FDA to have . . . at least an awareness of the
existence of substances that are independently determined to be GRAS, even in the absence of a
GRAS notice submitted by a company.”124 FDA’s acknowledgement is a serious
family sitting down to a breakfast of all three foods could end up consuming much more
of the ingredient than each company had anticipated.
Quinn & Young, supra note 66.
121
21 U.S.C. § 348(c)(5).
122
For an example of how such a cumulative effect assessment would be conducted, see generally
Maricel Maffini & Thomas Neltner, Brain Drain: The Cost of Neglected Responsibilities in Evaluating
Cumulative Effects of Environmental Chemicals, J. Epidemiol. Community Health 1-4 (2014). In their
analysis, they focus on chemicals likely to affect children’s brain development. They assembled a list
from FDA’s own toxicology database that the agency noted as affecting the brain, hypothalamus (the part
of the brain that interfaces with the endocrine system), or the thyroid. In the womb, the thyroid is
essential for brain development. Even transient disruptions of the thyroid could interfere with fetal brain
development in an inalterable manner. The list also included the results of the federal governments
Tox21 in vitro testing program of chemicals that either inhibited or activated thyroid hormone receptors.
FDA is active participate in the Tox21 program. Finally, it included pesticides that the European Food
Safety Authority determined may harm the thyroid or the nervous system.
123
This logic was also apparent to the U.S. Government Accountability Office when, in its 2010 report, it
recommended:
To better ensure FDA's oversight of the safety of GRAS substances, the Commissioner of
FDA should develop a strategy to require any company that conducts a GRAS
determination to provide FDA with basic information—as defined by the agency to allow
for adequate oversight—about this determination, such as the substance's identity and
intended uses, and to incorporate such information into relevant agency databases and its
public Web site.
GAO Report, supra note 37, at 34.
124
Comments from the Food and Drug Administration in GAO Report, supra note 37, at 62. FDA has
also acknowledged that it “cannot ensure that GRAS determinations that are not currently notified to FDA
are rigorous, robust, or consistent with agency’s proposed criteria.” Id.; see also Kindy, supra note 5

23

 understatement. Such awareness is not simply “informative,” it is essential for FDA to meet its
statutory obligations. When FDA’s acceptance of “secret GRAS” is considered in light of its
obligations under 21 U.S.C. § 348(c), it is clear that the FDA’s proposed rule is contrary to the
unambiguously expressed intent of Congress under the FAA.
FDA’s current interpretation of the law allows an expansive interpretation of GRAS that
contradicts the plain meaning of “general recognition of safety” and nullifies the statutory
definition of “food additive.” Its secret nature permits unknown and unproven additives to
bypass the statutorily mandated FAP process and makes it impossible for industry or the agency
to obtain the exposure data necessary to meet the safety requirements in the law. While FDA has
the power to resolve some questions left open by Congress in administering the law, it does not
have the power to “rewrite clear statutory terms to suit its own sense of how the statute should
operate.”125 Thus, FDA’s actions are unlawful and due no deference, as FDA may not make
“administrative constructions . . . contrary to clear congressional intent.”126
ii. The 1997 Proposal is not based on a permissible construction of the
FAA and therefore does not merit Chevron deference.
In the alternative, even if the statute were found to be ambiguous, FDA’s interpretation of
the GRAS exemption still would not merit deference under Chevron. Where a statute is
ambiguous, the question becomes whether the agency’s action is based on a permissible
construction of the statute.127 Thus, even if the FAA is considered ambiguous, FDA must still
operate “within the bounds of reasonable interpretation.”128 To be reasonable, the agency’s
interpretation must account for both “the specific context in which . . . language is used” and “the
broader context of the statute as a whole.”129 A statutory “provision that may seem ambiguous in
isolation is often clarified by the remainder of the statutory scheme . . . because only one of the
permissible meanings produces a substantive effect that is compatible with the rest of the
law.”130
Applying this framework, FDA’s interpretation of its obligations under the statute is
plainly unreasonable in light of the broader context of the statute as a whole. The FAA was
intended to create a system of regulatory oversight over new and potentially unsafe substances
added to foods: “An Act to protect the public health by amending the FDCA to prohibit the use
in food of additives which have not been adequately tested to establish their safety.”131 FDA’s
current interpretation of GRAS – that it permits a secret, virtually unlimited system of industry
determinations – is plainly incompatible with the rest of the law, as it nullifies the premarket
(quoting FDA’s Michael Taylor: “We simply do not have the information to vouch for the safety of many
of these chemicals[; w]e do not know the volume of particular chemicals that are going into the food
supply so we can diagnose trends[; and w]e do not know what is going on post-market.”).
125
Util. Air Regulatory Grp. v. E.P.A., 134 S. Ct. at 2446.
126
Chevron, 467 U.S. at 843.
127
Id. at 842-43.
128
Arlington v. FCC, 569 U.S. __, 133 S.Ct. 1863, 1868 (2013).
129
Robinson., 519 U.S. at 341.
130
United Sav. Assn. of Tex. v. Timbers of Inwood Forest Associates, Ltd., 484 U.S. 365, 371 (1988).
131
Food Additives Amendment of 1958, Pub. L. No. 85-929, 52 Stat. 1041 (1958).

24

 testing process and prevents FDA from exercising ongoing oversight of the safety of substances
added to foods. Thus, while the words “generally recognized as safe” may have multiple
possible constructions, when viewed in light of the statute as a whole, the 1997 proposal goes
well beyond the “bounds of statutory authority”132 and FDA’s interpretation does not merit
deference under Chevron step two.
1. In the FAA, Congress intended to establish a comprehensive
regulatory scheme overseeing the safety of substances added to
foods.
Under the 1997 proposal, FDA cedes far too much of its food additive oversight. The
agency’s construction of the law is patently unreasonable in light of the purpose of the FAA,
which was to create a comprehensive regulatory scheme overseeing the safety of all substances
added to foods. When the FAA was passed, both lawmakers and leading scientists were deeply
concerned about the food industry’s increasing use of untested food additives. The statute was a
response to the lack of information about the possible chronic risks of existing and future food
use chemicals.133 The entire purpose of the bill, in the words of FAA-sponsor Congressman
James Delaney of New York, was to change a system that made consumers “the guinea pigs for
anything that comes on the market.”134 He went on to remark at the congressional hearings on
the bill, “I say that we are probably playing Russian roulette with some of these things. We just
do not know. This bill simply provides for ordinary testing or pretesting before these new and
untested chemicals are added to our food supply.”135

132

Arlington, 133 S.Ct. at 1868.
H. Rep. No. 82-2356, at 27 (1952).
134
Chemical Additives in Food: Hearings on H.R. 4475 Before the Subcomm. On Health and Sci. of the
House Comm. On Interstate and Foreign Commerce, 84th Cong., 2d Sess. 30 (1956). One of the main
purposes of the legislation, according to the Report accompanying the House Bill, was “to protect the
health of consumers by requiring manufacturers of food additives and food processors to pretest any
potentially unsafe substances which are to be added to food.” H. Rep. No. 85-2284 (July 28, 1958). The
second purpose was to “advance food technology by permitting the use of food additives at safe levels.”
Id. This was a response to the previously existing law, which “entirely prohibit[ed] the use of these
additives [even] at safe levels.” Id. While this purpose promoted the use of additives, to a degree, it was
only the safe use thereof, as policed by adequate premarket assessment.
135
Id. at 30 (1956). The select committee headed by Delaney had undertaken an inventory of chemicals
then in use, and the legislation was intended to address the gaps in safety assessments. As Delaney noted,
“there are approximately 276 chemicals being used in food today the safety of which has not been
established to the satisfaction of many groups concerned with the health and safety of the public.” Id. at
29. In addition, concerns were being raised about “methods of processing” foods and the “chemical
changes” they incurred in substances. Id. The need for a comprehensive regulatory scheme to address
the gaps was emphasized throughout the hearings. For example, Rep. Delaney, who led a Congressional
investigation of the use of chemicals in foods by a select committee in
the 81st and 82nd Sessions of Congress, summarized the results of the select committee’s investigations
for the record at the hearing, noting that “…the committee believed that the public is in need of protection
against some irresponsible elements, as well as against the possible inadvertent mistakes of reputable food
processors and the premature enthusiasms of chemical manufacturers.” Id. at 27.
133

25

 Before the FAA was passed, FDA had only weak, post-market regulatory authority over
chemicals added to food under the FDCA. Section 402(a) of the FDCA provided that a food
shall be deemed adulterated if it “bears or contains any poisonous or deleterious substance which
may render it injurious to health.”136 To remove an additive from the market, FDA had to prove
that a substance was poisonous or deleterious and could render the food injurious to health.
Because of the substantial scientific support it would require to sustain such a finding, FDA
could enforce this provision only after consumers were already exposed to its risks.137 It could
take years for FDA to obtain the requisite proof, during which time it could not keep the
substance off the market or protect consumers.138
The FAA was thus passed specifically to shift this burden from the agency to the industry
– the new system “would put upon the processor rather than our Government the burden of
proving that a newly discovered substance which a processor of foodstuffs proposes to add to the
food we eat is safe.”139 In addition, it would replace the existing system with a premarket review
process overseen by FDA. Under the statute, a food additive could not be used unless and until
the agency deemed it safe for the use proposed by the sponsor.140 When reviewed in this
context, it is clear that the GRAS exemption was not intended to upset this balance; rather, it
provided additional flexibility so that FDA would retain authority over novel and risky
substances, while allowing substances with an undisputed and generally known record of safety
to be quickly cleared of concern and used.141
FDA’s conception of GRAS subverts the objectives of the FAA. Under the current
system, almost all chemical additives are self-determined GRAS by the food industry and
thereby bypass the food additive petition and premarket review process established by the
FAA.142 Once a substance is determined to be GRAS by industry, it may immediately be
marketed in foods without FDA’s approval or knowledge. In practice, FDA takes the position

136

21 U.S.C. § 342(a).
H. Rep. No. 85-2284, at 1 (1958) (“To prove an untested substance poisonous or deleterious may
require approximately 2 years or more of laboratory experiments with small animals and during this
period the Government cannot prevent the use of such a substance in food.”).
138
Id.
139
S. Rep. No. 85-2422, at 3 (1958). Despite numerous bills on the topic, the chairman of the Health and
Science Subcommittee opened the hearings in the House by stating, “There is basic agreement among
proponents of this legislation with regard to the need for additional legislation. The industries
concerned – those who manufacture the chemicals in question and those who use them in connection with
food products – and the Government agencies are in agreement that an advance determination by the
Government as to the acceptability of a chemical in connection with foods is desirable from the point of
view both of the industries concern, the Government and the consumer.” Chemical Additives in Food:
Hearings on H.R. 4475 Before the Subcomm. On Health and Sci. of the House Comm. On Interstate and
Foreign Commerce, 84th Cong., 2d Sess. 1-2 (1956).
140
S. Rep. No. 85-2422, at 2 (1958) (“The processor who wants to add a new and unproven additive
[must] accept the responsibility . . . of first proving it to be safe for ingestion by human beings.”).
141
Beyranevand, supra note 26 at 895.
142
Pew Charitable Trusts, supra note 4, at 1; Post Graphic at p. 21 supra.
137

26

 that it then must demonstrate the substance is not safe in order to revoke its GRAS status, and
any safety review is postmarket.143
Thus, the regulatory process FDA invented in its 1997 proposal for GRAS substances
returns us, practically speaking, to a pre-FAA world – a world in which the burden of proof was
on FDA and safety review occurred after substances were already marketed to consumers. The
fact that FDA has allowed this shift in the burden of proof144 violates the FAA, which was
expressly intended to shift the burden from FDA onto additive manufacturers to prove a
substance is safe.145 For this reason, the GRAS system is “inconsisten[t] with the design and
structure of the statute as a whole,” and is therefore an illegal interpretation of the FAA.146
Far from furthering the statutory purpose of the FAA, FDA’s GRAS system does
essentially nothing to protect the public from unsafe substances added to foods. The FAA, by its
plain meaning and structure, was meant to create a regulatory system to ensure the safety of
substances added to foods through required premarket testing and expanded FDA oversight.
However, under the 1997 proposal, the voluntary GRAS notification system places FDA in a

143

Toni Clarke, FDA Considering Revamping Food Additive Rules Amid Growing Safety Concerns,
Huffington Post, Aug. 7, 2013, http://www.huffingtonpost.com/2013/05/07/food-additives-rules-fdasafety-concerns_n_3229755.html (“Our system really puts the onus on us to prove harm," FDA
Commissioner Margaret Hamburg said at the Reuters Health Summit in New York. "It's perhaps a time to
look at what the legal framework looks like and what opportunities there are now to ask and answer
questions in new ways because of advances in science and technology.").
144
This is at least how it has been interpreted by industry. See, e.g., Comment of Grocery Manufacturers
of America re: Tentative Determination Regarding Partially Hydrogrenated Oils; Request for Comments
and for Scientific Data and Information (March 2014), at 7 (“[To revoke GRAS status,] FDA bears the
burden of demonstrating that there exists a severe conflict among experts with respect the safety of the
challenged use of a substance.”)
145
See S. Rep. No. 85-2422, at 3 (1958). Indeed, another clear theme of the hearings was that the law was
intended to shift the burden of proof regarding the safety of additives to the regulated industry and off the
FDA. Rep. Dies asked Rep. Delaney: “This would shift the burden to the industries to have the pretesting
done?” Rep. Delaney responded, “At least to show that they have made tests to prove that they are not
toxic, to the satisfaction of the Food and Drug Administration.” Chemical Additives in Food: Hearings
on H.R. 4475 Before the Subcomm. On Health and Sci. of the House Comm. On Interstate and Foreign
Commerce, 84th Cong., 2d Sess. 30 (1956). Rep. Delaney also quoted from the findings of the select
committee regarding the views of a food industry representative, Dr. R.C. Newton, Vice President of
Swift & Co., who, citing the risks to the reputations of companies from a “disaster,” noted that the current
burden of proof rested with FDA, and stated that “[a] simple change in the law to place the burden of
proof on the prospective user, for example, of the chemical, is essentially all that is needed.” The select
committee’s investigation headed by Rep. Delaney emphasized this overarching purpose as well in its
conclusion, proposing “that the public is entitled to greater protection with respect to foods it must
necessarily consume; and that such protection is not afforded by existing legislation under which the
Government may take no action until after the food has been placed upon the market and injury may have
occurred.” Id. at 31.
146
Nassar, 133 S. Ct. at 2528-29; see also Fed. Election Comm'n v. Democratic Senatorial Campaign
Comm., 454 U.S. 27, 32 (1981) (finding that “administrative constructions of the statute, whether reached
by adjudication or by rule-making,” must be rejected where they are “inconsistent with the statutory
mandate or . . . frustrate the policy that Congress sought to implement.”)

27

 quasi-regulatory role that Congress never intended.147 While the system allows industry to
submit notifications to FDA, the agency does not accord its own action any regulatory or
clearance status for the safe use of the substance.
Where FDA does raise real safety questions, manufacturers are permitted to withdraw
their notifications, and still may market the substances in question. The notification system also
gives producers little inducement to notify FDA in the case of potentially dangerous substances,
as it would invite regulatory scrutiny without much benefit for industry (as FDA never actually
attests to a determination of GRAS).148 Thus, the 1997 proposal actually decreases FDA
oversight over the most critical decisions impacting public safety, frustrating the “policy that
Congress sought to implement.”149 It therefore does not merit deference.
2. FDA’s current interpretation of GRAS is unreasonable in light
of the purpose of the FAA.
FDA’s current expansive interpretation of GRAS, which has overtaken the “food
additive” definition under the law, is plainly unreasonable in light of the purpose and history of
the FAA (as well as the statutory language, as discussed above). The legislative history indicates
that Congress intended the GRAS exemption to be used only where the probable consumption of
a substance was widely known and its safety publicly and clearly understood. FDA’s current
interpretation of GRAS directly conflicts with this intent.
The FAA was enacted in response to a 1952 Congressional Committee Report that
recommended amending the FDCA to require that chemicals used in foods be subject to the
same safety standards as those required for new drugs.150 The Committee Report found that only
about half of the 840 chemicals used in food were considered safe, and concluded that the
existing law failed to provide adequate safety assurances for food additives.151 At that time, new
drugs required premarket testing and FDA approval, while new food additives did not.152 The
FAA was thus passed to “prohibit the use in food of additives which have not been adequately
tested to establish their safety.”153 Representative Delaney specifically stated, when describing
the scope of the law, “I am talking of new and unused and untested chemicals and this bill
requires, to get down to the basic point, that when a chemical is added there be sufficient testing
In the words of James O’Reilly, a professor of food and drug law at the University of Cincinatti
College of Law, this system amounts to “a balloon in the shape of a regulator. We have the appearance
that there’s a regulator protecting us, but there’s not.” Quinn & Young, supra note 66.
148
Compare IOM Workshop Summary, supra note 28, at 3 (describing GAP process as the “primary
mechanism for manufacturers to protect themselves from FDA enforcement actions), with 62 Fed. Reg. at
18,950 (explaining that under proposed notification system, “FDA would not be in a position to affirm a
notifier’s conclusion that a use of a substance is GRAS”).
149
Fed. Election Comm'n, 454 U.S. at 32; see also Util. Air Regulatory Grp. v. E.P.A., 134 S. Ct. at 2442
(2014) (quoting Nassar, 133 S.Ct. at 2529) (agency interpretation “’inconsisten[t] with the design and
structure of the statute as a whole’ does not merit deference”).
150
H. Rep. No. 82-2356, at 27 (1952); Lars Noah & Richard A. Merrill, Starting from Scratch?:
Reinventing the Food Additive Approval Process, 78 B.U. L. Rev. 329, 336 (1998).
151
Beyranevand, supra note 26, at 893.
152
Noah & Merrill, supra note 149, at 336–37.
153
Pub. L. No. 85-518, 72 Stat. 1784 (1958).
147

28

 so that the Food and Drug Administration is satisfied that it does not contain elements that are
not harmful to the human system.”154
Recognizing that it would be impractical to require pre-market evaluation for innocuous
substances that had long been used in foods or were well known to be safe, Congress included an
exemption for substances with a well-recognized track record of safety. The law therefore
required premarket evaluation for safety with respect to additives that were not “generally
recognized among competent experts as having been adequately shown to be safe under the
conditions of their intended use,” or “generally recognized as safe” (“GRAS”).155
The GRAS exemption was narrowly tailored to achieve the aforementioned goal: it did
not exempt all substances in current use, as this would not have achieved the purpose of ensuring
the safety of all additives used in foods. Instead, it was understood at the time that the pretesting
requirement would deal “mostly with new additives and more occasionally with old ones” and
that the status of any substance – including GRAS substances – could always change in light of
updated science.156 Through the FAA, Congress sought to establish a comprehensive, farreaching regulatory scheme overseeing all substances added to foods – rather than applying only
to a small subset of additives. In other words, contrary to FDA’s current application, GRAS was
truly intended to be an exception to the FAP process, not the primary practice.
The legislative history also indicates that GRAS was meant to have a narrow
construction. The “generally recognized” clause was copied verbatim from the new-drug law
provision already in force under the original FDCA.157 Under the FDCA, the term “new drug”
means “any drug . . . the composition of which is such that such drug is not generally recognized,
among experts qualified by scientific training and experience to evaluate the safety and
effectiveness of drugs, as safe and effective for use under the conditions prescribed,
recommended, or suggested in the labeling thereof.”158 Any manufacturer of a new drug – one
not “generally recognized . . . as safe” – must file an application for approval to FDA before
delivering or distributing the drug in interstate commerce.159
In the new-drug context, the “generally recognized . . . as safe” language has never been
interpreted to allow industry-made safety determinations for novel pharmaceuticals – and had not
been at the time of the FAA’s enactment. The FAA’s use of the FDCA’s new drug language is a
clear indication that the GRAS exemption for food additives was intended by Congress to
154

Chemical Additives in Food: Hearings on H.R. 4475 Before the Subcomm. On Health and Sci. of the
House Comm. On Interstate and Foreign Commerce, 84th Congress, 2d Sess. 28 (1956).
155
21 U.S.C. § 321(s).
156
Charles Wesley Dunn, Initial Comments—for Food Law Institute, 13 Food Drug Cosm. L.J. 743, 744
(1958); 104 Cong. Record, 85th Cong, 2d Sess. 17412, 17420 (Aug. 13, 1958) (Letter from John L.
Harvey, Deputy Commissioner of FDA) (explaining that later scientific developments could always
change GRAS status to food additive status).
157
The language is also used in the Pesticide Residues Amendment of 1954. Pub. L. No. 83-518, 68 Stat.
511 (1954) (codified at 21 U.S.C. 346a(a)), repealed and replaced by Food Quality Protection Act of
1996, Pub. L. No. 104-170, 11 Stat. 1489 (1996).
158
21 U.S.C.A. § 321.
159
21 U.S.C.A. § 355.

