
[Federal Register Volume 80, Number 116 (Wednesday, June 17, 2015)]
[Notices]
[Pages 34650-34670]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2015-14883]


-----------------------------------------------------------------------

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Food and Drug Administration

[Docket No. FDA-2013-N-1317]


Final Determination Regarding Partially Hydrogenated Oils

AGENCY: Food and Drug Administration, HHS.

ACTION: Notice; declaratory order.

-----------------------------------------------------------------------

SUMMARY: Based on the available scientific evidence and the findings of 
expert scientific panels, the Food and Drug Administration (FDA or we) 
has made a final determination that there is no longer a consensus 
among qualified experts that partially hydrogenated oils (PHOs), which 
are the primary dietary source of industrially-produced trans fatty 
acids (IP-TFA) are generally recognized as safe (GRAS) for any use in 
human food. This action responds, in part, to citizen petitions we 
received, and we base our determination on available scientific 
evidence and the findings of expert scientific panels establishing the 
health risks associated with the consumption of trans fat.

DATES: Compliance date: Affected persons must comply no later than June 
18, 2018.

FOR FURTHER INFORMATION CONTACT: Mical Honigfort, Center for Food 
Safety and Applied Nutrition (HFS-265), Food and Drug Administration, 
5100 Paint Branch Pkwy., College Park, MD 20740, 240-402-1278, email: 
mical.honigfort@fda.hhs.gov.

SUPPLEMENTARY INFORMATION: 

Table of Contents

I. Background
II. Definitions and Scope, and Related Comments With FDA Responses
III. Discussion of Legal Issues, and Related Comments With FDA 
Responses
    A. GRAS
    B. Prior Sanctions
    C. Procedural Requirements
IV. Discussion of Scientific Issues, and Related Comments With FDA 
Responses
    A. Intake Assessment
    B. Safety
V. Citizen Petitions
VI. Environmental Impact
VII. Economic Analysis
VIII. Compliance Date and Related Comments With FDA Responses
IX. Conclusion and Order
X. References

I. Background

    In accordance with the process set out in Sec.  170.38(b)(1) (21 
CFR 170.38(b)(1)), we issued a notice on November 8, 2013 (the November 
2013 notice, 78 FR 67169), announcing our tentative determination that, 
based on currently available scientific information, PHOs are no longer 
GRAS under any condition of use in human food and therefore are food 
additives subject to section 409 of the Federal Food, Drug, and 
Cosmetic Act (the FD&C Act) (21 U.S.C. 348).
    FDA's evaluation of the GRAS status of PHOs centers on the trans 
fatty acid (TFA, also referred to as ``trans fat'') component of these 
oils. Although we primarily use the word ``oil'' when discussing PHOs 
in this document, partially hydrogenated fats (such as partially 
hydrogenated lard), are included within the definition of PHOs 
(discussed in section II) and therefore within the scope of this order, 
and references to ``oil'' in this document should be read in most cases 
to include fats. PHOs are the primary dietary source of industrially-
produced trans fatty acids (Ref. 1). As explained in the tentative 
determination (78 FR 67169), all refined edible oils contain some trans 
fat as an unintentional byproduct of their manufacturing process; 
however, unlike other edible oils, trans fats are an integral component 
of PHOs and are purposely produced in these oils to affect the 
properties of the oils and the characteristics of the food to which 
they are added. In addition, the trans fat content of PHOs is 
significantly greater than the amount in other edible oils. Non-
hydrogenated refined oils may contain trans fatty acids as a result of 
high-temperature processing, at levels typically below 2 percent (Ref. 
2). Low levels (below 2 percent) may also be found in fully 
hydrogenated oils (FHOs) due to incomplete hydrogenation (Ref. 3). 
Small amounts (typically around 3 percent) may be found in the fat 
component of dairy and meat products from ruminant animals (Ref. 4).
    FDA's tentative determination identified the significant human 
health risks associated with the consumption of trans fat (78 FR 67169 
at 67171). The tentative determination was based on evidence including 
results from a number of controlled feeding studies on trans fatty acid 
consumption in humans (Refs. 5 and 6), findings from long-term 
prospective epidemiological studies (Refs. 5 and 6), and the opinions 
of expert panels (Refs. 7, 8, 9, 10, 11, 12, 13, and 14). The latter 
included the 2005 recommendation of the Institute of Medicine (IOM) to 
limit trans fat consumption as much as possible while consuming a 
nutritionally adequate diet, recognizing that trans fat occurs 
naturally in meat and dairy products from ruminant animals and that 
naturally-occurring trans fat is unavoidable in ordinary, non-vegan 
diets without significant dietary adjustments that may introduce 
undesirable effects (Ref. 7). In addition, in the tentative 
determination FDA cited

[[Page 34651]]

a peer reviewed, published estimate of deaths and coronary events that 
would be prevented annually in the United States from elimination of 
remaining uses of PHOs from the food supply (Ref. 15). Given all this 
evidence, we tentatively determined that there is no longer a consensus 
among qualified experts that PHOs, the primary dietary source of IP-
TFA, are safe for human consumption, either directly or as ingredients 
in other food products.
    PHOs have a long history of use as food ingredients. The two most 
common PHOs currently used by the food industry, partially hydrogenated 
soybean oil and partially hydrogenated cottonseed oil, are not listed 
as GRAS or as approved food additives in FDA's regulations. However, 
these and other commonly used PHOs (e.g., partially hydrogenated 
coconut oil and partially hydrogenated palm oil) have been considered 
GRAS by the food industry based on a history of use prior to 1958. By 
contrast, the partially hydrogenated versions of low erucic acid 
rapeseed oil (LEAR oil; Sec.  184.1555(c)(2) (21 CFR 184.1555(c)(2)) 
and menhaden oil (Sec.  184.1472(b) (21 CFR 184.1472(b))) have been 
affirmed by regulation as GRAS for use in food. Partially hydrogenated 
LEAR oil was affirmed as GRAS for use in food (50 FR 3745 (January 28, 
1985)) through scientific procedures. Partially hydrogenated menhaden 
oil was affirmed as GRAS for use in food (54 FR 38219 (September 15, 
1989)) on the basis that the oil is chemically and biologically 
comparable to commonly used partially hydrogenated vegetable oils such 
as corn and soybean oils. FDA believes that partially hydrogenated LEAR 
and menhaden oils are not currently widely used by the food industry. 
We plan to amend these regulations in a future rulemaking.
    In the November 2013 notice, FDA requested additional data and 
scientific information related to our tentative determination and, in 
particular, requested comment on several questions (78 FR 67169 at 
67174). Interested persons were originally given until January 7, 2014, 
to comment on the notice. However, in response to several requests, we 
extended the comment period to March 8, 2014 (78 FR 79701 (December 31, 
2013)).
    We received over 6000 comments in response to the November 2013 
notice announcing our tentative determination, including over 4500 form 
letters. In addition to submissions from individuals, we received 
comments from industry and trade associations, consumer and advocacy 
groups, health professional groups, and state/local governments. Most 
comments generally supported the tentative determination or supported 
aspects of it. FDA also received numerous comments stating that 
although they agreed with FDA's efforts to further reduce trans fat in 
the food supply, they disagreed with our tentative determination 
regarding the GRAS status of PHOs. Of the comments that objected to the 
tentative determination, many disagreed with FDA's scientific analysis 
and offered alternative approaches to address trans fat in the food 
supply. Some comments addressed issues outside the scope of the 
tentative determination (such as disruptions to trade, taxation of 
foods, and requests for bans on other substances) and were not 
considered. We reviewed all comments that were submitted to the docket 
before arriving at the decision outlined in this order.
    We have arranged comments and our responses by topic throughout the 
remainder of this document. To make it easier to identify the comments 
and our responses, the word ``Comment,'' in parentheses, appears before 
the comment's description and the word ``Response,'' in parentheses, 
appears before FDA's response. Each comment is numbered to help 
distinguish between different comments. The number assigned to each 
comment is purely for organizational purposes and does not signify the 
comment's value or importance.
    The major provisions of this order are:
     PHOs are not GRAS for any use in human food.
     Any interested party may seek food additive approval for 
one or more specific uses of PHOs with data demonstrating a reasonable 
certainty of no harm of the proposed use(s).
     For the purposes of this declaratory order, FDA is 
defining PHOs as those fats and oils that have been hydrogenated, but 
not to complete or near complete saturation, and with an iodine value 
(IV) greater than 4.
     FDA is establishing a compliance date of June 18, 2018.

II. Definitions and Scope, and Related Comments With FDA Responses

    (Comment 1) Some comments requested that we define PHOs and clearly 
delineate them from FHOs. The comments suggested various parameters for 
defining these fats and oils, including setting a specification for 
trans fat content (e.g., a percentage) or using iodine value (IV; also 
interchangeably called iodine number).
    (Response) FDA agrees with the comments that we should define PHOs 
to differentiate them from FHOs, which are outside the scope of this 
order. When a fat or oil is hydrogenated, the degree of hydrogenation 
can be tailored to obtain the desired properties for the application. 
FHOs are produced by allowing the hydrogenation process to proceed to 
complete or near complete saturation to obtain a more solid fat. In 
practice, the reaction does not proceed to 100 percent completion, even 
when producing FHOs, and some degree of unsaturation unavoidably 
remains in the final fat or oil. Non-hydrogenated refined fats and oils 
generally contain trans fatty acids as an unavoidable impurity as a 
result of high-temperature processing, at levels typically below 2 
percent (Ref. 2). The IV of a fat or oil is not a direct measure of the 
TFA content, but is a measure of the degree of unsaturation. Thus, in a 
fat or oil that has been hydrogenated, a low degree of unsaturation 
(i.e., a low IV number) will correlate to a low level of TFA. FHOs with 
an IV of 4 or less generally contain trans fat at levels similar to 
non-hydrogenated refined fats and oils (less than 2 percent). By 
contrast, when the hydrogenation process is arrested before near 
complete saturation, trans fat content is typically higher, and IV is 
typically greater than 4.
    Based on data for FHOs that are currently available on the market, 
which are indicative of modern hydrogenation technology (Ref. 16), we 
define FHOs for the purposes of this order as fats and oils that have 
been hydrogenated to complete or near complete saturation, and with an 
IV of 4 or less, as determined by a method that is suitable for this 
analysis (e.g., ISO 3961 or equivalent). FHOs are outside the scope of 
this order. For the purposes of this order, we define PHOs as fats and 
oils that have been hydrogenated, but not to complete or near complete 
saturation, and with an IV greater than 4 as determined by a method 
that is suitable for this analysis (e.g., ISO 3961 or equivalent). 
These definitions will ensure that IP-TFA content in the food supply 
will be kept to the minimum amount feasible with current technology, 
except as otherwise authorized.
    (Comment 2) We received several comments requesting clarification 
on the scope of FDA's tentative determination, including whether it 
applies only to PHOs used in human food; whether it applies to 
ingredients that contain only naturally occurring trans fat, such as 
those ingredients derived from ruminant sources; and whether it applies 
to conjugated linoleic acid. We also received a citizen petition 
(discussed in section V) raising questions related to partially 
hydrogenated methyl ester of rosin.

[[Page 34652]]

    (Response) FDA wishes to clarify that this order applies only to 
PHOs used in human food, not animal feed, and applies to PHOs used as a 
food ingredient, which includes those uses sometimes considered 
processing aids or food contact substances (e.g., pan-release agents). 
By contrast, the use of PHOs as raw materials used to synthesize other 
ingredients is outside the scope of this order. We do not have specific 
information on the intake of industrially-produced trans fat from this 
source. There is no requirement that materials used to make food 
ingredients be GRAS themselves; rather, the resultant food ingredient 
must be safe for the intended conditions of use. The use of PHOs as raw 
materials to make other food ingredients may result in the 
incorporation of industrially-produced trans fats into those 
ingredients. When ingredients are synthesized using PHOs, and the 
ingredient is being used on the basis of a GRAS self-determination, 
reevaluation of such a determination may be appropriate in light of the 
health effects from the intake of trans fat that underlie our 
determination that PHOs do not meet the GRAS standard.
    This order does not apply to ingredients that contain only 
naturally occurring trans fat, such as those ingredients derived from 
ruminant sources.
    This order does not apply to the use of conjugated linoleic acid 
(CLA) as a food ingredient. CLA does not fit the definition of PHO. 
CLAs are a class of fatty acid isomers derived from linoleic acid and 
do not contain nonconjugated double bonds in a trans configuration nor 
are CLAs triglyceride molecules. On the other hand, PHOs are primarily 
mixtures of triglycerides, produced by partial hydrogenation and 
include at least one nonconjugated double bond(s) in a trans 
configuration (Ref. 16). Considering CLA to be distinct from PHOs is 
consistent with how FDA has previously defined trans fatty acids for 
nutrition labeling purposes, focusing on the presence of nonconjugated 
bond(s) in a trans configuration (see Sec.  101.9(c)(2)(ii) (21 CFR 
101.9(c)(2)(ii))).
    This order also does not apply to the use of partially hydrogenated 
methyl ester of rosin. Partially hydrogenated methyl ester of rosin 
does not fit the definition of PHO. Partially hydrogenated methyl ester 
of rosin is composed of resin acids that are chemically and 
structurally distinct from fatty acids found in PHOs. Resin acids are 
terpene-derived aromatic compounds that do not have long chain fatty 
acid components with cis/trans double bonds (Ref. 16).