29

 function similarly.160 Indeed, the new-drug language is repeatedly referenced throughout the
Congressional hearings on the FAA. For example, when discussing the GRAS exemption at the
Congressional hearings, a representative from the Grocery Manufacturer of America noted that
the vagueness of the GRAS language would not be problematic because “as the new drug section
experience suggests, unless the one using the ingredient is absolutely certain, the risk of criminal
penalties will result in his asking the FDA whether it thinks a new ingredient falls within the act
or is general[ly] recognized as safe.”161 This suggests that, in the new-drug context, the industry
would decline to invoke the GRAS exemption except in the very clearest circumstances for fear
of FDA’s strict enforcement.
The Congressional hearings on the FAA indicate that Congress assumed that the same
process would occur in the food additive context.162 However, because FDA enforcement in the
food additive space is virtually nil (due to inadequate resources, a self-created lack of
information about industry practices on GRAS substances, and FDA’s failure to hold GRAS
notifications to the safety standards applied for food additives), the GRAS exemption for food
additives has never been limited by industry fears of prosecution in the way it was and always
has been for new drugs.163 Instead of noticing this critical distinction in practice and addressing
it by narrowing the exemption appropriately to reflect Congress’s intent, FDA has ignored the
difference, and even worsened it, through its failure to enforce the law with penalties for
violators or to adequately police the substances permitted to be self-declared as GRAS by
industry. In the food area, in stark contrast to drugs, FDA’s interpretation of the GRAS
exemption has blown a hole in the law sufficient to undermine its clearly stated intent.
That GRAS was intended to be narrowly construed is amply supported by testimony from
the congressional hearings. During the Congressional hearings on the bill, a statement by
William Goodrich, then-Assistant General Counsel for FDA, regarding the “generally
recognized” language, demonstrates the intended scope of the GRAS exemption:

160

Chemical Additives in Food: Hearings on H.R. 4475 Before the Subcomm. On Health and Sci. of the
House Comm. On Interstate and Foreign Commerce, 84th Congress, 2d Sess. 203-205 (1956). (George P.
Larrick, Commissioner of FDA, describing agency’s experience applying GRAS exception for drugs).
161
Food Additives: Hearings Before the Subcomm. On Health and Sci. of the House Comm. On Interstate
and Foreign Commerce, 85th Cong. 213 (1957) (Testimony of Charles Wesley Dunn, on behalf of GMA).
162
Id.
163
FDA’s policing of GRAS substances is exceedingly rare. While a procedure exists for citizens to
petition the removal of an additive’s GRAS status, this process is also nonfunctional; of the 11 citizen
petitions submitted to the agency between 2004 and 2008, FDA has decided on the validity of these
concerns in only one case. Nine of these 11 petitions raised specific concerns about the safety of GRAS
substances. For example, a 2006 petition cited studies linking diacetyl (a GRAS substance used in
microwave popcorn) to severe respiratory reactions. Nine years later, FDA has still not responded to this
petition, and diacetyl remains both “GRAS” and on the market. GAO Report, supra note 37, at 23. The
remaining citizen petitions involved milk protein concentrate, partially hydrogenated vegetable oils,
aluminum-based food additives, salt, carbon monoxide gas in fresh meat packaging, carbon monoxide gas
in fresh tuna packaging, diacetyl, iodized salt, monosodium glutamate, carrageenan and similar
substances, and stevia extracts. Id.

30

 Mr. GOODRICH.

The point of the general language is to avoid pretesting of things
like table salt. We have to have the law like that.

Mr. DIES.

He would be perfectly safe in putting out table salt.

Mr. GOODRICH.

Because it is generally recognized.164

At the same hearings, then-Commissioner of the FDA George P. Larrick testified similarly:
We believe only those chemicals should be automatically exempted from the new
law which are recognized among competent experts as safe for their intended use.
This would make it unnecessary, for example, to do studies on table salt, but
would not approve the continued use, without scientific proof of safety, of the
synthetic emulsifiers now widely used in some fabricated foods.165
The testimony at the hearings clearly shows that the GRAS exemption was meant to be
limited to substances considered at that time to be both safe and uncontroversial. The hearings
also indicate clearly that novel chemical additives could never be GRAS because they have not
been used for an adequate amount of time for their safety to be generally recognized. For
example, a representative from the adhesive industry testified that, “it is certainly true that no
new substance can qualify as being used to a material extent or for a material length of time,
which is the ordinary or usual connotation of the words ‘generally recognized.’”166 In other
words, novel chemical additives could never be GRAS because their very novelty makes a broad
scientific consensus impossible to achieve.
The food industry’s own interpretation of their legal obligations under the Act indicates
that they also shared this understanding of the statutory language. At the hearings, a
representative from several major food industry groups (including the American Bakers
Association, the Grocery Manufacturers Association, and the National Restaurant Association)
declared industry’s support for the bill and issued the following principles:


We believe every substance not represented by long usage in the human diet
should be subject to question as an ingredient in food, and that this question

164

Chemical Additives in Food: Hearings on H.R. 4475 Before the Subcomm. On Health and Sci. of the
House Comm. On Interstate and Foreign Commerce, 84th Cong., 2nd Sess. 227 (1956) (Testimony of
William Goodrich, Assistant General Counsel for FDA). Note that the safety of salt was then considered
uncontroversial; yet sodium in the diet may not be GRAS for all current uses. See Petition to Revoke the
GRAS Status of Salt, submitted by CSPI to FDA Nov. 8, 2005, available at
http://www.cspinet.org/new/pdf/fda_salt_petition.pdf. This highlights the need for FDA’s active
reassessment of the safety of additives as required under the law.
165
Chemical Additives in Food: Hearings on H.R. 4475 Before the Subcomm. On Health and Sci. of the
House Comm. On Interstate and Foreign Commerce, 84th Cong., 2nd Sess. 197 (1956) (Testimony of
FDA Commissioner George P. Larrick).
166
Id. at 151 (Statement of John A. Gosnell, on behalf of the Adhesives Manufacturers Association of
America).

31

 


should be resolved by adequate animal experimentation to prove that its use in
food does not present a hazard to public health.
We believe every new substance proposed for use in human food should be
subjected to adequate pretesting by the manufacturer or user of the substance and
that such pretesting should be required by law.
We believe it to be a proper function of government to control those factors
which may affect adversely public health. Therefore, we believe the results of
animal experimentation in pretesting new substances proposed for use in food
should be reviewed and approved by the Food and Drug Administration before
the substance is allowed to be used in food sold to the public.167

Thus, even in the view of the regulated industry, the pretesting and FDA approval
requirements under the food additive petition process were intended to apply to all novel
chemical additives – or in their words, “substances not represented by long usage in the human
diet.” It is exceedingly clear from the legislative history that novel chemical additives were
never intended to fall within the GRAS exemption of the FAA.168 Indeed, FDA itself has more
recently recognized that novel additives warrant greater caution in toxicological testing than
known ones.169 The fact that the current system permits unknown and unproven substances to be
self-determined GRAS is an unreasonable interpretation of the statute in light of the original
purpose of the GRAS exemption and the overall purpose of the FAA.
3. “Secret GRAS” undermines the purpose of the FAA and
prevents the statute from working as Congress intended.
Finally, the legislative history makes clear that Congress, in passing the FAA, intended
FDA to monitor the long-term effects of additive exposure in order to identify and address less
obvious risks. However, the lack of transparency under “secret GRAS” makes it difficult or
impossible for FDA to perform this critical task. Such review is also hampered because FDA
lacks basic, comprehensive data on the public’s exposure to chemicals that would allow it to

167

Id. at 58 (Statement of Howard O. Hunter, President of the American Institute of Baking, Appearing
for Various Food Industry Associations) (emphasis added).
168
See Noah & Merrill, supra note 149, at 337 (finding that the legislative history of the FAA
demonstrates that “food processors and ingredient suppliers accepted the need for some method of
premarket FDA review of novel ingredients in processed food”).
169
U.S. Food and Drug Administration, Guidance for Industry: Summary Table of Recommended
Toxicological Testing for Additives Used in Foods (June 2006), at n.2,
http://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/ucm054658.h
tm#ftn2 (last accessed 3/3/15) (“Minimum toxicological testing recommended to support the safety of a
novel additive might include studies generally recommended for a Concern Level III additive, irrespective
of its chemical structure and exposure.”); see also U.S. Food & Drug Administration, Guidance for
Industry: Assessing the Effects of Significant Manufacturing Process Changes, Including Emerging
Technologies, on the Safety and Regulatory Status of Food Ingredients and Food Contact Substances,
Including Food Ingredients that Are Color Additives 14 (June 2014), available at
http://www.fda.gov/downloads/Cosmetics/GuidanceRegulation/GuidanceDocuments/UCM300927.pdf,
federal register notice at 79 Fed. Reg. 36,533 (June 27, 2014) [hereinafter “Nanotech Guidance”].

32

 connect trends in the food supply to trends in public health.170 Quite simply, FDA’s proposed
rule makes it impossible for the agency to do its job under the FAA, which clearly contemplated
ongoing FDA oversight and reassessment of all substances in food. The “secret GRAS”
permitted by the voluntary notification process is therefore unlawfully “incompatible” with “the
substance of Congress’ regulatory scheme,” and does not warrant Chevron deference.171
The legislative history clearly demonstrates that Congress intended FDA to be aware of
all substances added to foods, including those properly determined to be GRAS. Before 1958,
FDA could already police poisonous additives and adulterants. In passing the statute, Congress
was giving FDA increased control over additives in order to address less obvious or immediate
risks presented by such substances, such as potential carcinogenicity.172 In other words,
Congress was “not so much concerned with the acutely toxic compounds, whose harmfulness
can readily be detected, as with those chemicals which may produce harmful effects only after
being ingested for months or perhaps for years.”173 Instead, Congress was “most interested in . .
. the residuary toxicity that is found in many of the[] chemical additives” that may not be
“apparent [for] a day, or 2 days, or a week” after exposure.174
Moreover, the legislative history underscores that Congress meant for FDA to ensure the
safety of all substances – even those found initially to be safe that presented later problems or
substances that are adverse to health over the life of a human. As Congresswoman Leanor
Sullivan, of Missouri’s 3rd District, clarified during the hearings:
Mrs. SULLIVAN.

Am I correct that while the bill primarily deals with new chemical
additives, it does not preclude the Food and Drug Administration
from moving against existing additives if they have evidence to
question their safety?

Mr. HARRIS.

Yes; I will say that the gentlewoman is correct in her statement.175

Kindy, supra note 5 (quoting FDA’s Michael Taylor: “We do not know the volume of particular
chemicals that are going into the food supply so we can diagnose trends.” “We do not know what is going
on post-market.”).
171
FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 156 (2000).
172
See, e.g., Food Additives Amendment of 1958, P.L. 85-929, 52 Stat. 1041 § 409(c)(3)(A) (1958)
(“[N]o 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[.]”).
173
Chemical Additives in Food: Hearings on H.R. 4475 Before the Subcomm. On Health and Sci. of the
House Comm. On Interstate and Foreign Commerce, 84th Cong., 2nd Sess. 27 (1956) (Testimony of
Representative Delaney, quoting committee report).
174
Id. at 29.
175
104 Cong. Record, 85th Cong, 2d Sess. 17,412, 17,421 (Aug. 13, 1958).
170

33

 Representative Delaney similarly stated his view that, “if a chemical is harmful, even if it
has been used over a period of time, . . . it should be rejected.”176 Rather, “[t]he criterion should
always be: ‘Is this chemical safe for public consumption.’”177
At the same time, the FDA Deputy Commissioner made clear that the agency believed
even the most innocuous GRAS substances would be subject to its ongoing oversight. In a letter
to Congress, the Deputy Commissioner elucidated that the GRAS exemption was not a
grandfather clause; rather, “if later developments show that a substance previously considered
safe because of common use in food were, in fact, subject to question and not generally
recognized as safe, then the substance would become subject to the definition of a food additive
and would have to be cleared under the procedures of the proposed law.”178
The FAA thus intended to require ongoing FDA oversight over all chemicals added to
foods. GRAS status was never intended to be a permanent or true “grandfather clause,” as that
would defeat the food safety purpose of the statute.179 Instead, Congress was clear that the status
of any substance could always change in light of updated science. Indeed, Congress recognized
that while most additives are not chronically dangerous to consumers, it needed a system of
oversight that could ensure the safety of all substances, particularly those to which consumers are
exposed at low doses over long periods of time. Without such ongoing oversight, there would be
no way to identify latent effects linked to the food supply or their source. To perform such
oversight, as intended by Congress, FDA must be informed about what is in our food – and
therefore “secret GRAS” cannot possibly be a legal interpretation of the statute in light of its
overarching purpose.
FDA recognizes to this day that the FAA imposes an obligation on the agency to reassess
the GRAS status of substances when new evidence of harm is produced. Current FDA
regulations state that “[n]ew information may at any time require reconsideration of the GRAS
status of a food ingredient.”180 FDA recently reaffirmed this position in its tentative
determination regarding the GRAS status of PHOs, published in the Federal Register in 2013.

176

Chemical Additives in Food: Hearings on H.R. 4475 Before the Subcomm. On Health and Sci. of the
House Comm. On Interstate and Foreign Commerce, 84th Cong., 2nd Sess. 28 (1956).
177
Id. at 31.
178
104 Cong. Record, 85th Cong, 2d Sess. 17412, 17420 (Aug. 13, 1958) (Letter from John L. Harvey,
Deputy Commissioner of FDA). This contrasts with the prior-sanctioned clause, which functions more
like a true grandfather clause. In order to reverse a prior sanction, the FDA must meet the adulteration
standard under the FDCA. 21 C.F.R. § 181.1(b) (“Based upon scientific data or information that shows
that use of a prior-sanctioned food ingredient may be injurious to health, and thus in violation of section
402 of the Act, the Commissioner will establish or amend an applicable prior sanction regulation to
impose whatever limitations or conditions are necessary for the safe use of the ingredient, or to prohibit
use of the ingredient.”).
179
Frederick H. Degnan, Rethinking the Applicability and Usefulness of the GRAS Concept, 46 Food
Drug Cosm. L.J. 553, 554 (1991) (citing Chemical Additives in Food: Hearings on H.R. 4475 Before the
Subcomm. On Health and Sci. of the House Comm. On Interstate and Foreign Commerce, 84th Cong., 2d
Sess. 28 (1956)).
180
21 C.F.R. § 170.30(k)(1) (2012).

34

 There, the agency noted:
Importantly, the GRAS status of a specific use of a particular substance in food is
time-dependent. That is, as new scientific data and information develop about a
substance or the understanding of the consequences of consumption of a
substance evolves, expert opinion regarding the safety of a substance for a
particular use may change such that there is no longer a consensus that the
specific use is safe. The fact that the status of a substance under section 201(s) of
the FD&C Act may evolve over time is the underlying basis for FDA’s regulation
at § 170.38, which provides in part that FDA may, on its own initiative, propose
to determine that a substance is not GRAS. (See generally 36 FR 12093 (June 25,
1971) (issuance of 21 CFR 121.3, the predecessor of § 170.38)). Further, as stated
previously, history of the safe use of a substance in food prior to 1958 is not
sufficient to support continued GRAS status if new evidence demonstrates that
there is no longer expert consensus that an ingredient is safe.181
Although FDA acknowledges that it is required under the FAA to make an ongoing
review of GRAS substances to ensure they are still safe in the face of updated science and
exposure, its current GRAS proposal fails to create a system in which the agency is made aware
of all the substances it should be monitoring. Without knowing what is in our food, FDA cannot
perform the job given to it by Congress. Without information on substances and their dietary
exposure, FDA is incapable of measuring the more long-term, or more subtle, health effects that
led Congress to pass the FAA. Instead, FDA’s regulatory regime can only respond either to
acute risks or chronic risks with an obvious connection to a well-known source.182 By ensuring
that it is ignorant of the composition of the food supply, the agency has handicapped itself and its
policing of the safety of substances in food contrary to the original intent of the FAA.
As a result of this lack of information and oversight, FDA rarely actually reconsiders the
GRAS status of substances in light of new safety concerns – for example, the agency has never
revoked the GRAS status of any substance approved through the petition affirmation process, nor
has it retracted a “no questions” letter since 1997.183 This is likely in part because FDA, in its
181

Tentative Determination Regarding Partially Hydrogenated Oils; Request For Comments And For
Scientific Data And Information, 78 Fed. Reg. 67,169, 67,170 (Nov. 8, 2013) (emphasis added).
182
FDA acknowledged that chronic risks are notoriously difficult to connect to a source: “Currently, the
legal system does not ensure the optimum level of safety for foods because consumers who become ill
often do not know the reason for, or source of, their illness. Even in cases where consumers are aware that
their illness was contracted from a specific food, it is often difficult to determine who is ultimately
responsible for their illness, since the particular source of contamination is not known in many
circumstances.” U.S. Food and Drug Administration, Preliminary Regulatory Impact Analysis for the
proposed rules on Foreign Supplier Verification Programs (Docket No. FDA-2011-N-0143) and
Accreditation of Third-Party Auditors/Certification Bodies to Conduct Food Safety Audits and to Issue
Certifications (Docket No. FDA-2011-N-0146) under Executive Order 12866, Executive Order 13563,
the Regulatory Flexibility Act (5 U.S.C. 601-612), the Unfunded Mandates Reform Act of 1995 (Public
Law 104-4), and the Paperwork Reduction Act of 1995 (44 U.S.C. 3501-3520) 3-4 (2013), available at
http://www.fda.gov/downloads/AboutFDA/.../EconomicAnalyses/UCM363286.pdf.
183
GAO Report, supra note 37, at 22. This in spite of the fact that safety concerns have arisen regarding
many GRAS substances, such as parabens, salt, and sugar used in sugar-sweetened beverages.

35

 own words, “do[es] not know the volume of particular chemicals that are going into the food
supply so [it] can diagnose trends,” and it “do[es] not know what is going on post-market.”184
As a result, FDA lacks the data it would need to sensibly prioritize a public health approach and
assess overall exposure of particular chemicals or classes of chemicals – it does not have the
information it needs to form intelligent hypotheses about health trends and dietary exposure.185
Merely according final rule status to the proposal would similarly fail to enable the agency to
accomplish its statutory assignment to track the safety of substances on scientific or public health
grounds.186
Contrary to public expectations and its Congressional assignment to assure the safety of
the food supply, FDA has failed to meet its legal obligations under the FAA – and has proposed
continuing in that state, with weakened reporting and oversight. To perform its responsibilities
under the law, FDA must be informed of GRAS substances in use.187 Without information about
a chemical’s identity and toxicology, dietary exposure, and proposed use, FDA cannot assess its
health consequences over long periods of time – or even determine when and which substances
warrant review.188 Thus, so long as “secret GRAS” is allowed, FDA cannot meet its obligation
to ensure the safety of substances added to foods as required by the FAA.

184

Kindy, supra note 5 (quoting Michael Taylor).
The relevance of this is obvious. Post-market review cannot be sufficient when FDA has no means to
link public health trends to substances present in the food supply. Under the current system, FDA is
woefully unable to explore links between dietary exposure and outcomes. For example, recent research
points to the origins of childhood obesity as an outcome of the dynamic interplay of genetic, behavioral,
and environmental factors, with a compelling body of evidence suggesting that environmental exposures
affect a child’s risk of obesity. See Institute of Medicine, Examining a Developmental Approach to
Childhood Obesity: The Fetal and Early Childhood Years: Workshop in Brief 1 (March 2015), available
at http://www.nap.edu/openbook.php?record_id=21716. IOM, along with other partners, recently
convened a two-day workshop on the links between dietary exposure of additives and obesity, which
FDA did not participate in. See id. Without information about exposure and cumulative effects, FDA
cannot even begin to form appropriate questions about long-term exposure that Congress was so
concerned about when it passed the FAA in 1958 – and therefore fails to meet its obligations under the
statute.
186
For example, a recent study questioning the safety of emulsifiers – a substance with a very high
exposure in the food supply – illustrates the type of safety concern that Congress intended FDA monitor
when it passed the FAA. See Benoit Chassaing et al., Dietary emulsifiers impact the mouse gut
microbiota promoting colitis and metabolic syndrome, Nature 519, 92-96 (Feb. 2015),
http://www.nature.com/nature/journal/v519/n7541/full/nature14232.html.
187
Notably, both President Nixon and the FASEB committee presumed that the FAA permitted FDA to
conduct a comprehensive assessment of the safety of GRAS substances, as reflected by the
comprehensive review initiated during the 1970s. See supra at 8.
188
21 C.F.R. § 170.3(i). FDA told GAO that it maintains a database called the Priority-Based Assessment
of Food Additives (PAFA), containing administrative, chemical, and toxicological information about food
ingredients, including GRAS substances, in order to prioritize substances for assessment. However, FDA
could not provide any examples of a reconsideration of a GRAS substance derived from use of this
database. GAO Report, supra note 37, at 21. Regardless, this database is unlikely to include GRAS
substances that FDA has never been notified about, further illustrating the need for required notification,
as discussed infra.
185

36

 Commenters have noted that the legislative history of the FAA shows that “industry and
agency witnesses alike assumed that a manufacturer had the opportunity to make its own GRAS
determinations”189 – thus arguing that the FAA allows for the present “secret GRAS” system.
However, independent GRAS is not the same as secret GRAS. Even if independent GRAS
determinations are legal under the FAA, this does not mean that industry determinations totally
supplant FDA’s regulatory authority. FDA still has the responsibility, under the FDCA and the
FAA, to ensure the safety of all substances in the food supply. As a result, FDA has the
obligation to enact a regulatory scheme that allows it to meet its statutory mandate. As the
foregoing makes clear, FDA can only achieve its responsibilities when it is informed about the
composition of the food supply. Thus, the language of the FAA does not authorize independent
determinations to be made in secret or on insufficient evidence.
FDA has indicated that it does not believe it has the authority to require mandatory
notification of GRAS determinations.190 However, this is plainly not the case. Nowhere in the
statute does it say that FDA may not require such notification. Instead, the purpose of the statute
indicates that Congress intended FDA to have oversight over all substances added to foods. This
includes the authority to develop a regulatory program to ensure the purposes of the law are met;
in other words, “[FDA] is not merely an implementer of law, but an architect of the law as
well.”191
Again, while the language of the statute does not foreclose the possibility of independent
industry GRAS determinations, this does not usurp FDA of its authority (and obligation) to
oversee the safety of such substances. Indeed, FDA has frequently policed GRAS in the past, by
issuing regulations narrowing and defining its scope.192
In addition, in the 1997 proposal, it suggested that it would perform random audits in
order to ensure that GRAS determinations were made properly – a further indication that FDA
recognizes its own authority to regulate industry-made GRAS designations, and to assert
enforcement powers necessary to achieve its statutory and regulatory goals. It is unreasonable to
think that a program of random audits would be a sufficient or appropriate use of agency
resources where mere notification by industry would allow much more efficient allocation of
agency enforcement resources. Indeed, the secrecy of the current program means that the audits
would be truly random, and could not be targeted in any systematic way. Moreover, FDA and
the industry cannot meaningfully assess cumulative exposure and therefore the possible harm of
exposure. Mandatory notification is simply a more practical and efficient means of
accomplishing the same goal, and is therefore an implied authority under the FAA.

189

Degnan, Rethinking the Applicability and Usefulness of the Gras Concept, supra note 179, at 560.
GAO Report, supra note 37, at 19 (in response to GAO questions, “FDA officials stated that the
agency would have to seek authority from Congress in order to require all companies to inform it of their
GRAS determinations.”).
191
IOM Workshop Summary, supra note 28 at 4.
192
See, e.g., 23 Fed. Reg. at 9,512 (“In general, any substance added to a food which has no history of
common use as a food ingredient, should be regarded as a substance that is not generally recognized as
safe unless it has been scientifically tested and shown to be safe.”)
190

37

 FDA also has ample implied authority to address the statutory goals by requiring
mandatory notification. Section 701(a) of the FDCA vests the Secretary with the authority to
issue regulations for the efficient enforcement of the Act.193 Under this section, FDA may issue
regulations in order to “effectuate a congressional objective expressed elsewhere in the Act.”194
Even narrowly construed, this plainly gives FDA the authority to police the line between GRAS
and food additives to ensure that all food additives are subject to the statutory petition process
required in the Act.
Moreover, because a primary objective of the FDCA is the enhancement of public health,
the FDA’s rulemaking authority under Section 701(a) “has been broadly construed to uphold a
wide variety of assertions of regulatory power” and is valid so long as it is reasonably related to
an authorized regulatory objective.195 The objectives of the FAA in ensuring the safety of all
substances added to foods, codified as part of the FDCA, can only be achieved if exposure and
safety data are available and current for all substances added to foods. Thus, under 701(a), FDA
has the authority to require notifications, as a comprehensive inventory of substances added to
foods is an essential predicate to achieving its responsibilities under the FAA.
Indeed, courts interpreting 701(a) have evaluated practical enforcement considerations to
determine their role in carrying out the statutory scheme.196 In this case, FDA can only fulfill its
statutory obligations under the FAA if it is informed about usage and exposure of all substances
added to foods – thus, required notification would plainly be proper in order to “effectuate a
congressional objective” under the FAA. In fact, the lack of notification actually prevents
effective enforcement of FDA’s statutory obligations to consider exposure and cumulative
effects in making safety determinations. To keep such data current, companies must notify FDA
(and thereby each other) of their use of such substances.
Thus, FDA has both the authority – and a clear obligation – to issue regulations requiring
public and ongoing notification of all GRAS determinations.197 FDA’s program permitting
21 U.S.C. 371(a) (“The authority to promulgate regulations for the efficient enforcement of this
chapter, except as otherwise provided in this section, is vested in the Secretary.”).
194
Association of American Physicians and Surgeons, Inc. v FDA, 226 F. Supp. 2d 204 (D.D.C. 2002)
(citing Pharm. Mfrs. Ass'n. v FDA, 484 F. Sup. 1179, 1183 (D. Del. 1980) (upholding FDA regulation
requiring doctors to distribute information about estrogen-containing drugs to patients to whom the drugs
were prescribed because it was related to FDCA objective that consumers receive facts material to
consequences that may result from pharmaceuticals).
195
See Pharm. Mfrs. Ass’n, 484 F. Supp. at 1183; see also U.S. v. Nova Scotia Food Products, 568 F.2d
240, 246 (2d Cir. 1977) (finding that enforcement-related regulations are valid so long as they are
reasonably related to the purposes of the enabling legislation). In addition, in its final rule implementing
the seafood HACCP requirements, FDA broadly interpreted its authority under 701(a) as allowing it to
require record maintenance and access as effectuating a congressional objective to ensure safe conditions
in processing. See Degnan, FDA’s Creative Application of the Law, supra note 51, at 60.
196
See Nat’l Confectioners Ass’n v. Califano, 569 F. 2d 690, 693 (D.D.C. 1978) (finding that encoding
and recordkeeping requirements appropriately assisted the FDA in fulfilling its statutory mandate to
provide remedies for any adulterated foods present in interstate commerce, such as through product
recalls).
197
The FDA recently recognized such an authority when it included recordkeeping requirements for
recipes under the nutrition facts labeling proposed rulemaking. In the proposal, FDA found that “[t]he
193

38

 “secret GRAS” is therefore unreasonable in light of the statute as a whole, and does not merit
deference. Even if a court were to conclude that the agency’s prior interpretation of the law had
once been permissible, upon this record, as demonstrated below, it is no longer a reasonable
interpretation by FDA of its obligations.198
Congress’s purpose in passing the FAA was clear: to establish a comprehensive
regulatory scheme to protect the health of consumers from potentially unsafe substances added to
foods. The 1997 proposal, which all but eliminates the last shreds of FDA oversight, cannot
possibly achieve the objectives of the statute. Thus, in light of the purpose of the statute and
FDA’s responsibility to oversee the safety of all substances, the agency’s interpretation of GRAS
is unlawfully “inconsistent with the statutory mandate” laid out in the FAA, and does not merit
deference under Chevron.199
b. If FDA were to finalize the 1997 rule as proposed, it would be “arbitrary and
capricious” under the Administrative Procedure Act.
Finally, if FDA were to finalize the 1997 rule as proposed, this would be unlawfully
arbitrary and capricious under the Administrative Procedure Act. An agency rule is “arbitrary
and capricious” if an agency has “relied on factors which Congress has not intended it to
consider” or “entirely failed to consider an important aspect of the problem.”200 In this case, the
proposed rule has been in effect for nearly 20 years and FDA is fully aware that it is not
functioning as the agency intended, much less achieving the purposes of the FAA. Thus, if FDA
were to finalize the rule unchanged, it would be arbitrary and capricious due to a clear failure to
“consider an important aspect of the problem” under State Farm.201
The FDA’s rationale for creating the GRAS notification process was that the onerous
GAP process was discouraging individuals from requesting affirmation of self-determined
GRAS status.202 The agency predicted that a simpler process would provide incentive for
manufacturers to inform FDA of their GRAS determinations, resulting in “increased agency
awareness of the nation’s food supply and the cumulative dietary exposure to GRAS
authority granted to FDA under sections 701(a), 403(q), 403(a)(1) and 201(n) of the FD Act not only
includes authority to establish records requirements, but also includes access to such records. Without
such authority, the nutrient declarations for these specific nutrients that FDA has determined are
necessary to assist consumers in maintaining healthy dietary practices under section 403(q)(2)(A) of the
FD Act are, practically speaking, not enforceable.” Food Labeling: Revision of the Nutrition and
Supplement Facts Labels, 79 Fed. Reg. 11,880, 11,884 (2014).
198
The fact that FDA has long misinterpreted its obligations under the FAA does not change this, as the
Supreme Court has made clear that even “longstanding agency interpretations must fail to the extent they
conflict with statutory language.” Pub. Employees Ret. Sys. of Ohio v. Betts, 492 U.S. 158, 171, 109 S.
Ct. 2854, 2863, 106 L. Ed. 2d 134 (1989). Moreover, the 1974 Rulemaking and prior actions by the
Nixon Administration clearly demonstrate that FDA long viewed its authority over GRAS as permitting,
perhaps even necessitating, the agency to compile a comprehensive inventory of all GRAS substances.
199
Fed. Election Comm'n., 454 U.S. at 32 (explaining that courts must reject administrative rulemakings
“that are inconsistent with the statutory mandate or frustrate the policy that Congress sought to
implement”).
200
Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983).
201
Id. at 43-44.
202
62 Fed. Reg. at 18,941.

39

 substances.”203 FDA also predicted that the “streamlined” notification process would allow it to
redirect its resources to GRAS substances “that are a priority with respect to public health
protection” and to issuing industry guidance on food safety issues. Finally, it stated that because
both the GAP and notification procedures were voluntary, the substitution would be “neutral”
from a legal and regulatory perspective.204
The interim system has been in force since 1997 and ample evidence demonstrates that it
has not performed as the agency intended. While FDA predicted that the new system would
encourage greater reporting to the agency, it overlooked a crucial fact: the notification system
actually gives industry less inducement to notify FDA in the case of potentially dangerous
substances, because notice would invite regulatory scrutiny without concrete benefit for industry
(because FDA never actually attests to a determination of GRAS, as it did under the GAP
process).
At the same time, the interim rules have led to a dramatic shift away from the food
additive petition process, which has been functionally abandoned since the notification process
debuted. Thus, it is simply not the case that the 1997 proposal is “neutral.” Instead, FDA has
taken a program with considerable propensity to tolerate abuse and made abuse the norm, the
very “opposite of what the oversight law intended.”205
Before the notification system went into effect, independent GRAS existed, but occurred
only “occasionally.” Even in such cases, manufacturers would reportedly commission semipublic safety reviews by reputable scientific organizations.206 But rather than improving this
situation, the 1997 proposal has led to decreased transparency in the underlying science. As
FDA’s Michael Taylor noted, “the law was meant to increase public scrutiny of additive safety
by encouraging companies to publish their science in academic journals.”207 Instead, since 1997,
the bulk of GRAS determinations have been made by expert panels without peer-reviewed
publication.208

203

Id.
Id. at 18,951.
205
Kindy, supra note 5 (quoting Michael Taylor).
206
IOM Workshop Summary, supra note 28, at 24; see id. at 25 (describing private GRAS reviews as
“occasional”). In addition, the FEMA process appears to have changed significantly in the wake of the
1997 proposal. Originally, FEMA contracted with FDA to publish safety data underlying GRAS
determinations for flavorings. These reviews were published on a period basis during the 1970s and
formed the basis for the GRAS affirmation (GRASa) Program, which operated from 1975-1985.
Following the GRASa Program, there was an extreme decline in GRAS determinations for flavorings;
however, in the wake of the 1997 proposal, “there has been a significant increase in the number of
substances GRASed annually” by FEMA. See R.L. Smith et al., GRAS Flavoring Substances 24, Food
Technology 46-48 (June 2009).
207
Kindy, supra note 5.
208
Kahl Memo, supra note 67, at 5 (“In practice, approximately one half of the GRAS notices filed by
OFAS during the interim period included the findings of a panel of individuals who evaluated the data
and information that were the basis for the notifier’s GRAS determination. In most cases, these findings
were included as a report in the GRAS notice but were not published in peer-reviewed journals. In most
cases, OFAS responded to a GRAS notice containing the findings of a GRAS panel with a “no questions
letter”).
204

40

 Thus, it is clear under the 1997 proposal, FDA has failed to increase its awareness of
substances in the food system, their safety, and their dietary exposure.209 While the rate of
GRAS notifications is higher than that of GAP submissions, FDA remains totally unaware of an
estimated 1,000 substances currently added to foods. Because FDA is not notified about a
majority of GRAS substances nor their dietary exposure, the agency cannot rationally prioritize
its resources to address substances of public health concern (another purported justification in the
proposal). To finalize the proposed rule when it has not realized FDA’s justifications for
proposing it, and has in fact resulted in changes for the worse, would be indubitably “arbitrary
and capricious.”
There is also considerable evidence before the agency, as detailed in the next section,
about the failures of the system under the proposal and why it does not achieve the purposes of
the FAA. These failures were also raised by comments to the docket when the comment period
was reopened in 2010, over ten years after the interim system went into effect, and by an internal
review of FDA’s own experience with the program.210 It is also evidenced by the myriad
examples, provided below, of “GRAS” substances that are, properly considered, a food additive
– whether because they are novel, of questionable safety, or because their safety is not generally
recognized – and by the high number of GRAS notifications that are withdrawn before FDA can
publish its concerns regarding the industry’s notice. The fact that “food additives” and
substances of questionable safety are regularly introduced into the food supply via the GRAS
system – rather than the scheme statutorily mandated by Congress – demonstrates that the
proposed rule does not meet the requirements of the FAA. Instead, as explained above, the 1997
proposal undercuts the statutory objectives of the Act.
In sum, if FDA were to finalize the current proposed rule without revision, it would
disregard the fact that the system has prevented the statutory food additive petition process from
working as intended, has eviscerated the definitions of “safety” and “general recognition” thereof
under the statute, and has prevented effective FDA oversight of both food additives and GRAS
substances. It would thus plainly be “arbitrary and capricious” in violation of the Administrative
Procedure Act.
IV. Previous comments and studies in this docket demonstrate that the proposed
notification system does not achieve the purposes of the FAA.
The serious defects in the process described above were raised by an in-depth 2010 GAO
Report, in six peer-reviewed studies performed by the Pew Charitable Trusts,211 other reports by
Pew, and in the prior comments on this proposed rulemaking; these reports are in the docket and
FDA is well aware of them.212

GAO Report, supra note 37, at 13 ([W]ithout overseeing all companies’ GRAS determinations, FDA
is less informed about the nation’s food supply and consumers’ cumulative dietary exposure to GRAS
substances – both of which were viewed by FDA as beneficial potential outcomes of its 1997 proposal.”).
210
See generally Kahl Memo, supra note 67.
211
Four of which were published in the journal for the food industry’s professional society, the Institute
of Food Technologists.
212
Indeed, FDA reopened the docket in 2010 in response to the GAO report.
209

41

 a. A Key GAO Report found that FDA’s current GRAS system fails to ensure
safety of substances used in food as required by the FAA.
In 2010, GAO reviewed the GRAS program to determine the extent to which FDA’s
oversight of new GRAS determinations helps ensure the safety of these substances and the
continued safety of existing GRAS substances.213 The results of GAO’s review were
unambiguously abysmal. GAO found that FDA’s GRAS process failed to meet the purposes of
the FAA in the same respects that we highlight in these comments; specifically, that FDA’s
oversight process “does not help ensure the safety of all new GRAS determinations,” and that the
agency is not “systematically ensuring the continued safety of current GRAS substances” – the
central purpose of the FAA.214
GAO found FDA’s oversight process deficient because under the voluntary notification
system, FDA reviews only the GRAS determinations that companies choose to submit. Thus,
FDA’s role is limited because “the agency generally does not have information about other
GRAS determinations companies have made [as] companies are not required to inform FDA of
them.”215 Because notification is not required, industry applies an expansive interpretation of
GRAS that allows many food additives to avoid the regulation required by the FAA. GAO
expressed a particular concern about the safety of imported food products, noting that GRAS
substances may be manufactured anywhere in the world, and companies need not identify this
information as part of their GRAS determinations.216 The report also noted that this lack of
transparency makes FDA less informed about the nation’s food supply and cumulative dietary
exposure to GRAS substances. As a result, it is difficult for FDA to ensure the safety of GRAS
substances over time, a major concern of the legislators who passed the statute.217
Based on all of these deficiencies, GAO concluded that, “[b]ecause of the difficulty of
identifying GRAS substances as the source of food safety problems after they have entered the
food supply, FDA’s oversight of their safety would be improved if companies were required to
make the agency aware of their GRAS determinations.”218 GAO recommended that FDA
improve the system in many of the same ways that we recommend below, including that FDA
reassess the safety of GRAS substances in a more systematic manner.
GAO also found that FDA “has not taken certain steps that could help ensure the safety
of GRAS determinations,” such as issuing industry guidance on how to document GRAS
determination or monitoring companies to ensure that such determinations are made

213

GAO Report, supra note 37, at 1.
Id.
215
Id.
216
Id. at 13-14.
217
Id. at 12 (“Once a GRAS substance has entered the marketplace, FDA would find it difficult to
identify that substance as the potential source of a food safety problem, especially if FDA is unaware that
the substance has been determined to be GRAS. Food products may contain numerous ingredients,
including GRAS substances, making it difficult, if not impossible, for public health authorities to attribute
a food safety problem to a specific GRAS substance.”)
218
Id. at 12.
214

42

 appropriately.219 It recommended that FDA issue guidance on how to better document GRAS
determinations and develop a strategy to minimize the potential for conflicts of interest in
companies’ GRAS determinations. Last, it urged the agency to issue regulations requiring any
company that conducts a GRAS determination to provide FDA with basic information – as
defined by the agency to allow for adequate oversight – about this determination, such as the
substance's identity and intended uses, and to incorporate such information into relevant agency
databases and its public website.
b. The Pew Charitable Trusts’s studies found that FDA does not ensure the
safety of new and existing additives as required by statute.
In 2010, the Pew Charitable Trusts, a nonprofit, non-governmental organization,
launched a food additives project to evaluate the federal regulatory program overseeing food
additives.220 Its research, focused on the overall regulatory system, evaluated FDA’s ability to
fulfill its statutory obligation under the FAA to protect public health from unsafe chemicals
intentionally added to foods. The study involved experts from industry, academia, government
agencies, and public interest organizations. Like GAO, Pew concluded that the GRAS system is
plagued with systemic problems that prevent FDA from ensuring the use of food additives is
safe.221
Pew’s main critique was that, although FDA is supposed to make safety determinations
after considering cumulative effects of safety and exposure, it cannot do its statutorily required
job given that it has not been notified of: (1) an estimated 1,000 chemicals currently allowed in
food (secret GRAS); (2) actual usage for both these chemicals and the toxicologically related
chemicals it actually knows about; and (3) studies that suggest previously unknown health
effects.222 Because FDA lacks basic information needed to assess the safety of the thousands of
chemicals that have been cleared for use in food by industry, it only evaluates the safety of the
additives it happens to be aware of, through a program that is neither systematic nor scientific.223
Pew discovered that, as a result of this lax oversight, industry employs inadequate science
in its GRAS determinations. Pew found that most GRAS substances are not tested for safety in
accordance even with FDA’s most limited testing recommendations.224 For example, agency
guidance indicates that chemicals intentionally added to food should be fed to laboratory animals
to identify potential harmful effects, but Pew found that in the majority of cases, industrydetermined GRAS chemicals did not undergo any of this very basic type of testing.225 In
219

Id. at 1.
Pew Charitable Trusts, Food Additives Project, http://www.pewtrusts.org/en/archived-projects/foodadditives-project (last accessed Feb. 23, 2015).
221
Pew Charitable Trusts, supra note 93, at 6.
222
Id. at 7.
223
Id. at 8.
224
Id. at 9 n.72; Thomas G. Neltner et al., Data gaps in toxicity testing of chemicals allowed in food in the
United States, 42 Reproductive Toxicology 85, 88 (2013).
225
Neltner et al., supra note 224, at 90. In this report, Pew researchers found that less than 38% of FDAregulated additives have a published feeding study (the first test a scientist would do to evaluate the safety
of a chemical additive). For direct additives, only 21.6% have the feeding studies necessary to estimate a
220

43

 addition, Pew found that less than 38% of FDA-regulated additives have a published feeding
study.
Pew also found that FDA’s food additive (and GRAS) toxicology requirements were
severely out-of-date.226 According to FDA’s own scientists, advances in science have not been
incorporated into FDA’s practices, and the safety assessments currently used were developed in
the 1950s without sufficient updates.227 In the words of one FDA scientist, “[FDA] ha[s] not
tried to make [its] safety factor approach current with science.”228 Many of the scientific
deficiencies found in the 1982 report by FDA’s Select Committee on GRAS Substances
(SCOGS) persist, remaining unresolved to this day.229 In particular, Pew found that FDA’s food
additive and GRAS toxicology requirements were insufficient with regard to the following areas:







Behavioral Effects. FDA has failed to establish adequate testing standards to
assess the impact of additives on behavior. This puts the agency behind the U.S.
Environmental Protection Agency (EPA) and developed countries in the
Organization for Economic Cooperation and Development, which adopted such
methods and standards years ago.230
Endocrine systems. FDA has failed to identify standards and methods for how to
evaluate additives for potential endocrine disruption. Again, FDA is well behind
EPA in developing such practices.231
Vulnerable Subpopulations. FDA has failed to require consideration of additive
exposure to vulnerable subpopulations. Currently, its regulations only consider
infants. Pew recommends that this consideration be extended to pregnant women,
children, and those with hypersensitivity to certain substances, for example.
Absorption, distribution, metabolism, and excretion data. FDA currently
allows safety determinations to be made for chemicals without the data crucial to
understanding how humans process such chemicals, such as absorption,
distribution, metabolism, and excretion (ADME) data.

safe level of exposure, and only 6.7% have reproductive or developmental toxicity data in FDA’s
database. Id. at 88-90.
226
Maricel V. Maffini, Heather M. Alger, Erik D. Olson & Thomas G. Neltner, Looking Back to Look
Forward: A Review of FDA’s Food Additives Safety Assessment and Recommendations for Modernizing
its Program, 12 Comprehensive Reviews in Food Science and Food Safety 439, 449 (2013) (finding that
FDA’s food additives program has “not kept pace with the scientific developments of the last 20
y[ears]”). This is supported by FDA’s own recent internal review of the agency’s process for food
additives. See U.S. Food and Drug Administration, OFVM Report of the CFSAN Chemical Safety
Assessment Review Working Group, Appendix A at 11 (Apr. 2014).
227
See U.S. Food and Drug Administration, OFVM Report of the CFSAN Chemical Safety Assessment
Review Working Group, supra note 226, Appendix A at 11 (“Our conservative approach to food safety or
assessing the safety of food additives is also a weakness in a way because the safety factors that we rely
on when doing a safety assessment were developed in the early to mid 50’s. They are based on the science
and policy of that day. Since then they really have not been changed.”).
228
Id. at 12.
229
Maffini et al., supra note 226, at 449.
230
Pew Charitable Trusts, supra note 4, at 10; Maffini et al., supra note 226, at 450.
231
Pew Charitable Trusts, supra note 4, at 10; Maffini et al., supra note 226, at 450.