III. Discussion of Legal Issues, and Related Comments With FDA 
Responses

A. GRAS

    Section 409 of the FD&C Act provides that a food additive is unsafe 
unless it is used in accordance with conditions set forth in that 
section. ``Food additive'' is defined by section 201(s) of the FD&C Act 
(21 U.S.C. 321(s)) as any substance the intended use of which results 
or may reasonably be expected to result in its becoming a component or 
otherwise affecting the characteristics of any food, if such substance 
is not GRAS or otherwise excluded from the definition. Certain other 
substances that may become components of food are also excluded from 
the statutory definition of food additive, including pesticide 
chemicals and their residues, new animal drugs, color additives, and 
dietary ingredients in dietary supplements (section 201(s)(1) through 
(6) of the FD&C Act).
    A substance is GRAS if it is generally recognized, among experts 
qualified by scientific training and experience to evaluate its safety, 
as having been adequately shown through scientific procedures (or, in 
the case of a substance used in food prior to January 1, 1958, through 
either scientific procedures or experience based on common use in food) 
to be safe under the conditions of its intended use (section 201(s) of 
the FD&C Act). However, history of use prior to 1958 is not sufficient 
to support continued GRAS status if new evidence demonstrates that 
there is no longer a consensus that an ingredient is safe. See Sec.  
170.30(l) (21 CFR 170.30(l)) (``New information may at any time require 
reconsideration of the GRAS status of a food ingredient.'').
    FDA has defined safe as ``a reasonable certainty in the minds of 
competent scientists that the substance is not harmful under the 
intended conditions of use'' (Sec.  170.3(i) (21 CFR 170.3(i)), and 
general recognition of safety must be based only on the views of 
qualified experts (21 CFR 170.30(a)). To establish general recognition 
of safety, there must be a consensus of expert opinion regarding the 
safety of the use of the substance. See, e.g., United States v. Western 
Serum Co., Inc., 666 F.2d 335, 338 (9th Cir. 1982) (citing Weinberger 
v. Hynson, Westcott & Dunning, 412 U.S. 609, 629-32 (1973)). General 
recognition of safety does not require unanimous agreement. See, e.g., 
United States v. Articles of Drug * * * 5,906 boxes, 745 F.2d 105, 119 
n. 22 (1st Cir. 1984); United States v. Articles of Food and Drug 
(Coli-Trol 80), 518 F.2d 743, 746 (5th Cir. 1975) (``What is required 
is not unanimous recognition but general recognition.''); United States 
v. Articles of Drug * * * Promise Toothpaste, 624 F. Supp. 776, at 782-
3 (N.D. Ill. 1985) (``There is nothing in the statute to indicate that 
Congress intended `generally recognized' in other than its commonly 
understood meaning. The adverb, `generally,' is defined, inter alia, to 
mean . . . extensively, though not universally'' (internal quotations 
omitted)). Conversely, general recognition of safety does not exist if 
there is a lack of consensus among qualified experts that the use of a 
substance is safe. See, e.g., Coli-Trol 80, 518 F.2d at 746 (no general 
recognition of safety where there was ``no recognition of the safety . 
. . of these products at all''); Premo Pharmaceutical Laboratories v. 
United States, 629 F.2d 795, 803-4 (2nd Cir. 1980) (``genuine dispute 
among qualified experts'' precludes finding of general recognition, and 
no general recognition existed as a matter of law where there was a 
``sharp difference'' of expert opinion); United States v. Article of 
Food * * * Coco Rico, 752 F.2d 11, 15 n 6 (1st Cir. 1985) (substance 
was not GRAS as a matter of law based on existence of ``genuine dispute 
among qualified experts'' regarding safety of use); Promise Toothpaste, 
624 F. Supp. at 783 (court could not conclude whether a ``genuine 
dispute'' existed without considering the substance of the experts' 
opinions, such that a triable issue of fact existed regarding general 
recognition). See also United States v. Articles of Drug * * * 5,906 
Boxes, 745 F.2d 105, 119 n. 22 (1st Cir. 1984) (noting certain cases in 
which lack of general recognition was established as a matter of law 
and others in which there was a triable issue of fact regarding general 
recognition).
    Importantly, the GRAS status of a specific use of a particular 
substance in food may change as knowledge changes. For example, 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 the use 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 Sec.  170.38, which 
provides, in part, that we may, on our own initiative, propose to 
determine that a substance is not GRAS. (See generally 37 FR 6207 
(March 25, 1972) (proposal of 21 CFR 121.41, the

[[Page 34653]]

predecessor of Sec.  170.38); 37 FR 25705 (December 2, 1972) (issuance 
of 21 CFR 121.41); 35 FR 18623 (December 8, 1970) (proposal of 21 CFR 
121.3, the predecessor of Sec.  170.30); and 36 FR 12093 (June 25, 
1971) (issuance of 21 CFR 121.3)). Further, as stated in section I, 
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 (Sec.  170.30(l)).
    As noted in section III.A, under section 201(s) of the FD&C Act, a 
substance that is GRAS for a particular use in food is not a food 
additive, and may lawfully be utilized for that use without FDA review 
or approval. Currently, a GRAS determination may be made when the 
manufacturer or user of a food substance evaluates the safety of the 
substance and the views of qualified experts and determines that the 
use of the substance is GRAS. This approach is commonly referred to as 
``GRAS self-determination'' or ``independent GRAS determination.''
    Other substances that are GRAS may be identified in FDA regulations 
in one of two ways. Following the passage of the 1958 Food Additives 
Amendment, we established in our regulations a list of food substances 
that, when used as indicated, are considered GRAS. We made clear that 
this was not a comprehensive list. This list (commonly referred to as 
the ``GRAS list'') now appears at 21 CFR part 182. Thereafter, in 1972, 
we established the GRAS affirmation process through which we affirmed, 
through notice and comment rulemaking, the GRAS status of particular 
uses of certain substances in food. Regulations affirming the GRAS 
status of certain substances appear at 21 CFR parts 184 and 186. (As a 
general matter, we no longer affirm the GRAS status of substances 
through notice-and-comment rulemaking. In April 1997, we proposed to 
replace the voluntary GRAS affirmation petition process with a 
voluntary GRAS notification program, which would not involve rulemaking 
(62 FR 18938 (April 17, 1997)). At the time of the proposal, we 
initiated a pilot of the GRAS notification program, which continues to 
function. A firm may voluntarily submit information on a GRAS self-
determination to FDA for review through the GRAS notification program, 
but is not required to do so.)
    FDA received numerous comments on our tentative determination. Many 
related to the GRAS standard and what is needed to demonstrate that a 
substance is not GRAS. Many comments agreed with our determination that 
there is not a consensus among qualified experts that PHOs are safe for 
use in human food. However, there were also many comments that 
disagreed with FDA's tentative determination and stated that we did not 
adequately demonstrate that PHOs are not GRAS.
    (Comment 3) Some comments stated that FDA must show a ``severe 
conflict'' among experts about the safety of a substance in order to 
determine that PHOs are not GRAS.
    (Response) FDA disagrees that ``severe conflict'' is the relevant 
standard. As discussed in section III.A, general recognition of safety 
does not exist if there is a lack of consensus among qualified experts 
that the use of a substance is safe. We have considered all available 
information and determined that there is no longer a consensus among 
qualified experts that PHOs are safe for human consumption. To the 
extent there is disagreement among qualified experts about the safety 
of PHOs for human consumption, this genuine dispute regarding safety 
precludes a finding of GRAS.
    (Comment 4) Some comments focused on the idea that it may be 
possible to establish a threshold below which PHOs may be safely used 
in the food supply. One comment argued that there is no consensus among 
experts that PHOs are unsafe below some low threshold level of use.
    (Response) As discussed later in section IV.B.1, FDA does not agree 
that such a threshold has been identified based on the available 
science. Importantly, even if such a threshold could be identified, 
this alone would not meet the requirement of ``general recognition'' 
for uses below the threshold without there also being consensus among 
qualified experts that uses below the threshold are safe. (See United 
States v. 7 Cartons, 293 F. Supp. 660, 663 (S.D. Ill. 1968) (``an 
inference that safety might be shown by scientific testing and 
procedures'' is insufficient as a matter of law to demonstrate general 
recognition of safety), affirmed in relevant part, 424 F.2d 1364 (7th 
Cir. 1970).) FDA has no basis to conclude that there is any such 
consensus. FDA has previously revoked GRAS status under similar 
circumstances (51 FR 25021 at 25023, July 9, 1986; revoking GRAS status 
of sulfiting agents on fruits and vegetables intended to be served or 
sold raw to consumers; explaining that it was not possible to set a 
threshold for safe use based on available information). Moreover, we 
need not determine that there is a consensus that low level uses are 
unsafe to find that PHOs are not GRAS at low levels; we need only 
determine that based on available scientific evidence there is not a 
consensus among qualified experts that such uses are safe, as we do 
here. We acknowledge that scientific knowledge advances and evolves 
over time. We encourage submission of scientific evidence as part of 
food additive petitions under section 409 of the FD&C Act for one or 
more specific uses of PHOs for which industry or other interested 
individuals believe that safe conditions of use may be prescribed. We 
are establishing a compliance date of June 18, 2018 for this order to 
allow time for such petitions and their review.
    (Comment 5) One comment stated that FDA must demonstrate that each 
and every PHO, and every use of PHOs, is not safe.
    (Response) FDA disagrees. FDA need not demonstrate that PHOs are 
unsafe to determine that they are not GRAS, only that there is a lack 
of consensus among qualified experts regarding their safety. In 
addition, our consideration of PHOs as a class is justified because the 
available, relevant scientific evidence demonstrates an increased risk 
of coronary heart disease (CHD) attributable to trans fat (see section 
VI.B); PHOs are the primary dietary source of IP-TFA; and there is a 
lack of consensus among qualified experts that PHOs are safe for use in 
food at any level.
    (Comment 6) Some comments stated that, by determining that the use 
of PHOs are not GRAS because they contain a nutrient that increases 
risk of CHD, FDA would be calling into question the regulatory status 
of other food sources of trans fat.
    (Response) FDA disagrees. As noted in section II, this order does 
not apply to ingredients that contain naturally occurring trans fat 
(such as those ingredients derived from ruminant sources), fully 
hydrogenated oils, or edible oils that contain IP-TFA as an impurity. 
FDA has considered the available information and concluded that there 
is a lack of consensus among qualified experts that PHOs, as the 
primary dietary source of IP-TFA, are safe for use in human food. We 
may determine that the use of an artificial substance is not GRAS 
without necessarily making the same determination about naturally-
occurring versions of the substance. (See, e.g., 35 FR 7414 (May 13, 
1970) (Rescinding letters that had expressed opinions that certain uses 
of glycine and its salts are GRAS, and stating that such added 
substances are no longer GRAS in human food); 37 FR 6938 (April 6, 
1972) (Amino Acids in Food for Human Consumption; Proposed Conditions 
of Safe Use in Food and Deletion From GRAS List) (``[T]he mere natural

[[Page 34654]]