44

 Thus, Pew concluded that as a result of the minimal FDA oversight and scientific standards for
GRAS, these substances are not subject to adequate testing to ensure their safety for use in the
food supply.
Pew also found that the lack of FDA oversight results in rampant conflicts of interest
between industry and the experts making GRAS determinations. The organization reviewed the
industry’s GRAS notifications from 1997 to 2012 and found that “financial conflicts of interest
in these decisions are ubiquitous.” 232 The Pew report relied on criteria developed by the Institute
of Medicine in 2009, which identified a framework to assess the severity of conflicts of
interest.233 Applying this framework, Pew concluded that a manufacturer’s employee would
have the greatest likelihood of favoring a food substance as GRAS, while a member of a
standing expert panel selected by an independent third party would be least likely to do so.

Applying these standards to the 451 GRAS notices voluntarily submitted to the FDA
between 1997 and 2012, Pew found that 22.4% were made by an employee of an additive
manufacturer, 13.3% were made by an employee of a consulting firm selected by a
manufacturer, and 64.3% were made by an expert panel selected by the manufacturer or a
consultant to manufacturer.234 When manufacturers convened an expert panel, the panels were
small, typically only three people, all handpicked by the manufacturer or its hired consultant.235
Pew also found considerable overlap among the expert panels – at least 10 individuals
served on 27 or more panels, while one person served on 128 panels (of the total 290 panels
involved in GRAS determinations).236 In no case did a manufacturer use a standing expert panel
selected by an independent third party – the method Pew found least likely to involve a conflict
of interest short of FDA review. Thus, it is apparent that the current notification process does
little or nothing to curb conflicts of interest in GRAS determinations. In addition, there is no
basis to assume that the decisions withheld from agency review are better in terms of conflicts of
interest present in the determination process – in fact, they are likely worse.
Recognizing the significance of this problem, Pew drafted guidance on avoiding conflicts
of interest in GRAS evaluations and submitted it to the current docket on September 4, 2013.237
This guidance incorporated input from multiple reviewers and stakeholders in a public August
2013 workshop of experts, advocates and industry participants. Relying in part on FDA’s
Guidance for the Public, FDA Advisory Committee Members, and FDA Staff on Procedures for

232

Thomas G. Neltner et al., Conflicts of Interest in Approvals of Additives to Food Determined to Be
Generally Recognized as Safe, 173 JAMA Intern. Med. 2032, 2035 (2013).
233
Institute of Medicine (US) Committee on Conflict of Interest in Medical Research, Education, and
Practice, Conflict of Interest in Medical Research, Education and Practice (Bernard Lo & Marilyn Field,
eds. 2009).
234
Neltner et al., Conflicts of Interest, supra note 232, at 2034.
235
Id. at 2035.
236
Id. at 2035.
237
The Pew Charitable Trusts, Draft Guidance for Industry on Selection of Experts Conducting
“Generally Recognized as Safe” (GRAS) Evaluations, submitted to Docket No. FDA—1997—N—0020,
Substances Generally Recognized as Safe; Reopening of the Comment Period (Sept. 4, 2013).

45

 Determining Conflict of Interest and Eligibility for Participation in FDA Advisory Committees,
Pew recommended that:
1) Conflicts of interest should be disclosed to FDA for its review; and
2) An expert should not be permitted to conduct a GRAS evaluation if he or she would
be ineligible to serve on an FDA advisory committee without an agency waiver due to
disqualifying financial interests.
Since the FDA Guidance is primarily focused on financial factors, Pew identified several
additional factors that should be considered, drawn from similar policies adopted by food
industry groups and regulators.238 The draft notes that it “implements several of the U.S. GAO’s
recommendations from 2010 regarding the GRAS program.”239
As part of the review process for the draft guidance, Pew requested feedback from
workshop participants and outside reviewers (all submitted to the docket along with the draft
guidance). Some of these reflections are plaintive appeals for additional clarity from FDA from
scientists and practitioners.240 Lisa Lefferts, a Senior Scientist for CSPI, noted several other
238

In particular, the Draft Guidance prompts the manufacturer or end user to answer four questions to
evaluate whether the expert is qualified by scientific training and experience to evaluate whether a use of
a food substance may qualify for the GRAS exemption and provides additional detailed guidance for
evaluating responses for each question: 1) Does the expert have the necessary expertise? 2) Is the expert
too close to the issue to objectively assess the consensus of the scientific community? 3) Is the expert
disqualified because of serious financial conflicts of interest from exercising independent judgment? 4)
Are there non-financial conflicts of interest that need to be considered? Id. at 9.
239
Id. at 3.
240
For example, Nancy Rachman, a regulatory scientist with experience conducting risk assessments for
food additives and other chemicals, noted the following issues related to GRAS determinations:
1) “In addition to managing the potential for expert bias due to COI… two other aspects of any
regulatory process … are important for confidence in any decision: communication and
transparency… and steps to ensure the consistency and integrity of the scientific basis for
decisions. The less transparent a decision is, including its process and scientific basis, the greater
the concern about the potential for COI [conflicts of interest] will be. GRAS has real transparency
challenges…the food science, food toxicology and food regulatory community (including FDA)
have not done an adequate job of explaining some important aspects of GRAS.”
2) “The concept of ‘general recognition’ could be made much more transparent. It has always been
clear to me as a regulatory specialist and GRAS practitioner that an integral component of
‘general recognition’ is the use of well-established and accepted safety evaluation/risk assessment
approaches to the selection, analysis, interpretation and combining/weighting of evidence…Those
requirements and procedures are well established lore within the food science, food toxicology
and food law community, and reputable practitioners adhere to them. It would help improve
GRAS transparency if FDA would consolidate the various GRAS interpretive guidance materials
it has made available into a more comprehensive, ‘user-friendly’ and transparent format.”
Comments from Nancy J. Rachmann, GRAS consultant, to Tom Neltner, on Draft Guidance for Industry
on Selection of Experts Conducting “Generally Recognized as Safe” (GRAS) Evaluations 1-2 (Aug. 21,
2013).

46

 important aspects of the workshop’s presentations, including the following information that was
presented:
“A financial interest does not have to be great for the influence to be undue. Indeed,
social science research suggests that gifts of small value may influence decisions. It also
suggests that influence may operate without an individual being conscious of it.”241
Because conflicts of interest are not a problem of “deliberate corruption, but of
subconscious bias,”242 the social science research suggests that the only effective way to mitigate
conflicts of interest is to “realign incentives to eliminate conflicts.”243 For this reason,
mechanisms that create functional independence in researchers and their research outcomes –
such as separate funding – are recommended. Lefferts pointed out that other research-based food
regulators, such as those in Europe, require public disclosure of conflicts, and that requirements
for conflicts should be stricter than those for advisers of FDA given GRAS reviewers’ critical
role in determining the public’s exposure to ingredients. Moreover, Lefferts suggested that
proving an actual conflict of interest should not be necessary; instead, the presence of a potential
conflict should be “sufficient to raise questions about the integrity of the process and undermine
public confidence” in the safety of GRAS substances.244
Due to weak scientific requirements, nonexistent conflicts of interest standards, and
insufficient FDA oversight, the GRAS system does nothing to ensure the actual safety of
industry-determined GRAS substances. Pew concluded that these various deficiencies prevent
FDA from being able to oversee the safety of chemicals added to food as required by law.
c. Previous comments on this docket agree that the GRAS notification system
fails to ensure safety of substances added to food as required by the FAA.
Several consumer advocacy organizations submitted similar critiques as the GAO and
Pew in their comments on this proposal. Comments identified the lack of required notice, and
corresponding unavailability of necessary exposure data and scientific support, as the primary
weaknesses of the proposal, because they allow unapproved additives to avoid the review
required by the FAA. Thus, FDA is well aware of the problems with the 1997 proposal and it
would be arbitrary and capricious to finalize the rule as proposed.245
In a 2011 comment, Food & Water Watch urged FDA to adopt all of the
recommendations in the GAO report and condemned the proposed rulemaking as inadequate,
observing that the 1997 proposed rule falls short of what is necessary for adequate review and
that its “scope is inappropriately narrow [because it o]ffers only minor tweaks to a voluntary
system that does not offer real protection to consumers or an opportunity to participate in the

241

Comments from Lisa Lefferts, Senior Scientist, CSPI, to Tom Neltner, on Draft Guidance for Industry
on Selection of Experts Conducting “Generally Recognized as Safe” (GRAS) Evaluations (Aug. 1, 2013).
242
Id.
243
Id.
244
Id.
245
Envtl. Integrity Project v. E.P.A., 425 F.3d 992, 996 (D.C. Cir. 2005).

47

 process before substances are used.”246 Consumers Union commented in opposition to the
proposed notification system, because it fails to “ensure the safety of all substances added to
food,”247 which is the FDA’s statutory obligation under the FAA.
As Consumers Union noted, because “companies can make their own determination that
a substance is GRAS, not tell the FDA of that decision, and then start adding that substance into
food and selling it to consumers, . . . the system allows for potentially dangerous substances to
enter the food supply, without FDA’s knowledge or supervision.”248 As the International Center
for Technology Assessment (ICTA) pointed out in its 1997 comments on the proposed
rulemaking, the proposed notification system actually “reduce[s] the incentive for producers to
notify FDA, because notification would invite regulatory scrutiny without requiring the FDA to
attest to a determination of GRAS.”249
ICTA also pointed out that the lack of public notification prevents third parties (such as
consumer protection groups, competing companies, or scientific organizations) from being able
to alert FDA to potentially harmful substances in the food supply. Under the 1997 proposal,
“[t]he agency is essentially privatizing all new food additive data, an arbitrary and capricious
action which clearly contradicts Congressional goals of creating direct public oversight of the
food additive process.”250 As a result, the system fails to protect the safety of the food supply
and is a failure of FDA’s statutory obligations. Thus, to ensure the safety of all substances added
to food, as is the FDA’s responsibility, commenters contend that “companies, at a minimum,
should be required to notify FDA of any GRAS determinations they make.”251
Several groups expressed concerns about the fact that the current system does not make
dietary exposure estimates possible, which impedes an adequate finding of safety under the law.
NRDC, in its comment, recommends that GRAS safety determinations include published
exposure assessments.252 Food & Water Watch stated that companies should be required to
submit exposure data in their GRAS determinations so that industry, FDA, and citizens can make
a comprehensive assessment of the safety of substances.253
Finally, commenters noted that the “general recognition” standard for GRAS status is
significantly underdeveloped. As ICTA stated in its comments, “in yet another blow to food
safety oversight, the agency proposal broadens the description of ‘common knowledge’
246

Food & Water Watch, Comment on Docket No. FDA—1997—N—0020, Substances Generally
Recognized as Safe; Reopening of the Comment Period 1 (March 28, 2011).
247
Michael Hansen, Comments of Consumers Union on Food and Drug Administration (FDA) Proposed
rule on Substances Generally Recognized as Safe; Reopening Comment Period Docket No. FDA—1997—
N—0020 1 (March 28, 2011).
248
Id. at
249
International Center for Technology Assessment, Comments Concerning Docket No. 97 N-0103
“Substances Generally Recognized As Safe” 2 (July 16, 1997).
250
Id. at 3.
251
Hansen, supra note 247, at 1.
252
Natural Resources Defense Council, Comment on Docket No. FDA-1997-N-0020, Role of exposure
assessment in general recognition of safety and related issues 2 (Aug. 28, 2014).
253
Food & Water Watch, supra note 246, at 3.

48

 requirements so as to make the provision virtually meaningless.”254 For example, the technical
evidence requirements for the GRAS determinations have been expanded from “studies” to
include “scientific data and scientific information,” thereby “de-emphasizing and virtually
eliminating the requirement for peer-reviewed studies in deciding GRAS status.”255 This
severely weakens the “general recognition” requirement under the statute, as peer-reviewed and
published studies provide the expert and objective reviews essential for a “general recognition of
safety” finding. The commenters therefore emphasized the need to restore the centrality of
published, peer-reviewed scientific support in GRAS determinations.
Commenters also recommended that GRAS notices include information demonstrating
the independence of experts who generate data or analysis and that FDA should require
submitters to include a comprehensive discussion of information that may be inconsistent with a
GRAS conclusion.256 Without such requirements, any substance could masquerade as GRAS,
even one without an adequate record of safety. As the examples in the next section show, this is
in fact the case under the current system.
V. Case Studies: Many supposed “GRAS” substances today are in fact unapproved
food additives, in clear violation of Congressional intent under the FAA.
The 1997 proposal prevents FDA from being able to fulfill the regulatory role intended
by Congress because it leaves FDA without the information and oversight capability it needs to
ensure safety. The massive (and illegal) inadequacies of the current proposed GRAS system
mean that, in practice, many substances determined GRAS by industry are actually unapproved
food additives that FDA fails to adequately regulate or oversee under the statutorily required
FAP process. Even when a substance’s safety is questioned by independent experts
commissioned by the agency itself – as in the case of the SCOGS committee recommendations
in 1982 that FDA ignored, discussed above – it may be marketed to consumers without going
through the food additive petition process. Furthermore, because notifications are not required,
secret industry GRAS designations are made without the information on exposure and
cumulative effects that should be necessary to make such a determination.
The following section provides several current examples of how the 1997 proposal fails
to meet the requirements of the FAA: (1) the results of a FOIA request performed by CSPI with
regard to caffeine and caffeinated products; (2) a survey of withdrawn GRAS notifications; (3)
an FDA response letter on taste modifiers; (4) FDA’s lack of control regarding the possible use
of nanotechnology in the food supply; (5) several flavorings determined to be GRAS despite
being found carcinogenic by authoritative bodies; and (6) industry-determined GRAS substances
of dubious safety, such as Quorn-brand products. These case studies are only a small sampling
of how the 1997 proposal prevents effective regulation of substances added to foods and thus
directly contravenes the FAA.

254

International Center for Technology Assessment, supra note 249, at 3.
Id.
256
See Food & Water Watch, supra note 246; Hansen, supra note 247.
255

49

 a. The GRAS status of caffeine illustrates how current system impedes
informed and consistent GRAS determinations.
As the GAO and Pew studies and other comments on the docket noted, the 1997 proposal
makes meaningful oversight over substances added to foods impossible. This problem is starkly
illustrated by the GRAS status of caffeine. Although caffeine is limited in cola-type beverages to
72 milligrams per 12 ounces in a GRAS regulation,257 caffeine is a product that industry has selfdetermined to be GRAS – or even used without self-determining that it is GRAS – for many
other novel applications, even while it poses safety risks to consumers. FDA cannot effectively
regulate industry’s new and increased uses of caffeine under the current system because it lacks
information, such as exposure data, required to make ongoing assessments of safety.258
Caffeine added to foods and beverages has been linked in FDA’s own data to serious
adverse effects for consumers. In response to a FOIA request, CSPI found that between January
1, 2004, and March 10, 2014, FDA received 276 reports of adverse events associated with
energy drinks containing added caffeine.259 Thirty-four of those incidents resulted in death,
often due to heart failure. Of the remaining reports, 42 involved life-threatening injuries, 82
involved injuries that were characterized as “serious,” 115 resulted in hospitalization, 15 resulted
in disability, and one resulted in miscarriage.260 Consumers reported experiencing high blood
pressure, convulsions, paralysis, and brain damage due to lack of oxygen.261 These risks were
particularly acute in vulnerable populations such as teenagers, perhaps due to their comparatively
lower body weights and inexperience with caffeine consumption.262
257

21 C.F.R. § 182.1180.
Given that companies’ GRAS determinations are kept secret from the public and FDA, we rarely have
access to a range of self-determinations made for different uses of a substance by a wide swath of
companies. Because FDA requested and received this information for caffeine, and it became available to
CSPI through a Freedom of Information Act (FOIA) request, the use of caffeine in foods and beverages
presents a compelling case study in the failures that attend the FDA’s current policy permitting secret
GRAS determinations.
259
Results of CSPI FOIA Request, FOI 14-17 Report, Center for Food Safety and Applied Nutrition
Adverse Event Reporting System (CAERS) Reports Allegedly Related to Multiple Energy Drinks,
available at http://cspinet.org/images/foia-energy-drinks.pdf.
260
Id.
261
Id.
262
Anais Fournier, a 14-year-old girl, died from a cardiac arrhythmia due to caffeine toxicity after
consuming a Monster energy drink while she was at the mall with her friends. Autopsy Report of Anais
Fournier, Age 14, Hagerstown Maryland Resident Who Died After Consuming Monster Energy Drink.
The death of a 19-year-old in California and the brain damage suffered by an Oklahoma 16-year-old have
also been linked to the consumption of Monster beverages. James Mulcahy & Kevin A. Adams, Monster
Beverage Company —A Case Study of Vertical Distribution Landmines, Mulcahy LLP,
http://mulcahyllp.com/firmnews/practicenews/monsterbeveragecompanyacasestudyofverticaldistributionl
andmines.html. Additionally, government records show that two 25-year-olds and an 18-year-old died in
Canada, possibly because of their consumption of either Monster or Red Bull beverages. Energy drinks
suspected to have caused deaths of 3 Canadians, Toronto Star, Nov. 18, 2012,
http://www.thestar.com/news/canada/2012/11/18/energy_drinks_suspected_to_have_caused_deaths_of_3
_canadians.html. And while energy drinks and other products with added caffeine are oftentimes
marketed to athletes, a case report indicates that a 28-year-old man consumed three energy drinks and five
258

50

 These reports demonstrate that the consumption of foods and beverages with added
caffeine may have dangerous effects, and in June 2013, FDA requested information from
companies that had self-determined the GRAS status of caffeine in their products. FDA asked
each company how it had determined that its products and proposed uses met the standards for
GRAS. The companies’ responses relating to their self-determinations of GRAS status expose
glaring inconsistencies and deficiencies in their determinations, and provide a stark indication
that industry self-determined GRAS does little to ensure the safety of the food supply. This
sorry record – almost by itself – demonstrates the need for required notification and
recordkeeping, along with better-defined requirements for the science and exposure data
necessary to sustain a GRAS determination. Despite the discrepancies, we are unaware of any
action taken by FDA to date to enforce GRAS obligations for these companies or revoke the
GRAS determinations regarding these uses of caffeine.
The industry’s GRAS responses indicate a desperate need for commonly agreed-upon,
accurate, and updated exposure data – all information necessary for the FAA to function as
intended. When considering worst-case scenarios for ingestion of caffeine, many of the
companies assumed unrealistically low levels of exposure. Jelly Belly, for instance, noted
exposure levels resulting from just two to three servings of its “Extreme Sports Beans”
(caffeinated jelly beans),263 even though the product is sold in units that contain 24 packaged
servings and has an appearance that a consumer could easily confuse with non-caffeinated candy.
More troublingly, the companies fail to adequately consider whether consumers have
already ingested caffeine from other sources, as would be required for a meaningful assessment
of the cumulative effect of the substance. It is well within the realm of possibility to imagine a
situation in which a consumer drinks a coffee before consuming a large quantity of caffeinated
jelly beans, for instance. Consumption from several products would represent the upper level of
caffeine exposure presented by a product, but none of the companies that responded to FDA’s
request for information factored in caffeine exposure from other sources when considering
worst-case exposure scenarios.264 In fact, Jelly Belly specifically (and appallingly) stated that
other sources of caffeine need not be considered because consumers would replace their other
sources of caffeine with their beans. The company’s self-serving, illogical, and empirically
unsupported assumptions should be plainly insufficient to support a GRAS finding.

hours later lost consciousness while playing basketball. He was hospitalized and after three days died
following sudden cardiac arrest. S. Avci, R. Sarikaya & F. Buyukcam, Death of a Young Man after
Overuse of Energy Drink, 31 Am. J. Emergency Med. 1624 (2013).
263
Letter from John Di Giusto, General Counsel, Jelly Belly Candy Company, to Michael Landa,
Director, Center for Food Safety and Applied Nutrition (June 24, 2013) (on file with CSPI).
264
This ignores FDA’s guidance for intake estimates of direct additives, which requires the petition
process to consider “[a]ny anticipated increase in consumption from its petitioned use(s), if the food
additive is also a naturally occurring material. The concentration of the additive occurring naturally in
food(s) and an estimate of the level of consumption of the food(s) should be provided.” U.S. Food and
Drug Administration, Guidance for Industry: Recommendations for Submission of Chemical and
Technological Data for Direct Food Additive Petitions (March 2009), available at
http://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/IngredientsA
dditivesGRASPackaging/ucm124917.htm#intake.

51

 The responses that companies submitted to FDA regarding 90th percentile intake of
caffeine reveal further discrepancies and inadequacies with regard to exposure estimates and the
need for accurate exposure data. Kraft, for instance, assumed 90th percentile consumption levels
ranging from 2.3 mg/kg/day to 5.2 mg/kg/day, which in a 70 kg person would amount to 161 to
364 mg of caffeine.265 In contrast, Pepsi assumed 90th percentile caffeine levels at 392 mg per
day for adults over 20 and Coke cited a 90th percentile of 406.3 mg per day for adults over 18 in
its self-determination.266 Monster assumed an even higher 90th percentile intake of 8.6 mg/kg,
or 600 mg/day for a 70 kg person.267 In addition, Monster noted that per capita intake of caffeine
is approximately 300 mg/day,268 which is significantly higher than the lower range of 90th
percentile consumption assumed by Kraft. Monster was also the only company that mentioned
90th percentile caffeine intake for adolescents, which it stated to be less than 225 mg.269
These significant discrepancies in the most basic assumptions needed for a safety
assessment reveal that GRAS – even industry-determined GRAS – cannot work when companies
do not share information with FDA and each other. The chaos that results indicates that some
companies lack the updated and accurate exposure data necessary to make a GRAS
determination, and will tailor core scientific assumptions that are critical to public health to meet
their own perceived needs.
The data sources companies rely upon are also cause for concern. Jelly Belly, for
instance, examined outdated caffeine consumption data from nearly 60 years ago.270 Many of
the companies relied on data from foreign markets, where consumption trends may differ from
those in the United States. Monster cited a survey conducted by researchers at Penn State
University271 that found that four percent of Americans who consume caffeine consume energy
drinks.272 This rate is higher than rates found in studies cited by both Monster and Rockstar
265

Letter from Kraft Foods Group, Inc., to Michael Landa, Director, Center for Food Safety and Applied
Nutrition 9 (Aug. 5, 2013) (on file with CSPI).
266
Letter from Rene Lammers, Chief Science Officer, PepsiCo, to Michael Landa, Director, Center for
Food Safety and Applied Nutrition 2 (Aug. 1, 2013) (on file with CSPI); Letter from Rhona Applebaum,
Vice President and Chief Science & Health Officer, The Coca Cola Company, to Michael Landa,
Director, Center for Food Safety and Applied Nutrition (Aug. 1, 2013) (on file with CSPI).
267
Letter from Miriam Guggenheim, Counsel to Monster Beverage Company, to Michael Landa,
Director, Center for Food Safety and Applied Nutrition 6 (June 24, 2013) (on file with CSPI).
268
Id.
269
Id. at 7. Adolescents are a vulnerable population when it comes to caffeine. See Scientific Report of
the 2015 Dietary Guidelines for Americans Committee 8-9 (Feb. 2015) (“The marketing and availability
of high-caffeine beverages and products is on the rise. Unfortunately, only limited evidence is currently
available to ascertain the safety of high caffeine intake (greater than 400 mg/day for adults and
undetermined for children and adolescents) that may occur with rapid consumption of large-sized energy
drinks. Limited data suggest adverse health outcomes, such as caffeine toxicity and cardiovascular
events.... Limited or no consumption of high caffeine drinks, or other products with high amounts of
caffeine, is advised for children and adolescents.”).
270
Letter from John Di Giusto, supra note 263.
271
D.C. Mitchell, et al., Supported by Washington (DC): International Life Sciences Institute, Beverage
Caffeine Intakes in the U.S. (2013) [Poster presented at EB 2013 in Boston, ASN Meeting, April 22,
program number: 848.12, abstract number: 7769].
272
Letter from Miriam Guggenheim, supra note 267, at 7.

52

 based on Canadian consumption. One such study mentioned by the companies found that only
1.5 percent of Quebecois teenagers consumed energy drinks daily,273 while another found that
only one percent of Quebecois teenagers reported daily energy drink consumption – while small
in order, one of these estimates is fifty percent higher than the other, which would alter the
estimate significantly.274
The fact that the companies made such disparate and often outdated exposure
assessments demonstrates that safety determinations for GRAS may not be based on accurate or
updated data – which is hardly surprising given that the current system does not require
submission of exposure data. The current GRAS system fails to ensure that such information is
– or even can be – taken into account by companies when determining whether use of a given
substance is safe. Accurate exposure data could only be obtained if companies are required to
submit GRAS notifications along with the most accurate possible updated exposure estimates,
and FDA engages in an iterative process with companies to standardize such estimates, as it does
for food additive petitions (which must meet identical requirements related to probable
consumption and exposure).
Due to FDA’s failure to develop any regulatory definition of “harm,”275 the submissions
also fail to seriously consider important risks associated with the intended use of caffeine. For
example, while the companies consider lethal doses in their GRAS determinations, the amount of
caffeine known to produce non-permanent harm, such as anxiety, nausea, and sleeplessness –
which are less severe but still significant – received scant attention. The possibility for non-fatal
side effects, for instance, is not even addressed in Jelly Belly’s response. Kraft’s response notes
that consumption of an entire bottle its MiO caffeine concentrate would result in exposure of
approximately 1,000 mg of caffeine; according to Kraft’s own assessment, this amount presents
a reasonable likelihood of adverse effects in children.276 But Kraft disregards this danger and
instead assesses its product according to the more serious but less-likely-to-occur possibility that
it might cause a severe adverse event, which Kraft defines as one “requiring medical intervention
to manage, mitigate or prevent progression to a life-threatening situation.”277 This means that, in
the absence of a regulatory definition of harm, industry has literally defined “GRAS” as anything
that will not kill you – a grossly insufficient standard to protect public health.
Red Bull and Rockstar similarly devote negligible attention to the risk of harm in their
responses. Pepsi and Coke acknowledge the potential for caffeine to cause anxiety and sleep
273

Institut de la Statistique du Quebec, Frequence de consommation habituelle de
certaines boissons sucrees, eleves du secondaire, Tableau A3.2, L'Enqu~te quebecoise sur la sante des
jeunes du secondaire 2010-2011: TOME 1: Le visage des jeunes d'aujourd'hui: leur santti physique et
leurs habitudes de vie (2012).
274
Reseau du sport etudiant du Quebec, Enquete Quebecoise sur le Marketing de la Malbouffe: 10 000
Jeunes se Prononcent! (2011).
275
Maffini et al, supra note 226 at 441.
276
Letter from Kraft Foods Group, supra note 265, at 8. (Noting that the probability of adverse side
effects is considered “low-medium” for ingestion rates between 500 and 1,000 mg of caffeine and
“medium” for rates between 1,000 and 1,500 mg). This contrasts with the DSM-5 which states that
caffeine intoxication can occur “with low doses (e.g., 200 mg) in vulnerable individuals such as children,
the elderly, or individuals who have not been exposed to caffeine previously.”).
277
Id.

53

 disturbance but dismiss these concerns, citing studies that show that caffeine use is self-limiting,
although consumers would inevitably suffer adverse effects before deciding to limit their
intake.278 And Monster, though it mentions the potential behavioral effects of caffeine on
children in passing, ultimately disregards them. The current system ignores the potential effects
of chronic and sustained exposures, in clear violation of the FAA’s original intent to address
effects beyond acute toxicity.279
The potential for adverse effects is an especially glaring omission considering that
teenagers are the most frequent users of energy drinks and adverse effects are more likely to
occur in consumers who are caffeine-naïve and have lower body weights.280 It shows the need
for scientific requirements regarding vulnerable populations, as recommended by the Pew
studies. As Monster notes in its response, a study of caffeine-naïve children found that ingestion
of less than 50 mg of caffeine per day produced effects such as restless behavior and difficulty
sleeping.281 More pointedly, the varying definitions and lack of clarity point out an urgent need
for FDA to clarify the definition of harm in its final rule.282
Moreover, the responses demonstrate that the voluntary nature of the current regime
permits industry self-determinations to ignore the applicable scientific requirements for GRAS.
For example, the Jelly Belly submission indicates that it based its GRAS finding for Extreme
Sports Beans on the “common use” of caffeine before 1958. In its submission, Jelly Belly stated
that no further scientific assessment was required because caffeine was in common use before
the FAA.283 However, the interim rule clearly states that it is the nature of the use, not the mere
use of the substance in foods, which confers eligibility for the GRAS exemption based on use
before 1958. Thus, an “evaluation of whether an additional use of a substance that is GRAS
through experience based on common use in food is also GRAS requires a scientific procedures
GRAS determination when the use in question was not common prior to January 1, 1958.”284 As
caffeine was not used as an additive in jelly beans before 1958, Jelly Belly’s GRAS
determination must be based on scientific evidence, not common use. However, as the current
system is voluntary, companies often fail to follow applicable regulations and employ adequate
science.
The responses also confirm the potential for conflicts of interest in GRAS determinations
flagged by Pew. Kraft, for instance, used an in-house toxicologist to conduct the company’s
determination, who singlehandedly determined that 60 mg of caffeine per 8-ounce serving was
278

Letter from Rene Lammers, supra note 266, at 18; Letter from Rhona Applebaum, supra note 266.
Even the SCOGS, back in 1982, recognized the potential risks of caffeine. The majority conclusion
was that FDA should issue an interim food additive regulation requiring commitment, within a stated
period, that necessary testing would be undertaken (the substance continuing as GRAS during completion
of tests and evaluation of the results). The minority conclusion would require that safe conditions of use
be established. Select Committee on GRAS Substances, Insights on food safety evaluation 50, 52 (1982).
280
F. Castellanos & F. Rapoport, Effects of caffeine on development and behavior in infancy and
childhood: a review of the published literature, 40 Food & Chemical Toxicology 1235-1242 (2002).
281
Id.
282
Maffini et al., supra note 226, at 441 (noting that the lack of definitions of “harm” or “adverse effects”
in FDA rules and public guidance documents makes system unpredictable and difficult to administer).
283
Letter from John Di Giusto, supra note 263.
284
62 Fed. Reg. at 18950 (emphasis added).
279

54

 GRAS, even though that is 25% higher than the amount determined by FDA to be GRAS when
used in a cola beverage.285 Monster also made its own GRAS determination, and later hired
consultants to prepare a safety assessment “(t)o help ensure the independence of the GRAS
determination.”286 The rest of the companies commissioned consulting firms and toxicologists.
This financial relationship presents a conflict of interest because scientists may feel pressure to
reach favorable conclusions for the companies that employ them.287 These conflicts of interest
undercut the integrity of the determinations and underlying scientific assessments.
The industry’s limited assessment of adverse health effects and vulnerable
subpopulations, insufficient attention to updated science, use of woefully dated and incomplete
exposure assessments, and rampant conflicts of interest all demonstrate that the GRAS system as
it currently operates does not achieve the purposes of the FAA.
b. Scores of withdrawn GRAS notifications illustrate how the GRAS
notification system fails to ensure the safety of substances used in the food
supply.
As explained above, companies will sometimes withdraw their GRAS notifications rather
than waiting for FDA’s response to their submissions. Typically, if FDA has raised safety
questions about a substance, a company will withdraw its notification, and FDA will
subsequently disclose only that it has ceased to evaluate the GRAS notice at the company’s
request, without publishing the concerns that might have led to this request. Companies ask
FDA to cease evaluations of their GRAS determinations with alarming frequency. A review of

285

21 C.F.R. § 182.1180 (caffeine up to a level of 0.02 percent (200 ppm) is generally recognized as safe
(GRAS) for use in cola-type beverages).
286
Letter from Miriam Guggenheim, supra note 267, at 1.
287
Recent research has documented significant effects related to funding bias in the outcomes of studies,
and industry sources of funding for research have been found to have significant impacts on outcomes of
studies. See, e.g., A. Lundh et al., Industry Sponsorship and Research Outcome, The Cochrane
Collaboration (2013) (finding in a review of 48 papers on drugs and medical devices, that “industry
sponsored drug and device studies are more favorable to the sponsor’s products than non-indutsry
sponsored drug and device studies due to biases that cannot be explained by standard risk of bias
assessment tools.”); Lenard Lesser, Cara Ebbeling, et al., Relationship between Funding Source and
Conclusion among Nutrition-Related Scientific Articles, 4 PLoS Med 41-46 (2007) (finding that studies
sponsored exclusively by food/drinks companies were four to eight times more likely to have conclusions
favorable to the financial interests of the sponsoring company than those which were not sponsored by
food or drinks companies); Anke Huss et al., Source of Funding and Results of Studies of Health Effects
of Mobile Phone Use: Systematic Review of Experimental Studies, 115 Env. H. Persp. 1-4 (Jan. 2007)
(finding that studies of the health effects of cellular telephone use that were “funded exclusively by
industry … were least likely to report a statistically significant result.”); Paul Ridker & Jose Torres,
Reported Outcomes in Major Cardiovascular Clinical Trials Funded by For-Profit and Not-for-Profit
Organizations, 2000-2005, 295 JAMA 2726 (May 17, 2006) (finding that recent cardiovascular trials
funded by for-profit organizations are more likely to report positive findings than trials funded by nonfor-profit organizations: while non-profits’ studies favored new treatments in 49 percent of case, studies
funded by for-profits favored new treatments in 67 percent of studies, and approximately the same results
held when only randomized trials were considered).

55

 the list of GRAS notifications on FDA’s Web site indicates that 84 notifications have been
withdrawn out of a total 562 submissions – 15% of the total number of notifications.288
Withdrawal of notifications, conducted under the terms of FDA’s 1997 proposal,
seriously undercuts what little oversight FDA previously had retained over the GRAS process.
In many of these cases, FDA identifies serious concerns about a company’s GRAS notification,
yet a company is allowed to continue marketing and selling its product for consumption, without
publication of the agency’s concerns for public or third-party review.289
In 2014, NRDC made a FOIA request for communications between FDA and the
manufacturers behind 16 withdrawn GRAS notifications, to ascertain the types of concerns that
would prompt manufacturers to ask the agency to stop reviewing a GRAS notice. In its review
of the FOIA documents, NRDC found that FDA does review notifications and ask tough
questions. When a notification raises serious safety concerns, FDA tells the company that it will
reject a notice if it is not voluntarily withdrawn. The company may then withdraw the
notification and use the product anyway, without any online publication of FDA’s concerns.290
In such a case, both the public and other manufacturers are deprived of important information
that might affect future decisions.
In a 2014 report, NRDC described several chemicals that have been used in foods after
manufacturers withdrew their GRAS notifications in response to FDA’s questions. For example,
epigallocatechin-3-gallate (EGCG) is a chemical used in beverages. Its manufacturer declared
the chemical GRAS despite evidence that it may cause leukemia in fetuses, and short-term study
results showing that it affected the thyroid, testis, spleen, pituitary, liver, and gastrointestinal
tract.291 The notice also did not explain potentially dangerous interactions with sodium nitrite, a
common preservative, or with acetaminophen, a common ingredient in over-the-counter
painkillers. A company has submitted and withdrawn GRAS notices for EGCG twice. Despite

288

U.S. Food and Drug Administration, GRAS Notices,
http://www.accessdata.fda.gov/scripts/fdcc/?set=GRASNotices (last accessed Feb. 25, 2015). In the FDA
experience memo, spanning the period from 1998-2009, FDA estimates that 16% of notices are
withdrawn. See Kahl Memo, supra note 67, at 10.
289
Neltner & Maffini, supra note 66, at 9. This report provides the following quote from an email
procured from the FOIA request:
The next day, [notifier] called and asked whether [notifier] would have an option to
withdraw the notice rather than receive a letter that the notice did not provide a basis for a
GRAS determination. I replied that this was an option. On September 4, [notifier] asked
whether [notifier] could still sell its [name] product if it withdrew its GRAS notice.
Consistent with my response to her earlier question about marketing [name], I said yes.
Id. at 3 (quoting FDA officer summarizing telephone conversations with manufacturer regarding its
GRAS notice review).
290
Id. at 9. An FDA reviewer stated that FDA “cannot require anything, as this is a voluntary program
and we don’t want to frighten anyone away.” Id.
291
U.S. Food & Drug Administration, GRAS Notice Inventory, GRN No. 225 and NRDC, Main FDA
Response to FOIA regarding GRN No. 225 (2014), www.nrdc.org/food/files/chemicals-in-food-FOIAMain.pdf.

56

 the red flags identified by FDA, other companies still use EGCG in their products: NRDC
identified 25 food products with EGCG as a named ingredient.292
Gamma-amino butyric acid (GABA) is a neurotransmitter that was self-determined
GRAS by industry for use in beverages, chewing gum, coffee, tea, and candy. GABA was
declared GRAS despite the fact that estimated exposures were in excess of what the
manufacturer considered safe. In addition, the submitter relied on unpublished safety studies,
and failed to consider existing exposures.293 When FDA identified these concerns, the company
withdrew its notice. While the company told NRDC that it would not market the product for use
in food, it continued to use the ingredient in dietary supplements. NRDC also identified five
food products marketed by other companies with GABA as a named ingredient, including bottled
tea and nutrition bars.294
Sweet lupin protein, fiber, and flour have been declared GRAS by an Australian firm for
use in baked goods, dairy products, gelatin, meats, and candy, despite concerns that the
chemicals could cause allergic reaction in people with peanut allergies.295 In communications
with the firm, FDA noted that a warning label would be insufficient to alert consumers to this
risk, and the company withdrew its GRAS notification. Despite FDA’s concerns, sweet lupin is
listed as an ingredient in more than 20 food products, none of which include a warning for those
with peanut allergies.296 These products are likely marketed by other companies who have not
been alerted to FDA’s concerns, as they are not publicized under the current system.
Finally, theobromine was declared GRAS by a U.S. company for use in bread, cereal,
beverages, chewing gum, tea, soy milk, gelatin, candy, and smoothies, despite having an
estimated consumption rate that was more than five times the safe consumption level reported by
the company’s consultant. On top of this already considerable cause for alarm, the notification
did not provide explanations for various safety concerns raised by animal testing of theobromine,
including testicular degeneration and delayed bone formation.297 Nevertheless, NRDC found
that theobromine was a named ingredient in more than 20 food products, including isotonic
waters, nutrition bars, and diet foods.298
These examples show that even when the GRAS process is theoretically “working” –
when FDA actually receives and reviews GRAS notifications – it does little to actually ensure
the safety of our food supply. Manufacturers are permitted to withdraw their notifications and
292

Neltner & Maffini, supra note 66, at 9.
U.S. Food & Drug Administration, GRAS Notice Inventory, GRN Nos. 257 and 259, and NRDC,
Main FDA Response to FOIA regarding GRN Nos. 257 and 259 (2014),
www.nrdc.org/food/files/chemicals-in-food-FOIA-Main.pdf.
294
Neltner & Maffini, supra note 66, at 9.
295
FDA, GRAS Notice Inventory, GRN Nos. 262, 263, and 264, and NRDC, Main FDA Response to
FOIA regarding GRN Nos. 262, 263, and 264 (2014), www.nrdc.org/food/files/chemicals-in-food-FOIAMain.pdf.
296
Neltner & Maffini, supra note 66, at 10.
297
FDA, GRAS Notice Inventory, GRN No. 340, and NRDC, Main FDA Response to FOIA regarding
GRN No. 340 (2014), www.nrdc.org/food/files/chemicals-in-food-FOIA-Main.pdf.
298
Neltner & Maffini, supra note 66, at 10.
293

57

 continue to use the product despite legitimate concerns raised by FDA. Moreover, the
ingredients may be marketed by other companies – and, in fact, because FDA’s concerns are not
published, the third-party companies may never know about the safety questions raised by the
agency. This means that FDA’s process does nothing to prevent unsafe products from entering
the food supply or being adequately labeled. This demonstrates both the vacuity of the
proposal’s process for GRAS and the importance of making GRAS notifications both required
and public. In addition, the fact that significant safety concerns have been raised in the scientific
community regarding these substances means that they should not be able to use the GRAS
process in the first place.
c. The current GRAS system permits novel “food additives” to bypass the
statutory food additive process.
We have identified many examples of novel and/or questionable substances that cannot
meet any possible interpretation of the “GRAS” language in the FAA, but nonetheless have
apparently been self-determined GRAS by industry, rather than going through the statutory FAP
process.
i.