presence of an amino acid in unprocessed foods in free or combined (as 
protein) form does not qualify it as safe for addition in a pure form 
as a component of a formulated or processed food''), 38 FR 20036 (July 
26, 1973) (Amino Acids in Food for Human Consumption; Conditions of 
Safe Use in Food and Deletion From GRAS List); 47 FR 22545 (May 25, 
1982) (Cinnamyl Anthranilate; Proposed Prohibition of Use in Human 
Food) (acknowledging ``the presence of other cinnamyl and anthranilate 
derivatives naturally in food and in natural substances used to flavor 
food'' but proposing to prohibit only cinnamyl anthranilate); 50 FR 
42929 (October 23, 1985) (Cinnamyl Anthranilate; Prohibition of Use in 
Human Food)).
    (Comment 7) One comment stated that Congress, through the Nutrition 
Labeling and Education Act of 1990 (NLEA) (Pub. L. 101-535), prescribed 
labeling as the sole vehicle for achieving the nutritional policy 
objective of shifting dietary patterns to reduce the risk of 
multifactorial chronic diseases such as CHD. The comment argued that 
FDA's use of its food additive authority with respect to PHOs and their 
effect on risk of CHD is not within FDA's legal authority. Some 
comments characterized the tentative determination as a new approach or 
a change in interpretation, arguing that FDA has not previously 
addressed health concerns related to nutrient intake through the FD&C 
Act's food additive provisions. In support of the argument that FDA has 
changed its interpretation of the applicability of the food additive 
provisions of the FD&C Act, one comment cited a statement by FDA in 
rulemaking regarding health claims that ``where the only safety issue 
is an increased risk of chronic disease from excessive consumption, the 
safety provisions of the act would not provide regulatory sanctions 
against such components of food, at least if they have not been added 
to foods'' (58 FR 2478 at 2490 (January 6, 1993)).
    (Response) FDA disagrees with these comments. FDA may properly 
address such health risks using the food additive authorities in the 
FD&C Act (sections 201(s), 409, and 402(a)(2)(C) of the FD&C Act). The 
broad language of the food additive definition in section 201(s) of the 
FD&C Act covers ``any substance'' added to food, including nutrients. 
Nothing in the FD&C Act or its legislative history suggests that the 
food additive definition should be interpreted in a way that limits its 
applicability as the comment suggests. On the contrary, the legislative 
history of the Food Additives Amendment of 1958 (Pub. L. 85-929) 
emphasizes the broad applicability of sections 201(s), 409, and 
402(a)(2)(C) of the FD&C Act, which apply to ``any substances the 
ingestion of which reasonable people would expect to produce not just 
cancer but any disease or disability'' (S. Rep. No. 2422, at 11 (1958), 
as reprinted in Vol. 14, Legislative History of the Food, Drug & 
Cosmetic Act and its Amendments, at 923 (1979)). In fact, we have 
previously taken action regarding health risks related to nutrients 
using these authorities (55 FR 50777 (December 10, 1990) (determining 
certain Vitamin K Active Substances not GRAS); and 38 FR 20036 (July 
26, 1973) (establishing conditions of safe use for amino acids for 
nutritive purposes and deleting them from GRAS list)). We also have 
previously applied these authorities to substances presenting increased 
health risks related to chronic multifactorial diseases, such as cancer 
(50 FR 42929 (October 23, 1985) (prohibiting use of cinnamyl 
anthranilate in food); and 34 FR 17063 (October 21, 1969) (prohibiting 
use of cyclamates in food)).
    With respect to the comment citing a statement from a final rule on 
health claims, FDA does not agree that this statement shows any change 
in FDA's position, as it was explicitly limited to situations that did 
not meet the food additive definition because the components discussed 
``have not been added to foods.'' The statement is consistent with 
FDA's current understanding of the law.
    Moreover, FDA disagrees with the argument that FDA must address 
health risks related to PHOs through food labeling requirements rather 
than through the food additive provisions of the FD&C Act. The NLEA 
amended the FD&C Act to provide, among other things, for certain 
nutrients and food components to be included in nutrition labeling. 
Section 403(q)(2)(A) and (q)(2)(B) (21 U.S.C. 343(q)(2)(A) and 
(q)(2)(B)) of the FD&C Act state that the Secretary of Health and Human 
Services (the Secretary) (and, by delegation, FDA) can, by regulation, 
add or delete nutrients included in the food label or labeling if he or 
she finds such action necessary to assist consumers in maintaining 
healthy dietary practices. We have used this authority to require 
labeling of trans fat content (68 FR 41434 (July 11, 2003); see also 
Sec.  101.9(c)(2)(ii) and Sec.  101.36(b)(2)(i)) (21 CFR 
101.36(b)(2)(i)). Although we may further address trans fat through 
labeling requirements in the future, labeling is not the only method by 
which we may address health risks related to trans fats, and more 
specifically health risks related to PHOs, the primary dietary source 
of IP-TFA. Nothing in the NLEA suggested that its passage limited the 
preexisting food additive provisions in the FD&C Act, or that the food 
additive provisions did not apply to nutrients and chronic 
multifactorial disease under appropriate circumstances. On the 
contrary, as the comment noted, the NLEA contained a clause stating 
that ``[t]he amendments made by this Act shall not be construed to 
alter the authority of the Secretary of Health and Human Services . . . 
under the [FD&C Act]'' (NLEA section 9).
    The FD&C Act's nutrition labeling and food additive provisions are 
two different kinds of authority, with different standards, and we may 
choose among available approaches to a public health problem when the 
FD&C Act provides multiple options. See, e.g., Chevron U.S.A. Inc. v. 
Natural Resources Defense Council, 467 U.S. 837, 865-6 (1984) (``While 
agencies are not directly accountable to the people, the Chief 
Executive is, and it is entirely appropriate for this political branch 
of the Government to make such policy choices--resolving the competing 
interests which Congress itself either inadvertently did not resolve, 
or intentionally left to be resolved by the agency charged with the 
administration of the statute in light of everyday realities''); United 
States v. Mead Corp., 533 U.S. 218, 227 (2001) (``agencies charged with 
applying a statute necessarily make all sorts of interpretive 
choices''). There is no ``conflict'' between the FD&C Act's nutrition 
labeling provisions and food additive provisions as the comment 
suggests. It is also worth noting that we have previously determined 
that a use of a substance is not GRAS while rejecting a labeling-based 
approach to the health risks presented by that use (51 FR 25021 (July 
9, 1986) (final rule revoking GRAS status of sulfiting agents on fruits 
and vegetables intended to be served or sold raw to consumers); and 50 
FR 32830 (August 14, 1985) (proposal to revoke GRAS status of sulfiting 
agents on fruits and vegetables intended to be served or sold raw to 
consumers)).
    (Comment 8) Some comments stated that the expert panels we cited in 
the tentative determination (i.e., the Institute of Medicine/National 
Academy of Sciences (IOM/NAS), American Heart Association, American 
Dietetic Association, World Health Organization, Dietary Guidelines 
Advisory Committee, and the FDA Food Advisory Committee Nutrition 
Subcommittee) were not experts qualified by scientific training and 
experience to evaluate the safety of substances in food. The comments 
also

[[Page 34655]]

stated that these expert panels were not convened for the purposes of 
evaluating the safety of PHOs and did not make determinations regarding 
the GRAS status of PHOs. Therefore, the comments argued that the 
conclusions of these panels do not demonstrate a lack of consensus 
among qualified experts that PHOs are GRAS.
    (Response) FDA disagrees with these comments. The expert panels we 
cited were composed of scientists qualified by relevant training and 
experience to review literature on trans fat consumption, because of 
their nationally recognized and established expertise in the area of 
food and nutrition. For example, the Food and Nutrition Board at IOM/
NAS is a recognized national resource for recommendations on health 
issues, and the Dietary Guidelines Advisory Committee members are 
nationally recognized experts in nutrition and health. These panels' 
evaluations and conclusions raised significant questions about the 
safety of trans fat, thus showing that there is no consensus among 
qualified scientific experts that PHOs are safe, because PHOs are the 
primary dietary source of IP-TFA. The safety information reviewed by 
the panels is further discussed in section IV.B.2. We consider that the 
conclusions of the panels demonstrate that there is a ``lack of the 
proper reputation . . . for safety of the food additive among the 
appropriate experts.'' Coli-Trol 80, 518 F.2d at 746. Further, whether 
the panels were convened specifically to make a GRAS determination is 
irrelevant; the purpose of the panels was to review the available data 
on health risks associated with consumption of trans fat. Moreover, the 
expert panel conclusions are not the only evidence upon which we rely 
for this determination, and conclusions of an expert panel are not 
required to establish general recognition of safety or its absence.
    (Comment 9) Several comments stated that the expert panels we cited 
considered nutritional science and not safety.
    (Response) FDA disagrees that the panels were not considering 
safety data; panels were considering data from controlled trials and 
observational studies on trans fat consumption that showed adverse 
effects on risk factors (e.g., effects on cholesterol) and increased 
risk of CHD (see section IV.B.2 for further discussion on expert panel 
reviews). As discussed in more detail in section III.A, FDA regulations 
define ``safe'' as ``a reasonable certainty in the minds of competent 
scientists that the substance is not harmful under the intended 
conditions of use'' (Sec.  170.3(i)), and data showing a potential 
relationship between a nutrient (or any other substance added to food) 
and disease are safety data. Studies reviewed by expert panels showed 
that trans fatty acids cause significant health risks. Such studies are 
safety data.
    (Comment 10) One comment stated that FDA should hold the 
manufacturer initially introducing the food or ingredient into 
interstate commerce responsible for compliance with a determination 
that PHOs are not GRAS, and that distributors should not be responsible 
for determining whether foods they merely distribute contain PHOs.
    (Response) Although we are mindful of the need to focus our 
enforcement efforts, those needs do not change the underlying law or 
FDA's legal authority. Food that is adulterated may be subject to 
seizure and distributors, manufacturers, and other parties responsible 
for such food may be subject to injunction. We recognize that 
manufacturers who have previously added PHO to food, rather than other 
parties such as distributors who merely receive and sell finished 
foods, are the members of the food industry who will be most directly 
affected by this order, and we intend to focus our outreach and 
enforcement resources accordingly. However, we remind distributors and 
other members of the food industry that they have an obligation to 
ensure that the food they manufacture, distribute, sell, or otherwise 
market complies with the FD&C Act.
    (Comment 11) Some comments requested that FDA take a position 
regarding the effect of this order on state and local laws regarding 
PHOs.
    (Response) There is no statutory provision in the FD&C Act 
providing for express preemption of any state or local law prohibiting 
or limiting use of PHOs in food, including state or local legislative 
requirements or common law duties. As with any Federal requirement, if 
a State or local law requirement makes compliance with both Federal law 
and State or local law impossible, or would frustrate Federal 
objectives, the State or local requirement would be preempted. See 
Wyeth v. Levine, 555 U.S. 555 (2009); Geier v. American Honda Co., 529 
U.S. 861 (2000); English v. General Electric Co., 496 U.S. 72, 79 
(1990), Florida Lime & Avocado Growers, Inc., 373 U.S. 132, 142-143 
(1963); Hines v. Davidowitz, 312 U.S. 52, 67 (1941). We decline to take 
a position regarding the potential for implied preemptive effect of 
this order on any specific state or local law; as such matters must be 
analyzed with respect to the specific relationship between the state or 
local law and the federal law. FDA believes, however, that state or 
local laws that prohibit or limit use of PHOs in food are not likely to 
be in conflict with federal law, or to frustrate federal objectives.