Taste Modifiers

As described above, the FAP process was intended to apply to novel, unfamiliar additives
– as these substances could not have a track record of safety necessary to qualify as “GRAS.”
Yet, under the current system, many novel substances are used in foods without a FAP. For
example, in 2013, CSPI wrote to FDA seeking information about the labeling requirements for a
novel group of substances known as “taste modifiers,” including, specifically, 4-amino-5,6dimethylthieno[2,3-d]pyrimidin-2(1H)-one & hydrochloride salt and 3-[(4-amino-2,2-dioxido1H-2,1,3-benzothiadiazin-5-yl)oxy]-2,2-dimethyl-N-propylpropanamide. The presence of these
chemicals is not obvious from food labels because the food industry claims – wrongly in our
view – that they are encompassed by the labeling term “artificial flavors,” despite the
manufacturer’s acknowledgment that they do not have any taste of their own. Therefore, their
use in the food supply – and the extent of this use – is a mystery to consumers.
FDA is apparently content to let it remain mysterious. In response to CSPI’s letter on
taste modifiers (two years after receipt), FDA wrote that the substances in question had been
determined GRAS by industry, that the manufacturer had not submitted any information on the
substances to FDA, and therefore FDA had not conducted an evaluation of the substances.299
These types of chemicals, 4-amino-5,6-dimethylthieno[2,3-d]pyrimidin-2(1H)-one &
hydrochloride salt and 3-[(4-amino-2,2-dioxido-1H-2,1,3-benzothiadiazin-5-yl)oxy]-2,2dimethyl-N-propylpropanamide, which were lab-created to modify how humans perceive
flavors, plainly constitute the kind of “newly discovered substances” that Congress was
contemplating when it passed the FAA. That such complex, novel compounds have been selfdetermined as GRAS illustrates how industry uses the GRAS exemption to circumvent the food
additive petition process required by Congress. Furthermore, the fact that their toxicology, use,
and exposure levels are not publicly known should preclude any possible “general recognition”
299

Letter from Yinqing Ma, Center for Food Safety and Applied Nutrition, U.S. Food & Drug
Administration, to Michael F. Jacobson, Executive Director, CSPI (Feb. 9, 2015) (on file with CSPI).

58

 of their safety, as required by the statute and FDA regulations. Their industry-determined GRAS
status is conclusive proof that the current system is not working at intended by Congress.
ii. Nanomaterials
In addition to taste modifiers, we have concerns – shared by GAO – about nanomaterials
being used in the food supply without FDA oversight. Nanomaterials are another example of a
novel food additive that may masquerade as GRAS rather than being approved via the statutory
food additive petition process. Engineered nanomaterials are created through nanotechnology –
the manipulation of materials at a molecular scale that enhances the resulting nanomaterials’
physical properties.300 While the underlying chemical structure of a substance is not changed by
the engineering process, its physical properties may change. As GAO noted in its 2010 report,
nanotechnology presents potential challenges to the regulation of food safety because companies
may conclude that their engineered nanomaterials are GRAS without informing FDA.301
Since the GAO report was issued in 2010, FDA has issued a voluntary guidance on
nanotechnology in food ingredients, suggesting that such products are ineligible for selfdetermined GRAS;302 however, this guidance is not binding.303 Furthermore, because all GRAS
notifications are voluntary, FDA has no way of knowing the full extent to which engineered
nanomaterials have entered the U.S. food supply, despite the fact that they are precisely the type
of “food additives” warranting pre-market testing contemplated by Congress in passing the FAA.
In its guidance, FDA notes that it would be “prudent” for manufacturers to consult the
agency on changes in manufacturing processes that include “novel” and “emerging” technologies
such as nanotechnology (which, according to FDA, results in “new properties not seen in
traditionally manufactured food substances”304). While the agency fails to require
consultation,305 FDA clearly understands the risks of allowing the industry to experiment on the
public, stating: “[t]he consequences (to consumers and to the food industry) of broadly
distributing a food substance that is later recognized to present a safety concern have the
potential to be significant.”306 This is precisely the concern that drove Congress to enact the
FAA in the first place – and is therefore yet another stark illustration of how the statutory food

300

GAO Report, supra note 37, at 2.
Id. at 26.
302
Nanotech Guidance, supra note 169.
303
As the guidance indicates: “[i]t does not create or confer any rights for or on any person and does not
operate to bind FDA or the public. You can use an alternative approach if the approach satisfies the
requirements of the applicable statutes and regulations.” Id. at 1. It further provides that “FDA’s
guidance documents, including this guidance, do not establish legally enforceable responsibilities. … The
use of the word should in Agency guidelines means that something is suggested or recommended, but not
required.” Id. at 6. As there are no “statutes or binding regulations” specific to nanotechnology and FDA
has interpreted its own obligations on the safety of additives and GRAS so haphazardly, as above, this
text in this context is essentially meaningless.
304
Id. at 14.
305
Id. at 5.
306
Id.
301

59

 additive petition process has been undone by FDA’s interpretation of its obligations under the
law.
Nanotechnology also reveals the limitations of FDA’s current scientific standards for
testing of GRAS substances. Nanomaterials may have a unique effect on the human body, as
changes in the “physical and chemical properties of a food substance can affect its bioavailability
through altered absorption, distributions, metabolism and excretion of the substance in the body
[and] affect the level at which toxic effects may occur,” including such changes as “particle
size.”307 FDA concedes that its standard toxicological guidance may not address “known
toxicological endpoints” that are a concern for some compounds or class of compounds, and
states that “in such cases, it is the responsibility of chemical manufacturers and food industry end
users to develop appropriate protocols to address particular safety issues” and that “historically,
such persons have been advised” to consult FDA.308 In short, nanotechnology clearly alters the
metrics for risks to human health, and FDA’s current scientific standards do not address some of
these new and emerging risks.
FDA’s analysis makes that clear that particular testing to measure the impacts of
nanotechnology would be needed to assure the safety of such substances. 309 Specifically, due to
the variability shown in a body of research regarding the toxicology of nanotechnology,
particular kinds of in vitro tests would be required to validate the results of safety testing for
nanomaterial foods substances.310
In the guidance, the agency states that it has not to date received GRAS affirmation
petitions or notices for any uses of nanomaterials in food, but that where “safety questions are
raised that experts would need additional data to resolve and such data are not generally
recognized, the criteria for GRAS would not be satisfied for the use of such food substances.” 311
The agency further stated that it was not aware of any food substances intentionally engineered
on the nanometer scale “for which there are generally available safety data sufficient to serve as
the foundation for a determination that the use of a food substance is GRAS.”312

307

Id. at 13.
Id.
309
Id. at 15 (“The variation in biological activity that may result from engineering food substances in the
nanometer range may raise questions about the applicability of traditional safety tests for these materials.
Thus, as with any studies to support the safety of food substances, studies to establish the safety of food
substances manufactured using nanotechnology should have been appropriately validated for these
materials. Notably, variability has been reported when traditional toxicity tests have been used to assess
nanomaterials. Because of this variability, and because the physiochemical properties of an individual
nanomaterials may require adjustment to a particular assay, validation (single and/or multi-laboratory) of
traditional in vitro toxicity tests will ensure that the results are meaningful and appropriate to the safety
assessment of the nanomaterial food substance. FDA anticipates that method validation for nanomaterial
food substances would include parameters such as accuracy, precision, sensitivity, specificity,
repeatability and biological relevance.”)
310
Id.
311
Id. at 15-16.
312
Id.
308

60

 Thus, FDA is suggesting that nanomaterials cannot be GRAS – a stunningly logical
conclusion, under the law – but it is doing so in a nonbinding guidance in the context of a
program that requires no notice to FDA of the use of any substance self-determined by the food
industry to be GRAS. This development is so far outside the contemplation of the FAA that it is
a farce built upon a fallacy. In fact, the agency currently has no way of knowing whether
nanotechnology is being used in food; even after issuance of its guidance, these substances could
be in the food supply and FDA would never know. Most troublingly, the guidance from FDA
likely (and foreseeably) undermined the agency’s ostensible purpose: by signaling that expensive
additional testing sensitive to the risks of nanotechnology would be required to demonstrate the
safety of a nanotech food substance in any notification to the agency, FDA ensured that any use
of such substances would remain a food industry secret.
However, FDA’s invocation of a limitation on GRAS based on an assessment of risks
and the inadequacy of data sufficient for a general safety is instructive. It shows a way forward,
should the FDA finally breathe meaning and life into its moribund authorities. The fact that
FDA limits, in its guidance, the eligibility of nanotech food substances for a GRAS designation
is good precedent for an appropriate additional narrowing of GRAS eligibility with binding rules
that bring the program in line with both historical indicators of Congressional intent and the clear
statutory design.313
d. The lack of appropriate limits on “GRAS” designations allows substances
whose safety is questioned in the scientific community or unsupported by the
scientific literature to impermissibly bypass the food additive process.
By the plain language of the statute, a substance whose safety has been questioned by the
scientific community cannot be “generally recognized as safe.” However, we are aware of many
cases of allegedly GRAS substances whose safety has been seriously questioned by the scientific
community or is unsupported by the scientific literature. These examples demonstrate that the
1997 proposal allows industry to make GRAS designations for substances whose safety is not
“generally recognized” by experts – in plain violation of the FAA. These substances, by
definition, should not fall within the GRAS exemption and, because of questions about their
safety raised in the scientific community, should only be approved for use in food via the
congressionally mandated food additive petition process. Instead, their GRAS status illustrates
how the system permits substances whose dangers have been recognized by authoritative bodies,
or whose safety has not been demonstrated by science, to bypass the statutory system without
FDA’s knowledge or approval.
i. Flavorings

313

FDA made similar findings in a nonbinding industry guidance regarding substances added to
beverages and dietary supplements. See U.S. Food and Drug Administration,
Guidance for Industry Considerations Regarding Substances Added to Foods, Including Beverages and
Dietary Supplements 5 (Jan. 2014) (stating that it was “concerned that some of the novel substances that
are being added to conventional foods . . . may cause the food to be adulterated because these added
substances may not be GRAS for their intended use and are not being used in accordance with a food
additive regulation prescribing conditions of safe use”).

61

 Many flavorings have been determined to be GRAS by industry despite the formal
designation of a government authority that they induce cancer when ingested by man or animal.
Most of these GRAS determinations were originally made in the 1960s, and were not reassessed
when later scientific studies revealed the problems. For example:


2-Phenylphenol/o-Phenylphenol was determined as GRAS for use as a seasoning
or flavoring,314 despite being declared a carcinogen by California Prop 65 in
2000.315



Quinoline was determined GRAS in 1975 for confectionary and baked goods,
puddings, meats, and soups.316 In 2001, EPA IRIS found it likely to be
carcinogenic in humans, echoing California Prop 65’s finding that it was
carcinogenic in 1997.317 Industry did eventually remove the GRAS designation
for quinoline, but not until late 2014, nearly two decades after the Prop 65
finding.318



Trans,trans-2,4-hexadienal was found to be GRAS in 1974 and 2003;319 IARC
found it possibly carcinogenic to humans in 2012.320

It is unknown how many dangerous or carcinogenic substances have been selfdetermined as GRAS without any notification to FDA or the public.

314

FEMA Expert Panel and Staff (P. Newberne, R.L. Smith, J. Doull, V.J. Feron, I.C. Munro, P.S.
Portoghese, W.J. Waddell, B.M. Wagner, C.S. Weil, T.B. Adams and J.B. Hallagan), GRAS Flavouring
Substances 19 (FEMA No. 3906-3963), Food Technology, Vol. 54, No. 6, (June 2000).
315
California Office of Environmental Health Hazard Assessment (OEHHA), Chemicals Listed Effective
August 4, 2000 as Known to the State of California to Cause Cancer: o-Phenylphenol (2000), available at
http://oehha.ca.gov/prop65/out_of_date/800Not.html.
316
B.L. Oser and R.A. Ford on behalf FEMA Expert Panel, Recent Progress in the Consideration of
Flavouring Ingredients Under the Food Additives Amendment, GRAS Substances (FEMA No. 34453476), Food Technology, Vol. 29, No. 8 (Aug. 1975).
317
EPA IRIS, Quinoline (CASRN 91-22-5), http://www.epa.gov/iris/subst/1004.htm (last accessed
2/21/15); OEHHA, Evidence On The Carcinogenicity Of Quinoline And Its Strong Acid Salts (1997),
available at www.oehha.ca.gov/prop65/hazard_ident/pdf_zip/quinolin.pdf.
318
FEMA Upgrades GRAS Interim List, Removes Quinoline (Nov. 2014),
http://connection.ebscohost.com/c/articles/98900741/fema-updates-gras-interim-list-removes-quinoline.
319
B.L. Oser and R.A. Ford on behalf FEMA Expert Panel, Recent Progress in the Consideration of
Flavouring Ingredients Under the Food Additives Amendment: 8. GRAS Substances, (FEMA No. 34243444), Food Technology, Vol. 28, No. 9 (Sept. 1974); FEMA Expert Panel (R.L. Smith, S.M. Cohen, J.
Doull, V.J. Feron, J.I. Goodman, L.J. Marnett, P.S. Portoghese, W.J. Waddell, B.M. Wagner, and T.B.
Adams), GRAS Flavouring Substances 21 (FEMA No. 4024-4068), Food Technology, Vol. 57, No. 5
(May 2003).
320
IARC, 2,4-Hexadienal, IARC Monograph – Volume 101-012, 391-405 (2012), available at
http://monographs.iarc.fr/ENG/Monographs/vol101/mono101-012.pdf.

62

 Equally worrying, many industry-made GRAS determinations are based entirely on
unpublished safety data. For example, in 2013, the Flavor and Extract Manufacturers
Association (FEMA) issued GRAS determinations for four new flavorings:


FEMA GRAS No. 87: trans-6-Octenal (also known as (E )-6-Octenal or (E )-Oct-6enal)



FEMA GRAS No. 4798: 2-(((3-(2,3-Dimethoxyphenyl)-1H -1,2,4-triazol-5yl)thio)methyl)pyridine (also known as 2-((5-(2,3-Dimethoxyphenyl)-2H -1,2,4triazol-3-ylthio)methyl)pyridine or Pyridine, 2-(((3-(2,3-dimethoxyphenyl)-1H -1,2,4triazol-5-yl)thio)methyl))



FEMA GRAS No. 4802: (S)-1-(3-(((4-amino-2,2-dioxido1Hbenzo[c][1,2,6]thiadiazin-5-yl)oxy)methyl) (also known as piperidin-1-yl)-3methylbutan-1-one)



FEMA GRAS No. 4809: 2-(4-Methylphenoxy)-N -(1H -pyrazol-3-yl)-N -(thiophen-2ylmethyl)acetamide (also known as N -(1H -pyrazol-5-yl)-N-(thiophen-2-ylmethyl)-2(ptolyloxy)acetamide)

When nonprofit groups sought data on these new substances, the organizations found no
relevant published safety data, toxicology data, or exposure data.321 When the groups asked
FEMA for published data on the four substances, it responded that there was none.322 This
practice of relying on unpublished data cannot meet any definition of “general recognition” of
safety. As FDA is aware, FEMA fails to apply the published safety data requirement to their
GRAS determinations.323 That a trade group can further enlarge the already-oversized GRAS
loophole to extend to substances that have been labeled by scientific experts as known hazards to
human health is only more evidence that FDA’s current practices are profoundly inadequate
under the law.
ii. Mycoprotein (“Quorn”)
Mycoprotein is another example of a novel substance that poses obvious and evident
risks to consumers, yet was permitted to bypass the food additive petition process (and be
marketed to consumers) due to FDA’s overexpansive conception of GRAS. The case of
mycoprotein also illustrates the need for a more reaching definition of harm, to ensure that
GRAS determinations consider vulnerable subpopulations and effects that disrupt the ordinary
routines of consumers, rather than considering only acute health risks or irreversible harm.

321

Letter from NRDC, CSPI, CFS, Center for Environmental Health, and Environmental Working Group
to Michael Taylor Re: Flavors and FEMA GRAS Program (Sept. 18, 2014) (on file with CSPI).
322
Id. It did, however, offer to make 7,000 pages of unpublished data available to NRDC for a processing
fee of over $1,000.
323
Letter from NRDC, CSPI, CFS, Center for Environmental Health, and Environmental Working Group
to Michael Taylor, Deputy Commissioner for Foods, FDA, Re: Flavors and FEMA GRAS Program (Sept.
18, 2014), at Appendix C (on file with CSPI).

63

 Mycoprotein is a manufactured form of fungus that is the base of Quorn-brand meat and
poultry substitutes. The ingredient was developed in the United Kingdom and originally
marketed there in 1985. Marlow, the manufacturer of Quorn foods petitioned the FDA in 1986
to approve mycoprotein as a food additive. However, because the FDA never reached a final
decision on that petition, in 2001 the manufacturer notified FDA that, consistent with the FDCA,
it had concluded that mycoprotein was GRAS. The notice was supported by an analysis by a
panel of “experts” convened by Marlow. In 2002, FDA issued a “no questions” letter in
response to the GRAS notice,324 but has never formally approved the product as a food
additive.325
Despite its supposed GRAS status, adverse effects associated with Quorn are frequent.
An analysis of 2,007 adverse reaction reports and previously published papers found that the
“mycoprotein” in Quorn-brand meat substitutes causes food intolerances and allergic reactions.
Gastrointestinal (GI) symptoms, including vomiting and diarrhea, typically occurred 1–3 hours
after consumption of the product.326 Symptoms ranged from mild nausea to vomiting severe
enough to warrant medical attention.327 Some consumers “said they broke out in hives and had
breathing difficulties – anaphylactic reactions.”328 The GI episodes associated with Quorn were
described as “violent”; victims “would vomit so hard it could break the blood vessels in their
eyes.”329 It is not known if the reactions are due to food allergies, intolerances, or both. The
analysis also found that the “mycoprotein” in Quorn-brand meat substitutes caused apparent
allergic reactions within 4 hours in 312 people. One of those individuals died after eating a
Quorn food.
In addition, a double-blind, placebo-controlled study conducted in 1977–78 by the
developer of mycoprotein, found that nine of 200 people who were fed mycoprotein two times in
324

Letter from Alan M. Rulis, Director, Office of Food Additive Safety, Center for Food Safety and
Applied Nutrition, FDA, to Stuart M. Pape, Attorney for Marlowe Foods (Jan 7, 2002), available at
http://www.fda.gov/Food/IngredientsPackagingLabeling/GRAS/NoticeInventory/ucm154623.htm.
325
The no questions letter suggests that the GRAS notice replaced the food additive petition:
In 1986, when Marlow submitted FAP 6A3930, most of the data and information that
Marlow relies on to conclude that the intended use of mycoprotein is safe were not
generally available to the public. Given the circumstances in 1986, there was no basis at
that time to consider whether the intended use of mycoprotein could be GRAS. In
contrast, at the present time most of the data and information that Marlow relies on to
conclude that the intended use of mycoprotein is safe are generally available to the
public. In addition, these data and information have been reviewed by a panel of
individuals that Marlow considers qualified by scientific training and experience to
evaluate the safety of substances added to food. Given the circumstances at present, FDA
has no questions about Marlow's view that mycoprotein is GRAS under the conditions of
its intended use.
Id. The regulatory timeline for Quorn is an explicit example of how industry-made GRAS
designations have entirely replaced the statutory food additive petition process.
326
Michael F. Jacobson, Janna dePorter, & Qian Yuan, Food Intolerances or allergic reactions linked to
Quorn-brand foods (forthcoming).
327
Id.
328
Kindy, supra note 5 (quoting Michael Jacobson).
329
Id.

64

 each of four weeks suffered GI symptoms that appeared to be due to mycoprotein, including five
“minor” symptoms and four “more severe” (mostly violent vomiting, difficulty breathing, severe
abdominal pains) or a rash, while only one out of 100 people fed a control food reported
significant symptoms.330
These reports and studies raise real questions about whether Quorn is safe for use in
foods as an additive at all – particularly as it is a non-essential product. However, whether or not
these adverse effects should prevent its use in foods altogether, it is clear that they raise
sufficient questions about the safety of Quorn to preclude any general consensus or “general
recognition” of safety in the scientific community and thus should have been subject to the full
food additive petition process. As with caffeine, the GRAS status of Quorn also demonstrates
the need for FDA to provide a definition of “harm” that takes less-serious, but still disruptive,
effects into account when reviewing the GRAS designation of substances added to foods.
VI. In finalizing the rule, FDA must limit the categories of substances eligible for GRAS
determinations and set clearer requirements for the scientific standards.
As the foregoing makes clear, FDA cannot legally finalize the rule as proposed, as it
violates Congress’s unambiguous intent of the FAA and would be arbitrary and capricious under
the APA in light of the substantial evidence already before the agency. Instead, FDA must revise
its proposal and reopen the docket in order to establish a meaningful regulatory system that
accomplishes the requirements of the FAA as intended by Congress.
Specifically, FDA must strengthen its oversight of substances added to foods in two major
respects. First, FDA must limit the GRAS safety determinations with binding standards and
establish an effective enforcement regimen for such standards. Under the current scheme, the
GRAS exemption has nullified the food additive petition process under the FAA and rendered
the law toothless. To fix this, the current rulemaking should delimit eligibility for GRAS and set
clearer requirements for the “general recognition of safety” standard to ensure that “food
additives” are subject to the requirements of the FAA.
Second, FDA must make GRAS notifications both mandatory and public. Due to
missing data on the composition of the food supply, neither FDA nor the food industry can
actually conduct the safety evaluations required by the law for both GRAS substances and food
additives. It should also establish a requirement for updated GRAS notifications to monitor the
use of chemicals in the food supply, so that the agency, industry and public can use common
exposure baselines for chemicals and classes of chemicals.
a. FDA should limit the GRAS concept as Congress intended.
First, FDA must address the limits on GRAS in its rulemaking. The 1997 proposal has
nullified the food additive regulatory system under the FAA by allowing GRAS to overtake the
“food additive” definition. FDA should narrow the GRAS exemption by issuing a final rule,
330

R.H.M. Research Limited, Large scale volunteer trial on mycoprotein (A3/5) (1977); R.H.M.
Research Limited. Supplementary studies on subjects who consumed A3/5 mycoprotein in a large scale
volunteer trial conducted November/December 1977 (1978) (Documents on file at the FDA).

65

 binding on future GRAS determinations, that: (1) GRAS cannot apply to novel ingredients since
a scientific consensus has not yet developed; (2) GRAS cannot apply to substances flagged as a
risk to human health by authoritative scientific bodies; (3) GRAS status must be based on a
regulatory definition of “harm,” to be developed by FDA; (4) GRAS determinations must be
based on published and peer-reviewed science; (5) GRAS must be evaluated based on adequate
science for the chemicals being assessed and their anticipated maximum exposure levels; and (6)
GRAS must be determined by experts without conflicts of interest.
i. GRAS cannot apply to novel ingredients since a scientific consensus
regarding their safety has not yet developed.
As detailed above, the legislative history of the FAA unambiguously indicates that the
GRAS exemption was never intended to apply to novel or “newly discovered” substances. FDA
has an affirmative obligation to more narrowly define the scope of the GRAS exemption going
forward in order to ensure that all novel food additives are subject to the food additive petition
process as intended by Congress in passing the FAA. Neither the existing nor the proposed
regulations delineate this category; to maintain consistency with the statute, FDA must
promulgate a final rule making clear that novel substances cannot be self-determined GRAS.331
As noted above, courts interpreting an identical legal standard in the drug context have drawn
this self-evident and obvious conclusion.332 To meet the “general recognition of safety” standard
in the FAA, GRAS chemicals must be familiar substances whose safe use in the food supply is
generally known and well established.
ii. GRAS cannot apply to substances flagged as risks to human health by
authoritative scientific bodies or questioned in the scientific
community.
In addition to excluding novel additives from the GRAS exemption, FDA must further
define the extent of scientific agreement required to meet the GRAS standard. FDA and the
courts have interpreted the FAA to require a high level of scientific consensus for a GRAS
determination.333 However, such consensus has not been required in practice. The standard is a
For example, Canada defines a “novel food” in its regulations, in part, as “(a) a substance, including a
microorganism, that does not have a history of safe use as a food; (b) a food that has been manufactured,
prepared, preserved or packaged by a process that (i) has not been previously applied to that food, and
(ii) causes the food to undergo a major change. . . .” C.R.C., c. 870 § B.28.001.
332
See, e.g., United States v. An Article of Drug . . . "Bentex Ulcerine", 469 F.2d 875, 880 (5th Cir. 1972)
(observing that if a substance is generally recognized as safe, one would expect to “find in the medical
literature over a period of years support for this premise from wide experimentation and study”); United
States v. Articles of Drug . . . Hormonin, 498 F. Supp. 424, 431-32 (D.N.J. 1980) aff'd sub nom. Appeal of
Carnrick Labs., Inc., 672 F.2d 902 (3d Cir. 1981) and aff'd sub nom. United States v. Articles of Drug,
672 F.2d 904 (3d Cir. 1981) (finding that “unawareness of a product among qualified experts precludes a
finding of general recognition”).
333
21 C.F.R. 170.30(a) (1997) (GRAS “requires common knowledge about the substance throughout the
scientific community knowledgeable about the safety of substances directly or indirectly added to food”);
Weinberger v. Hynson, Westcott & Dunning, Inc., 412 U.S. 609, 632 (1973) (interpreting “generally
recognized” language in the drug context to require “expert consensus” based on “substantial evidence”);
United States v. An Article of Drug Consisting of 4,680 Pails, More or Less, Each Pail Containing 60
331

66

 general recognition of safety by competent scientists; a substance cannot possibly be GRAS if it
has been identified as harmful by authoritative entities within the scientific community.
In the 1997 proposal, FDA explains that “general recognition” requires consensus in the
scientific community without “severe” conflict.334 This suggests that the “general recognition”
standard is not merely whether a scientist compensated by the food industry decides that
something is safe; instead, to be GRAS, the safety of a substance is demonstrated by general
agreement among the scientific community.335 To that end, FDA must issue a rule stating that
the GRAS exemption cannot apply to any substance whose safety has been called into question
by expert authorities. This is essentially a self-executing requirement under the law, regardless
of FDA’s role.336 The statute does not require any agency review or pronouncements on the
matter; it simply is the case that GRAS eligibility is unavailable for substances that lack a
consensus of safety among knowledgeable experts.
However, given the decades of agency neglect of this clear statutory principle, FDA
should acknowledge this limitation explicitly and make clear its meaning: if certain authoritative
bodies,337 such as the U.S. Environmental Protection Agency, the Institute of Medicine, the
European Food Safety Authority, the National Toxicology Program, the International Agency for
Research on Cancer, or the State of California (or other state authorities), have declared a
substance to be a hazard to human health or a carcinogen, the substance cannot be selfdetermined GRAS, as such a conclusion is definitive proof of serious discord among
“knowledgeable experts” and constitutes an absence of a “general recognition” of safety.338
This does not mean that an additive flagged by an authoritative body is automatically
barred from use in foods; it would simply fall within the definition of a food additive under the
statute as not GRAS and need to go through the formal food additive petition process to be
approved for use in food. This would narrow the GRAS exemption as intended by the statute
without requiring additional work by FDA in evaluating the safety of the most-disputed, and
potentially dangerous, substances being proposed for entry into the food supply. It would also
Packets, Etc., 725 F.2d 976, 985 (5th Cir. 1984) (“‘[G]eneral recognition’ requires a two-step showing:
first, that there is general recognition in fact, i.e., that there is an expert consensus that the product is
effective; and second, that the expert consensus is based upon “substantial evidence” as defined in the Act
and in FDA regulations.”); United States v. Undetermined Quantities of Various Articles of Drug . . .
Equidantin Nitrofurantoin Suspension, 675 F.2d 994, 1000 (8th Cir. 1982) (finding “a genuine dispute
concerning the safety and effectiveness of a drug product . . . precludes a finding of “general
recognition”) (quoting United States v. Articles of Drug (Hormonin), 498 F.Supp. at 431-32).
334
62 Fed. Reg. at 18939
335
Id.
336
Neal Fortin, PHOs: Illegal Additives Since 2005, Food Technology, Nov. 2014, at 92.
337
FDA has already acknowledged, in the 1997 proposal, the role such bodies can play in supporting a
GRAS finding. 62 Fed. Reg. at 18941. They should have equal force in disproving a general recognition
of safety.
338
Indeed, FDA has already acknowledged safety conclusions in its decisions in the past, including its
proposal to de-GRAS trans fat, which relied on an Institute of Medicine analysis on PHOs. 78 Fed. Reg.
at 67,169. FDA concluded that the IOM’s questioning of the safety of PHOs meant that the substance
was no longer “generally recognized as safe” in the scientific community. Id. FDA should formalize this
practice in a final rulemaking.

67

 immediately and definitively bring the intended and much-needed lucidity to the regulatory
scheme. For clarity (rather than because it is statutorily required), FDA should also maintain and
publish a regularly updated list of bodies and determinations that it explicitly incorporates into
this standard and invite the public to submit additional bodies and determinations for its list.
iii. FDA must provide a definition of what constitutes “harm” to human
health.
Safety is defined in the regulations as “a reasonable certainty in the minds of competent
scientists that the substance is not harmful under the intended conditions of use.”339 However,
there is no corresponding definition of what may be considered “harmful.”340 The examples of
mycoprotein and caffeine illustrate an immediate need for a definition of harm that protects
vulnerable subpopulations and encompasses transient effects that disrupt the ordinary routines of
consumers. The definition of harm should not be interpreted narrowly to mean only effects that
require medical intervention. Instead, the definition should include both transient and permanent
effects, as well as acute and chronic effects.
Whether or not such adverse effects should prevent use of a substance in foods, it is clear
that they raise sufficient questions to preclude a general consensus or “general recognition” of
safety in the scientific community. Thus, by failing to define “harm” under its regulations, FDA
has provided an additional means of abusing the GRAS concept. FDA should issue a definition
of harm that ensures that effects which are transient or affect only vulnerable subpopulations are
taken into account when determining the GRAS designation of substances added to foods.
iv. GRAS determinations cannot be based on unpublished research and
must instead be based on peer-reviewed, published science.
FDA should also clarify the standard for “general recognition,” to help industry identify
the types of substances that cannot fall within the category. General recognition of additive
safety requires consensus in the scientific community. Such consensus is impossible to achieve
if a chemical’s use is unknown to the scientific community and to FDA. Thus, the general
recognition standard necessarily requires a basis in peer-reviewed and published science.341
Under the previous regulations, “[g]eneral 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.”342 The 1997 proposal expanded the language to include
“generally available and accepted scientific data, information, methods, or principles, which

339

21 C.F.R. § 170.3(i).
Beyranevand, supra note 26, at 889.
341
United States v. An Article of Drug Consisting of 4,680 Pails, 725 F.2d at 987 (finding that the
substantial evidence required for a general recognition of safety “consists of adequate and well-controlled
studies that must be generally available to the scientific community”).
342
21 C.F.R. § 170.30.
340

68

 ordinarily are published.”343 This revision elevates unpublished studies from corroboration to
primary support.344
To meet the requirements of general recognition, FDA must take steps to ensure that all
GRAS determinations are based on published and peer-reviewed studies. That recommendation
is supported by comments in the docket submitted by other interest groups; for example, NRDC
noted the need for published toxicological and exposure data in making GRAS determinations.345
Thus, in its final rule, FDA should restore the pre-1997 requirements for actual studies, both
toxicological and exposure, to support GRAS determinations, and the primacy of published
studies in furnishing sufficient evidence. This proposal could be accomplished on the current
docket, as FDA’s proposal has been opposed by commenters, and returning to the pre-1997
standard would therefore be a “logical outgrowth” of the pending proposal.346
v. GRAS must be evaluated based on adequate science for the chemical
being assessed.
FDA must also strengthen the evidentiary requirements for GRAS determinations. As
FDA states in its existing regulations, the standard of safety for GRAS substances is identical to
the standard that applies to food additives.347 Therefore, a manufacturer’s decision must be
consistent with the food additive regulations and comply with FDA’s guidance on food additive
testing, including the Redbook and other relevant sources from the agency. However, this is not
the case in practice.
Under FDA guidance, safety evaluations for chemicals added directly to food348 involve
assigning the substance to a Concern Level (i.e., low (I), intermediate (II) or high (III)) based on
the substance's toxicological potential predicted from its chemical structure and an estimation of
cumulative human exposure.349 Yet, according to Pew, only 1 percent of GRAS notifications
even identified a concern level, and most GRAS substances are not tested for safety even with

343

74 Fed. Reg. at 18960.
This is illustrated by the FEMA letter, which notes that the GRAS determinations for four flavorings
made in 2013 were based on no published studies, exposure data, or toxicological data. Letter from
NRDC, CSPI, CFS, Center for Environmental Health, and Environmental Working Group to Michael
Taylor, Deputy Commissioner for Foods, FDA, Re: Flavors and FEMA GRAS Program (Sept. 18, 2014),
at Appendix B (on file with CSPI).
345
NRDC Comment, supra note 252, at 2; see also International Center for Technology Assessment,
supra note 249, at 3 (noting need for peer-reviewed published studies).
346
Envtl. Integrity Project v. E.P.A., 425 F.3d 992, 996 (D.C. Cir. 2005) (“An agency's proposed rule and
its final rule may differ only insofar as the latter is a ‘logical outgrowth’ of the former.”).
347
21 C.F.R. § 170.30(b) (GRAS status based on scientific procedures “require[s] the same quantity and
quality of scientific evidence as is required to obtain approval of a food additive regulation for the
ingredient.”).
348
This does not include indirect additives such as those used in packaging.
349
Food and Drug Administration, Guidance for Industry: Summary Table of Recommended
Toxicological Testing for Additives Used in Food (June 2006), available at
http://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/IngredientsA
dditivesGRASPackaging/ucm054658.htm.
344

69

 FDA’s lowest concern level testing recommendations.350 FDA must thus emphasize that the
science underlying GRAS safety determinations must be the same as for food additives, and that
it will evaluate submissions on this basis.
In addition, as detailed above, FDA’s standards for evaluating the safety of additives are
severely out of date. In its final rule, FDA must require testing for potential endocrine
disruption, behavioral effects, and developmental neurotoxicity, and include additional safety
requirements for vulnerable subpopulations (such as infants, pregnant women, the elderly, people
with compromised immune systems, diabetes, alcoholism, or food sensitivities, among others).
FDA must also require absorption, distribution, metabolism, and excretion (ADME) assessments.
These standards should be based on the Redbook and other governmental guidance on modern
risk assessment principles and updated on a regular basis to ensure that they track contemporary
scientific knowledge. FDA should also provide a clear guide for how to assemble and evaluate
the scientific evidence underlying a GRAS determination.
vi. GRAS determinations must be made by experts without conflicts of
interest.
The GRAS exemption is meant to apply only where there exists a “consensus” among
scientists that the use is generally recognized as safe. Under the 1997 proposal, companies can
demonstrate this “consensus” in a variety of ways, such as scientific review articles, expert
panels, reports by authoritative bodies, or by some combination of these methods.
Problematically, the 1997 proposal treats industry-funded “expert panels” as the equivalent of
more public and inherently credible forms of assuring safety, such as determinations by
authoritative bodies and peer-reviewed published articles. As a result, food manufacturers often
make GRAS safety decisions by relying solely or mainly on their own experts, despite the
inherent conflicts of interest.351
FDA has not issued any guidance to minimize potential for conflicts of interest for
companies to help ensure that the members of their expert panels are independent in their
determinations of GRAS status.352 Furthermore, because FDA does not require companies to

350

For example, agency guidelines say that chemicals intentionally added to food should be fed to
laboratory animals to identify potential harmful effects, but Pew found that in the majority of cases,
industry-determined GRAS chemicals did not undergo this very basic type of testing. Neltner et al. Data
gaps, supra note 224, at 90; see also McQuate & Kraska, supra note 6 (“The Redbook identifies many
tests that may be needed to assess the safety of a food additive; however, most new food ingredients that
have successfully emerged from GRAS notifications do not have the full complement of tests presently
prescribed by the Redbook.”).
351
Elizabeth Weise, Experts who decide on food additives conflicted, USA Today, Aug. 18, 2013,
http://www.usatoday.com/story/news/nation/2013/08/07/food-additives-conflict-of-interest/2625211/
(“The companies hire a consulting firm to get experts for them and then the experts review the
information that's available and then they write a letter to FDA saying this additive should be considered
GRAS,” says Marion Nestle, a professor of nutrition at New York University. “There are whole
companies that are in the business of recruiting scientists to sign off on these things,” she says. “That was
one of the amazing findings of this paper.”)
352
GAO Report, supra note 37, at 14.

70

 provide information about independence in their GRAS notices, FDA “does not know whether
the determinations of companies’ expert panels are arrived at independently.”353
FDA regulations also state that GRAS “may be based only on the views of experts
qualified by scientific training and experience to evaluate the safety of substances directly or
indirectly added to food.”354 However, the agency has never explained how to identify the pool
of potential experts that represent the scientific community and ascertain whether they are
“qualified” to evaluate the safety of a substance’s use in food.
Thus, the current system thus provides no assurances the safety of GRAS substances are
actually “common knowledge,” as required by law – and conflicts of interest are ubiquitous.355
As FDA’s Michael Taylor acknowledged to The Washington Post in August 2014, saying with
regard to a company’s self-determination of EGCG’s safety (see, infra, at 55):
“This is the opposite of what the over[s]ight [sic] law intended,” the FDA’s
Taylor said. “Although informing the FDA is voluntary,” he said, “the law was
meant to increase public scrutiny of additive safety by encouraging companies to
publish their science in academic journals.” . . . “The assessments need to be
based on publicly available information where there is agreement among
scientists,” he said. “It has got to be more than three employees in a room looking
at information that is only available to them.”356
This 1997 system does not ensure that so-called GRAS substances are actually “generally
recognized” as safe under the law. This is inconsistent with published case law on the identical
standard in the drug context, which states that the evidence required for a general recognition of
safety “does not consist of the expressed opinions of experts hired to testify on behalf of one
party or the other.”357
The agency has indicated an interest in undertaking such a review of policy on
conflicts.358 Rather than issuing more toothless voluntary guidance, however, FDA should
publish its policy as part of a binding rule. The issue is central to the credibility of GRAS
determinations and deserves treatment as such.
As mentioned in the GAO report, the devastating review by Pew of conflicts issues, and
by other commenters to this docket, the notion that regulated industry is well equipped to
undertake an objective scientific study of the very products it is seeking to bring to market is

353

Id. at 15.
21 C.F.R. § 570.30.
355
GAO Report, supra note 37, at 15 (noting that the small community of experts means that panels
inevitably have corporate or financial affiliations that could bias their decisions).
356
Kindy, supra note 5.
357
United States v. An Article of Drug Consisting of 4,680 Pails, 725 F.2d at 987.
358
McQuate & Kraska, supra note 6 (noting that FDA indicated a commitment to issuing guidance on
conflicts of interest by end of 2014).
354

71

 naïve at best and dangerous to consumers at worst.359 Recent research documents significant
effects related to funding bias in the outcomes of studies, and industry sources of funding for
research have been found to have significant impacts on outcomes of studies.360 Bias is apparent
not only in outcomes, but in the myriad decisions that go into the categorizations and questions
asked, and the multiple other informed judgments that shape study design.361
This is a particular problem with regard to the secret GRAS process. Funding bias is
both inescapable and utterly unchecked in such a system. For these reasons, it is incumbent upon
the agency to craft a system of oversight for GRAS self-determinations that does the most it can
do, given the industry’s clear interests and its role as a source of funding for these studies, to
counteract the funding bias that would otherwise undermine the integrity of the GRAS program.
Pew, with input from multiple reviewers and stakeholders in a public August 2013
workshop of experts, advocates and industry participants, drafted guidance on avoiding conflicts
of interest in GRAS evaluations and submitted it to the current docket on September 4, 2013.362
We recommend that FDA adopt the suggestions from the Pew guidance document submitted to
this docket in 2013 as a binding final rule.363 These measures, made enforceable and with the
agency’s active oversight, would help to ensure that experts have necessary expertise, are
sufficiently removed from the issue to ensure objective assessment of scientific consensus, and
do not have financial or non-financial conflicts of interest.
We further recommend that FDA require any company that conducts a GRAS
determination to provide FDA and the public with critical information about the experts who
participated in making the determination, their sources, and their findings, as well as
compensation arrangements. The need for this part of the final rule is articulated by other
comments on the docket, which asserted that GRAS notices should include information
demonstrating the independence of experts who generate data or analysis underlying the GRAS
finding, as well as the Pew guidance draft.364
In addition, FDA must establish an active role in monitoring and evaluating the
sufficiency of steps taken to address conflicts of interest in GRAS determinations. As noted
above, the literature on funding bias is clear and shows the skewing effect of funding
relationships upon outcomes; the current GRAS system is far worse than the research on general
bias, which concerns pee-reviewed, published study outcomes, and demonstrably rife with both
actual and potential conflicts. FDA therefore should develop objective measures for bias in