B. Prior Sanctions

    We stated in our tentative determination that we were not aware 
that FDA or U.S. Department of Agriculture (USDA) had granted any 
explicit approval for any use of PHOs in food prior to the 1958 Food 
Additives Amendment to the FD&C Act, and requested comments on whether 
there was knowledge of an applicable prior sanction for the use of PHOs 
in food (78 FR 67169 at 67174). We received various comments on this 
topic. We are not making a determination regarding the existence of any 
prior sanctions for uses of PHO in this order. This order is limited to 
our determination regarding the GRAS status of PHOs. We intend to 
address any claims of prior sanction in a future action.

C. Procedural Requirements

    Under 5 U.S.C. 554(e) (section 5(d) of the Administrative Procedure 
Act (APA)), an agency, ``in its sound discretion, may issue a 
declaratory order to terminate a controversy or remove uncertainty.'' 
The APA defines ``order'' as ``the whole or a part of a final 
disposition, whether affirmative, negative, injunctive, or declaratory 
in form, of an agency in a matter other than rulemaking but including 
licensing'' (5 U.S.C. 551(6)). The APA defines ``adjudication'' as 
``agency process for the formulation of an order'' (5 U.S.C. 551(7)).
    FDA's regulations, consistent with the APA, define ``order'' to 
mean ``the final agency disposition, other than the issuance of a 
regulation, in a proceeding concerning any matter . . .'' (Sec.  
10.3(a) (21 CFR 10.3(a)). Our regulations also define ``proceeding and 
administrative proceeding'' to mean ``any undertaking to issue, amend, 
or revoke a regulation or order, or to take or not to take any other 
form of administrative action, under the laws administered by the Food 
and Drug Administration'' (Sec.  10.3(a)). Moreover, our regulations 
establish that the Commissioner may initiate an administrative 
proceeding to issue, amend, or revoke an order (21 CFR 10.25(b)).
    FDA's regulations also set forth a process by which we, on our own 
initiative or on the petition of an interested person, may determine 
that a substance is not GRAS. Specifically, FDA may initiate this 
process by issuing

[[Page 34656]]

a notice in the Federal Register proposing to determine that a 
substance is not GRAS and is a food additive subject to section 409 of 
the FD&C Act (Sec.  170.38(b)). The notice must allow a period of 60 
days for comment. If, after review of comments, FDA determines that 
there is a lack of convincing evidence that a substance is GRAS or is 
otherwise exempt from the definition of a food additive in section 
201(s) of the FD&C Act, FDA will publish a notice thereof in the 
Federal Register (Sec.  170.38(b)(3)). Such a notice ``shall provide 
for the use of the additive in food or food contact surfaces as 
follows: (1) It may promulgate a food additive regulation governing use 
of the additive[;] (2) It may promulgate an interim food additive 
regulation governing use of the additive[;] (3) It may require 
discontinuation of the use of the additive[;] (4) It may adopt any 
combination of the above three approaches for different uses or levels 
of use of the additive'' (Sec.  170.38(c)).
    On our own initiative, we began an administrative proceeding to 
formulate a 5 U.S.C. 554(e) declaratory order to remove uncertainty 
regarding the GRAS status of PHOs. Accordingly, we published a notice 
in the Federal Register, consistent with Sec.  170.38(b), communicating 
our tentative determination that PHOs are no longer GRAS for any use in 
food, and allowed 60 days for comments (78 FR 67169 (November 8, 
2013)). We later extended the comment period for an additional 60 days 
(78 FR 79701 (December 31, 2013)).
    In the tentative determination, FDA noted that two PHOs had been 
affirmed by regulation as GRAS for use in food (78 FR 67169 at 67171; 
the partially hydrogenated versions of low erucic acid rapeseed oil 
(LEAR oil; Sec.  184.1555(c)(2)) and menhaden oil (Sec.  184.1472(b)). 
We also noted that the nature of some of the products for which there 
are standards of identity is such that PHOs historically have been used 
in their manufacture in conformance with those standards (78 FR 67169 
at 67171). However, we also noted that no food standard of identity 
requires the use of PHOs and, therefore, industry's ability to comply 
with any standard would not be prevented by a change in the regulatory 
status of PHOs. As discussed in section III.B, two standards of 
identity explicitly mention PHOs in allowing partially hydrogenated 
vegetable oil as an optional ingredient; the standards of identity for 
peanut butter (Sec.  164.150 (21 CFR 164.150)) and canned tuna (Sec.  
161.190 (21 CFR 161.190)). Because these standards do not require the 
use of PHOs, industry's ability to comply with them would not be 
prevented by a change in the regulatory status of PHOs. In addition, 
our labeling regulations explicitly address ingredient designations for 
PHOs (Sec.  101.4(b)(14) (21 CFR 101.4(b)(14))).
    This final determination is a 5 U.S.C. 554(e) declaratory order 
regarding the status of PHOs. Consistent with Sec.  170.38(b)(3), we 
have reviewed the comments received and determined that there is a lack 
of convincing evidence that PHOs are GRAS. Thus, consistent with Sec.  
170.38(c)(3), we are publishing a notice thereof in the Federal 
Register that requires discontinuation of the use of these additives. 
Moreover, we are providing advance notice of our intention to undertake 
rulemaking with respect to the uses of PHOs explicitly permitted for 
use by regulation and other conforming changes.
    (Comment 12) Some comments argued that FDA must determine the GRAS 
status of PHOs through notice-and-comment rulemaking.
    (Response) FDA agrees that we must conduct rulemaking to revise 
Sec. Sec.  184.1555(c)(2) and 184.1472(b), which explicitly permit the 
use of partially hydrogenated LEAR oil and partially hydrogenated 
menhaden oil, respectively. FDA will also consider taking further 
action to revise regulations regarding the standards of identity for 
peanut butter (Sec.  164.150(c)) and canned tuna (Sec.  
161.190(a)(6)(viii)), the regulation regarding ingredient designations 
for PHOs (Sec.  101.4(b)(14)), and nutrition labeling regulations 
regarding trans fats (Sec. Sec.  101.9(c)(2)(ii) and 101.36(b)(2)(i)). 
We note that although trans fat does occur naturally in some product 
groups such as dairy foods, it is only likely to be present at levels 
at or above 0.5 g per serving in products containing PHOs.
    We do not agree that we must determine the GRAS status of PHOs 
generally via rulemaking. FDA may properly make such a determination in 
an order, as we have chosen to do here. This is not the first time FDA 
has issued a declaratory order when determining that a substance is not 
GRAS and is a food additive. See 55 FR 50777, 50778 (Declaratory Order 
regarding Vitamin K Active Substances in Animal Food, issued under 21 
CFR 570.38, the regulation for animal food that parallels Sec.  170.38 
for human food).
    We have authority to administer the statutory provisions of the 
FD&C Act that are most relevant to this determination, namely, are 
sections 201(s), 402(a)(2)(C), and 409 of the FD&C Act. Section 201(s) 
of the FD&C Act defines a food additive, in part, as a substance that 
is not GRAS, and section 402(a)(2)(C) of the FD&C Act establishes that 
food bearing or containing a food additive that is unsafe within the 
meaning of section 409 of the FD&C Act is adulterated. Section 409 of 
the FD&C Act establishes that a food additive is unsafe for the 
purposes of section 402(a)(2)(C) of the FD&C Act (and therefore 
adulterated) unless certain criteria are met, such as conformance with 
a regulation prescribing the conditions under which the additive may be 
safely used. Section 409 of the FD&C Act also sets forth a process by 
which we administer the review of food additive petitions and may 
establish regulations prescribing conditions of safe use for such 
additives. Thus, we have explicit statutory authority to review, 
approve, and deny food additive petitions.
    Because it is necessary to determine whether the use of a substance 
is GRAS as part of identifying it as a food additive, it is implicit in 
this statutory structure that we also have the authority to determine 
whether the use of a substance is, or is not, GRAS. The statute does 
not explicitly provide the procedure we must use to make such 
determinations. Thus, we may choose to use either rulemaking or 
adjudication. ``The choice between rule-making or declaratory order is 
primarily one for the agency regardless of whether the decision may 
affect policy and have general prospective application.'' (See Viacom 
v. FCC, 672 F.2d 1034, 1042 (2nd Cir. 1982). See also SEC v. Chenery, 
332 U.S. 194, 203 (1947); NLRB v. Wyman-Gordon Co., 394 U.S. 759 
(1969); NLRB v. Bell Aerospace Co., 416 U.S. 267, 294 (1974); Almy v. 
Sebelius, 679 F.3d 297, 303 (4th Cir. 2012); City of Arlington, Texas 
v. FCC, 133 S. Ct. 1863, 1874 (2013); Qwest Servs. Corp. v. FCC, 509 
F.3d 531, 536-37 (D.C. Cir. 2007) (``Most norms that emerge from a 
rulemaking are equally capable of emerging (legitimately) from an 
adjudication, and accordingly agencies have very broad discretion 
whether to proceed by way of adjudication or rulemaking'' (internal 
citations and quotations omitted)).
    Determining that PHOs are no longer GRAS for use in human food in a 
declaratory order issued as a product of informal adjudication is well 
within FDA's discretion under the FD&C Act and the APA. Whether PHOs 
are GRAS for use in human food is a ``concrete and narrow question[] of 
law the resolution[] of which would have an immediate and determinable 
impact on specific factual scenarios'' (City of Arlington v. FCC, 668 
F.3d 229, 243 (5th Cir. 2012)). (See also Qwest Servs. Corp.,

[[Page 34657]]