GAO Report, supra note 37, at 34 (calling on FDA to “[m]inimize the potential for conflicts of interest
in companies’ GRAS determinations, including taking steps such as issuing guidance for companies on
conflict of interest and requiring information in GRAS notices regarding expert panelists’ independence”)
360
See sources cited supra note 287.
361
David Michaels, It's Not the Answers That Are Biased, It's the Questions, The Washington Post, July
15, 2008, http://www.washingtonpost.com/wp-dyn/content/article/2008/07/14/AR2008071402145.html.
362
Pew Charitable Trusts, Draft Guidance for Industry, supra note 237.
363
Id.
364
Food & Water Watch, supra note 246, at 3; Hansen, supra note 247, at 5.
359

72

 outcomes based on the funding bias literature, monitor conflicts in GRAS determinations to
assess such bias, and publish its findings in a public report.
Should adoption of a binding, mandatory regime of disclosure and substantive checks on
conflicts prove to be insufficient to address bias in outcomes, FDA must develop a more
proactive approach to preventing predictable conflicts of interest. Either through its existing
regulatory authority or by soliciting authority from Congress, it should establish a group of
credentialed and independent expert reviewers for GRAS notifications, funded by pooling
industry user fees, and base GRAS determinations on their reviews. Ex parte communication
and funding between the pool of experts and the food industry should be barred. As the Pew
report identifies, this is the gold standard for safety. There is considerable doubt that any
industry-funded process – however well monitored by FDA – could produce unbiased scientific
determinations sufficient to assure public safety.
b. FDA must require public notice of GRAS determinations to assure that the
burdens of general recognition and ongoing assessment of safety are met.
Even these appropriate limits on the GRAS category will not provide meaningful
protection to the public health unless FDA is notified about such substances. In addition to
limitations on what may be found GRAS, FDA must make changes to the system to facilitate
ongoing assessment and oversight over existing GRAS substances as intended by Congress
under the FAA. To achieve these ends, FDA must: (1) make GRAS notifications mandatory; (2)
require public notice of all GRAS determinations; (3) require regular submission and public
notice of updated and public exposure assessments; and (4) require documentation and
recordkeeping of GRAS determinations.
i. To achieve Congress’s intent under the FAA, GRAS notifications
must be mandatory.
Under FDA’s proposed notification system, a majority of “food additives” are never
declared to the agency. Without knowledge of “GRAS” substances currently in use, FDA cannot
execute its ongoing oversight over substances added to foods as intended by Congress in passing
the FAA. Moreover, no “general recognition of safety” standard can be satisfied where exposure
data is unknown.
As detailed above, Congress intended FDA to undertake an ongoing review of all
substances added to foods to ensure they remained safe in the face of updated science and
exposure.365 Congress was clear that GRAS status could always change in light of new
information on safety and risks. Moreover, the legislative history of the FAA indicates that this
365

104 Cong. Record, 85th Cong, 2d Sess. 17412, 17420 (Aug. 13, 1958) (Letter from John L. Harvey,
Deputy Commissioner of FDA) (explaining that “if later developments show that a substance previously
considered safe because of common use in food were, in fact, subject to question and not generally
recognized as safe, then the substance would become subject to the definition of a food additive and
would have to be cleared under the procedures of the proposed law”). This has been recognized by FDA
in its regulations, which state that [n]ew information may at any time require reconsideration of the
GRAS status of a food ingredient.” 21 C.F.R. § 170.30(k)(1) (2012).

73

 was precisely the type of oversight role that Congress intended FDA to develop – to monitor the
potential risks related to long-term exposure of substances added to foods. However, the lack of
required notification under the current system makes this, practically speaking, impossible, as
FDA is only made aware of substances when and if industry chooses to submit a voluntary
notification.
Moreover, even with respect to GRAS substances about which industry has notified the
agency, FDA is unable to assess the risks because it is unaware of actual usage and exposure, of
studies and adverse events that indicate possible (and previously unknown) adverse health
impacts, or of data for classes of chemicals or chemicals that act similarly in the body. Because
FDA lacks basic information needed to assess the safety of thousands of chemicals currently
used in food, it cannot fulfill its statutory obligation to make ongoing reevaluations of the safety
of such substances.
FDA’s proposal for a voluntary notification system certainly allows the agency to review
some GRAS determinations; however, because it is not mandatory, this review is limited. FDA
generally will not be alerted to “secret” GRAS substances in the marketplace or their dietary
exposure.366 This means that once a secret GRAS substance has entered the marketplace, FDA
would find it difficult to identify that substance as the potential source of a food safety
problem.367 By making GRAS notifications voluntary, FDA is “relinquishing a vital component
of its important role in food safety oversight [and] creating a situation in which public oversight
of food additives is virtually impossible.”368
The lack of notification also makes it difficult or impossible for FDA to make ongoing
safety assessments. FDA has claimed at times that it can address safety concerns that arise postmarketing by responding to concerns raised by the public or industry. In theory, information
brought to the agency’s attention could prompt the agency to reconsider the safety of a GRAS
substance, as it has done with PHOs. However, this mechanism fails under the current system in
several respects. First, it is entirely reactive – safety is only assessed after a danger has arisen
and the public has already been exposed to its risks. This is directly contrary to the purpose of
the law, which was intended to fix a system in which risks could only be identified after
substances were already sold to consumers. Second, industry and the public cannot alert FDA to
safety concerns when they are not aware of those concerns themselves. This is particularly
problematic given that these ingredients may or not be listed on labels (or listed under unfamiliar
names), making it difficult for anyone to link their exposure to adverse effects.
We thus agree with GAO’s conclusion that “FDA’s oversight of additive safety would
thus be improved if companies were required to make the agency aware of their GRAS
366

For example, officials representing one international marketer of food indicated the company makes
about 5 GRAS determinations each year without notifying FDA. These are usually new uses of
substances that have been deemed GRAS for other uses. In another case, a company began marketing a
purified version of stevia, a plant-based sweetener, as a GRAS substance before submitting a notice to
FDA and before FDA had indicated it had no questions about other GRAS notices related to stevia. GAO
Report, supra note 37, at 12.
367
Id. at 12.
368
International Center for Technology Assessment, supra note 249, at 1.

74

 determinations” because it would provide FDA with at least some information about GRAS
substances to help with investigations of food safety problems that may arise after such
substances have already been marketed.369 Consumers Union similarly found that “companies,
at a minimum, should be required to notify FDA of any GRAS determinations they make.”370
We urge the FDA to heed our concerns, shared by the GAO and Consumers Union, and make
GRAS notifications mandatory. Only mandatory notification would enable FDA to meet its
obligations under the FAA to maintain ongoing oversight over GRAS substances in order to
ensure their continuing safety in the face of developing science and exposure levels.
ii. Mandatory notifications must include accurate and current exposure
data.
Under the current regulations, GRAS status based on scientific procedures “require[s] the
same quantity and quality of scientific evidence as is required to obtain approval of a food
additive regulation for the ingredient.” 371 The 1997 proposal clarified that a GRAS substance is
“neither more safe nor less safe than approved food additives.372 For food additives, “safe or
safety” means “a reasonable certainty in the minds of competent scientists that the substance is
not harmful under the intended conditions of use.”373 In making this determination, the agency
must consider the probable consumption of the substance and its cumulative effect in the diet,
taking into account any chemically or pharmacologically related substance or substances in such
diet.374
The law, regulations, and guidance make clear that a GRAS determination requires an
exposure assessment based on the probable consumption of the substance and its cumulative
effects. As risk a function of both hazard and exposure, any “general recognition of safety”
cannot be determined without knowledge of exposure. However, without required notice of
GRAS determinations, the probable consumption of the substance and its cumulative effects
cannot be measured by experts, industry, or FDA – and the safety of the substance cannot be
“generally recognized.” To achieve the requirements and purpose of the statute, FDA and
industry must be able to estimate dietary exposure for all GRAS substances.
Thus, FDA must require companies to publicly submit exposure data along with their
GRAS determinations. A finding of general recognition of safety requires consensus in the
scientific community based on accurate toxicology and exposure data, including natural
exposure. General awareness of the substance and its exposure can only be achieved when FDA
is notified by industry about the use and existence of GRAS substances, and their overall dietary
exposure. To facilitate accurate exposure data and ongoing safety reassessments, FDA also must
require companies to regularly update their exposure assumptions in light of new information
and to assert under law that no new scientific information or adverse events reported by

369

GAO Report, supra note 37, at 12.
Hansen, supra note 247, at 1.
371
21 C.F.R. § 170.30(b).
372
62 Fed. Reg. at 18942.
373
21 C.F.R. § 170.3(i).
374
21 U.S.C. § 348(c)(5); 21 C.F.R. § 170.3(i).
370

75

 consumers related to that substance have come to light, particularly with regard to potential
health harms.375
This requirement is supported by prior comments to the docket. For example, NRDC
recommends that a GRAS safety assessment include published exposure assessments.376 Food &
Water Watch states that companies should be required to submit exposure data in their GRAS
determinations so that industry, FDA, and citizens can make a comprehensive assessment of the
safety of substances.377 In its final rule, FDA should also outline a mechanism the agency will
employ to corral and standardize exposure estimates for GRAS substances that are of potential
public health concern.
Finally, FDA should require notifiers to provide regular updates of exposure and safety
data. Under the current regulations, “FDA does not know to what extent, or even whether,
companies track evolving scientific information about substances they have determined to be
GRAS.”378 FDA should thus require companies to periodically update their exposure
assumptions in light of new information in the database and to assert under law that no new
scientific or consumer safety information related to that substance has come to light. In
particular, FDA should require submitters to disclose adverse studies and information in GRAS
notifications as the regulations specify that an assessment of the safety of a food substance
involves “an evaluation of information about its safety and functionality including all studies and
tests of a food additive on animals and humans and all studies and tests of a food additives for
identity, stability, purity, potency, performance and usefulness.”379
iii. Mandatory GRAS notifications must be public.
In addition to requiring submission of notification and exposure data, FDA should also
require public notice of all GRAS notifications and the information provided therein. GRAS
determinations made without notice to the agency (of which there are many) remain outside a
third party’s independent evaluation, because they are not made aware of its use or existence.
Academic experts, consumer groups, and scientific organizations do and should play an
important role in helping FDA oversee the safety of GRAS ingredients; indeed, independent
scientific research has contributed to FDA’s reconsideration of GRAS substances in the past.380
As GAO noted, “without knowledge of companies’ GRAS determinations, third parties, such as
the ones that have filed citizen petitions in the past, do not have the opportunity to investigate the
375

This requirement should apply to food additive petitions as well as GRAS notifications, for related
reasons.
376
NRDC Comment, supra note 252, at 2.
377
Food & Water Watch, supra note 246, at 3.
378
GAO Report, supra note 37, at 25. For example, representatives of one company told GAO that while
they review the status of their GRAS ingredients and keep up with the scientific literature, they do not
generally share these finding with FDA. Id.
379
21 C.F.R. § 171.1 (h(4)).
380
GAO Report, supra note 37, at 24. In addition, third party information may affect FDA decisions on
the GRAS notifications it does receive; under the proposed regulations, FDA may question a GRAS
notification if it is “aware of information that is not included in the notice but raises important public
health issues.” 62 Fed. Reg. at 18,950.

76

 potential health effects of such GRAS substances, leaving an additional gap in the oversight of
their continued safety.”381
As the ICTA pointed out in its 1997 comments, the current system “mak[es it] virtually
impossible for the public to fill the gaps left by the agency’s abdication of oversight review.”382
With “no easily accessible collection of data made available to the public,” making it difficult for
FDA, additive manufacturers, and public interest groups to monitor food safety. Furthermore,
given the lack of information provided to FDA, FOIA requests will yield limited information
because relevant documents remain with the producers. This is illustrated by CSPI’s FOIA
requests with regard to taste modifiers and caffeine. For taste modifiers, FDA was unable to
provide any information for some substances, and extremely limited information on the others.
Regarding caffeine, the only reason CSPI could obtain any information at all was because FDA
had conducted its own inquiry. The companies had not submitted any GRAS determination data
voluntarily. This is particularly troubling given FDA’s insistence that it makes GRAS
reassessments in response to concerns raised, in part, by the public interest community.383 Thus,
to facilitate ongoing oversight over GRAS substances, FDA must require industry to submit
public notice of its GRAS determinations and exposure data in its final rule.
FDA should also require companies to maintain active and accurate registrations for
GRAS substances in a public database in order to better track exposure data. This proposal is
supported by GAO, which recommended in its report that FDA issue regulations requiring any
company that conducts a GRAS determination to provide FDA with basic information about this
determination, such as the substance's identity and intended uses, and to incorporate such
information into relevant agency databases and its public Web site. Such public, consolidated
disclosures will enable FDA, GRAS notifiers, and the interested public to assess the safety of
new GRAS chemicals and proposed food additives based on accurate and current cumulative
exposure estimates.
iv. FDA should require manufacturers to maintain proper
documentation to support their GRAS determinations.
In its 1997 proposal, FDA stated that it would be prudent for companies to maintain
documentation of their GRAS determinations and suggested that FDA would monitor
compliance by conducting random audits of data and information maintained by the companies.
381

GAO Report, supra note 37, at 24. Even when outside groups are made aware of GRAS substances
and call attention to safety concerns, FDA has failed to respond to such concerns. While FDA claims that
information brought to its attention could prompt the agency to reconsider the safety of a GRAS
substance, it has historically not responded to the concerns of individuals and consumer groups. For
example, FDA has largely not responded to the concerns raised through 11 citizen petitions submitted
between 2004 and 2008. Nine of these 11 petitions raised specific concerns about the safety of GRAS
substances or the way they are used in food. As of 2012, FDA had only responded to 3 of these petitions
– all denials. Monica Eng, Citizen complaints not addressed by FDA, Chicago Tribune (Aug. 25, 2012),
http://articles.chicagotribune.com/2012-08-25/health/ct-met-food-ingredients-popcorn20120825_1_microwave-popcorn-fda-citizen-petitions.
382
International Center for Technology Assessment, supra note 249, at
383
GAO Report, supra note 37, at 21.

77

 In practice, FDA has never conducted such audits and thus has no way of ensuring that industry
conducts proper recordkeeping.384 Moreover, FDA has never explained what the proper level of
documentation would be.385 In finalizing the rule, FDA should develop criteria for the required
documentation underlying industry GRAS determinations. As FDA indicated an intention to
conduct random audits in the 1997 proposal, specifying rules for documentation would be a
reasonable outgrowth of the proposed rule. In addition, the agency should renew its commitment
to random auditing to ensure that companies maintain proper recordkeeping practices, and that
company records reflect data in FDA notifications.
c. FDA’s Final Rule should outline a comprehensive safety program.
In addition to fixing the process for GRAS substances moving forward, FDA must
address the secret GRAS determinations it has illegally permitted to enter the food supply. To
that end, it must bring them out of the shadows and classify them. There are many possible ways
to address the data gaps that FDA has created – and allowed industry to create – since 1997 and
prior. The food industry, recognizing the system’s clear weaknesses, has begun its own
approach, including developing a database, compiling safety information, and developing
standards.386 However, the law requires FDA to assure the safety of the food supply, and it is
both technically feasible and legally imperative that it take charge of any such endeavor. To
demonstrate that fixing the program is both possible and desirable, we propose some steps below
that the agency could take. We anticipate that development of a remedial program will be a
public process, and welcome ideas that assure public safety and secure the goals of the law.
To address the holes in the current system, FDA’s final rule should require submissions
of GRAS self-determinations in a reasonable time, with penalties for noncompliance, from all
food chemical manufacturers and food industry end users regarding their heretofore secret GRAS
self-determinations. These submissions should include the toxicological and exposure data
underlying these determinations. FDA should review the sufficiency of this documentation and
terminate the GRAS status of facially inadequate notifications, after a reasonable back-and-forth
with submitters, referring them instead, if appropriate, to the food additive petition process.
For substances in the current GRAS notifications database, FDA should require
submitters of substances that withdrew notices, or received an “insufficient basis” letter from the
agency, to notify FDA by a date certain about their use of that substance or other self-determined
GRAS substances and to submit appropriate documents regarding their safety determinations.
For submitters that received a “no questions” letter, FDA could prompt submitters to fill any
needed data gaps in submissions by a date certain to align submissions with the revised
documentation requirements. Previously completed GRAS affirmation or food additive petitions
should also be updated by companies according to the new documentation requirements.
All of these data should be incorporated into a modernized, centralized database. FDA
should also incorporate the data into a centralized data management system, such as the
Chemical Evaluation and Risk Estimation System (CERES), and use these databases to prioritize
384

Id. at 17.
Id.
386
See note 6 supra.
385

78

 toxicology and exposure data gaps and safety concerns, as it indicated it would do in the 1997
proposal. Once it has acquired use and exposure data, FDA will be able to identify and respond
to potential safety issues, including the long-term effects of substances and the potential
cumulative impacts of substances in combination with each other.
Most critically, the agency’s rulemaking could create and publish for comment a system
of risk-based review priorities for a combined database of GRAS and food additives. This
would, as the 1997 rule promised, direct FDA’s application of resources to the riskiest
substances in the food supply and equip the agency to assess the factors directed in the statute to
determine safe use. FDA should address the concern levels as follows:


For substances or uses above a specified risk concern level, based on the database,
FDA should clarify that the substance or use is not GRAS and require the notifier
to use the food additive petition process.



For classes of GRAS chemicals that pose risks to human health when their use
exceeds a certain level above current exposure, FDA should use the data to assess
and harmonize exposure levels as required by the law, for agency and industry use
in notifications. FDA should ask the food industry to submit an updated, but
simplified, set of data that would be needed to compute cumulative effect and
probable consumption levels for those substances, if the agency lacks such data.
FDA should compute naturally occurring exposures as well. These estimates
should also be subject to public comment. Should overall exposure exceed a
specified level of concern, FDA should open a public rulemaking on the GRAS
status of that chemical or class of chemicals.



Finally, for the lowest risk substances, FDA should compile and publish an
updated list of clearly non-hazardous GRAS substances based upon its review of
ingredient lists and industry and public submissions.

In addition, FDA should develop specific standards with regard to substances with novel
toxicology or that pose unique risks, for example, substances whose effects are not linear to dose
(such as endocrine disruptors), that pose specific concerns for vulnerable subpopulations, or that
pose risks in combination with other chemicals. The final rule should establish particular testing
requirements for such substances, as appropriate.
The database of secret GRAS substances and supporting documentation, like the GRAS
notification and affirmation processes, should be published, so that the public may assist the
agency in prioritizing the investigation of chemicals according to the exposure and risks to public
health, as the law intended. To keep these determinations current, FDA should ask for ongoing
public comments on which currently listed substances on GRAS lists are potentially hazardous
and should be reassessed. FDA should also assign a schedule for expiration of substances and
require renewal of notifications on an appropriate and periodic basis.

79

 VII. Conclusion: FDA Can and Should Solve the GRAS Problem
The 1997 proposal undermines the goal of the FAA: to protect public health. It fails to
ensure the safety of new or existing GRAS substances and allows food additives masquerading
as GRAS to avoid the agency’s oversight. FDA’s current system for GRAS is thus an abrogation
of its responsibilities under the FAA, and undermines consumer confidence in food additive
safety. If FDA were to finalize the rule as proposed, it would be a failure of its statutory
responsibilities; it must instead take this opportunity to revise the proposal in order to create a
regulatory system that meets the requirements under the statute.
Taken together, the changes we urge would transform the GRAS system from badly
broken to functional. FDA’s regulation of GRAS has waxed and waned since the 1958
Amendment, and our proposal provides an effective, even elegant, solution. Independent
determinations of the safety of substances in the food supply would remain, but would come out
of the shadows into the public eye, where they belong. By reasserting its legitimate power over
the scope and substance of safety determinations with reasonable definitions and requirements,
FDA could restore the rightful position of GRAS in the statutory scheme and reestablish public
confidence in the agency’s oversight. These changes would accomplish what FDA hoped for –
and failed to achieve – in its 1997 proposal. More importantly, they would equip the agency
with tools to effectively monitor public exposures and to analyze evolving questions impacting
public health in real time and proactively, just as Congress intended.

80