509 F.3d at 536-37; Chisholm v. FCC, 538 F.2d 349, 364-66 (D.C. Cir. 
1976); American Bar Association, A Guide to Federal Agency Adjudication 
8 (Jeffrey B. Litwak, ed., 2012) (Agency order to withdraw certain food 
from the market, which has particular applicability and future effect, 
provided as an example of adjudication)). We are issuing this 
declaratory order to remove uncertainty as to the status of PHOs as 
food additives. The order is a product of an informal adjudication that 
included notice to affected parties via publication of the tentative 
determination in the Federal Register and an opportunity for affected 
parties to be heard by submitting comments to the Agency. Such 
procedures are appropriate for the formulation of declaratory orders. 
(See, e.g., Weinberger v. Hynson, Westcott and Dunning Inc., 412 U.S. 
609, 626 (1973); American Airlines v. Dep't. of Transportation, 202 
F.3d 788, 796-797 (5th Cir. 2000). See also Lubbers, Jeffrey S. and 
Blake D. Morant, A Reexamination of Federal Agency Use of Declaratory 
Orders, 56 Admin. L. Rev. 1097, 1112-1114 (2004) and cases cited 
therein). Moreover, ``adjudicatory decisions are not subject to the 
APA's notice-and-comment requirements'' (Blanca Telephone Co. v. FCC, 
743 F.3d 860 (D.C. Cir. 2014)).
    Issuance of a declaratory order is also consistent with our 
regulations (Sec.  170.38(c)(3)), which provide that we may publish a 
notice in the Federal Register that requires discontinuation of the use 
of these additives, and do not specify that we must do so through 
rulemaking. Notably, other subsections of Sec.  170.38(c) mention 
promulgation of regulations, but Sec.  170.38(c)(3), providing for 
prohibition of use, does not. Moreover, when we make a determination 
under Sec.  170.38 that a substance is not GRAS, we must take one (or a 
combination) of the actions listed in Sec.  170.38(c). See 
Heterochemical Corp. v. FDA, 741 F. Supp. 382, 384 (E. D. N.Y. 1990).
    The purpose of a declaratory order is ``to develop predictability 
in the law by authorizing binding determinations which dispose of legal 
controversies without the necessity of any party's acting at his peril 
upon his own view'' (U.S. Department of Justice, Attorney General's 
Manual on the Administrative Procedure Act (1947) at 59, reprinted in 
Federal Administrative Procedure Sourcebook (William F. Funk et al. 
ed., ABA Section of Administrative Law and Regulatory Practice 3rd ed. 
2000)). Members of industry are not, as some comments suggested, faced 
with a choice between complying with a non-binding statement of policy 
and facing enforcement action. This is not a statement of policy. This 
declaratory order has the force and effect of law.
    (Comment 13) Some comments assumed that this order was a statement 
of policy, and, on that basis, argued that this action violates Due 
Process requirements.
    (Response) As explained in our response to comment 10, that 
assumption is incorrect. Further, FDA's order and the process used in 
its formulation raise no Due Process concern.
    (Comment 14) Some comments argued that FDA did not conduct a full 
Regulatory Impact Analysis in issuing the tentative determination.
    (Response) As discussed previously in this section, this final 
determination is a declaratory order issued as the result of informal 
adjudication to remove uncertainty regarding the status of PHOs. We 
have prepared a memorandum (Ref. 17) updating our previous estimate of 
economic impact published in the November 2013 notice, using 
information available to us as well as information we received during 
the comment period. See discussion in section VII. Further, we have 
stated our intention to conduct rulemaking regarding uses of PHOs in 
our existing regulations, and such rulemakings will be subject to the 
procedural requirements pertaining to rulemaking.
    (Comment 15) One comment stated that FDA must provide a more 
detailed justification for this action than what was provided in the 
tentative determination because it is a change in FDA's position 
regarding PHOs and industry has a substantial reliance interest in the 
GRAS status of PHOs.
    (Response) In the tentative determination (78 FR 67169 at 67172) 
and in this order, FDA has explained the factual findings supporting 
this action in detail. In section IV.B, we describe how the scientific 
evidence, and consensus among qualified experts regarding the safety of 
PHOs, has changed over time. We are not changing our interpretation of 
the GRAS standard or the relevant regulations. We are determining that 
PHOs are no longer GRAS by applying the GRAS standard to current 
scientific evidence and the views of qualified experts about the safety 
of PHOs. Moreover, reliance interests are implicated whenever FDA makes 
a determination that removes a substance from the food supply that has 
been previously used in food. FDA is aware of such concerns; however, 
the statutory standard for GRAS does not allow FDA to consider the 
extent to which industry has relied on GRAS uses of a substance. We 
encourage industry to submit food additive petitions under section 409 
of the FD&C Act if industry believes that it is possible to establish, 
by regulation, safe conditions of use of PHOs. We are establishing a 
compliance date of June 18, 2018 for this order to allow time for 
submission of such petitions and their review and approval, if 
applicable requirements are met.

IV. Discussion of Scientific Issues, and Related Comments With FDA 
Responses

A. Intake Assessment

    In the November 2013 notice, we discussed dietary intake of trans 
fat from PHOs, estimated in 2010 and updated in 2012 (78 FR 67169 at 
67171). The intake assessment was done for four reasons: (1) To 
determine the impact of the 2003 labeling rule and subsequent 
reformulations; (2) to assist in our review of the citizen petitions, 
which are discussed in section V; (3) to consider strategies for 
further trans fat reduction, if warranted; and (4) to better understand 
the current uses of PHOs and identify products that still contain high 
levels of trans fat. Our determination regarding the GRAS status of 
PHOs relies on an analysis of whether PHOs meet the GRAS standard based 
on available scientific evidence; the intake assessment was not the 
basis for this determination.
    In 2012, we estimated the mean trans fat intake from the use of 
PHOs to be 1.0 grams per person per day (g/p/d; 0.5 percent of energy 
based on a 2,000 calorie diet \1\) for the U.S. population aged 2 years 
or more. We also estimated intake for high-level consumers (represented 
by intake at the 90th percentile), as well as a ``high-intake'' 
scenario that assumed consumers consistently chose products with the 
highest trans fat levels. We received a number of comments on our 
intake assessment, including comments on assumptions, methodology, and 
recommendations for future studies.
---------------------------------------------------------------------------

    \1\ (1.0 g/p/d x 9 kcal/g x 100)/2,000 kcal/d = 0.5% of energy.
---------------------------------------------------------------------------

    (Comment 16) One comment challenged FDA's statement that intake of 
trans fat did not significantly change between 2010 and 2012. The 
comment indicated that the intake of trans fat from the use of PHOs 
decreased by roughly 23% in that time period due to significant 
reformulation efforts by the food industry.
    (Response) FDA agrees that a comparison of the assessments from 
2010 and 2012 demonstrates that reformulation has occurred and intake 
has decreased. While the intake estimates did show a 23 percent

[[Page 34658]]

decrease in trans fat intake between 2010 and 2012 (1.3 g/p/d to 1.0 g/
p/d), this change is small compared to the 3.3 g/p/d difference between 
FDA's intake estimate in the 2003 trans fat labeling final rule of 4.6 
g/p/d and the 2010 estimate of 1.3 g/p/d (about a 72 percent decrease). 
This was the context for the statement in the tentative determination 
that, ``We do not consider this to be a significant change in the 
overall dietary intake of trans fat since 2010. However, it suggests a 
continued downward trend in the dietary intake of trans fat.''
    (Comment 17) Many comments stated that a substantial number of 
products have been reformulated since the 2012 intake assessment and 
that we should revise our intake assessment for trans fat before 
issuing our final determination on the GRAS status of PHOs.
    (Response) FDA agrees that reformulation efforts by industry are 
continuing. However, the 2012 intake assessment was intended to be a 
snapshot in time and was based on products containing PHOs that were in 
the market at that time, and was done for the reasons described 
previously in this section. Given the evidence FDA has reviewed and our 
determination that PHOs are not GRAS for any use in human food, an 
updated intake assessment for trans fats from PHOs is not needed at 
this time. Our determination that PHOs are not GRAS for use in human 
food does not rely on the intake assessment.
    (Comment 18) Some comments stated that FDA should not use the 
``high intake scenario'' as justification for a determination that PHOs 
are not GRAS. Related comments stated that the intake for the highest 
level consumers should be determined directly rather than using worst-
case scenario assumptions.
    (Response) FDA disagrees that the high intake assessments provide 
justification for our determination regarding the GRAS status of PHOs; 
the determination is based on our assessment of whether any use of PHOs 
in human food meets the GRAS standard, based on available scientific 
evidence. Our determination did not rely on the intake assessment.
    (Comment 19) Several comments stated that FDA's estimate did not 
calculate intake from animal products that contain trans fat, and that 
FDA should update the intake assessment to include the intake of total 
trans fat from both ruminant sources and IP-TFA. The comments noted 
this was necessary to understand if dietary recommendations are being 
met. One comment indicated that a recent publication suggests that the 
intake of trans fat from ruminant sources may be decreasing, thereby 
indicating a more inclusive review of dietary intake of trans fat is 
warranted. Another comment stated that we did not consider the 
cumulative effect of trans fat because it did not present data on 
intake from all sources, including ruminant TFA.
    (Response) Our study was designed to assess trans fat intake from 
the use of PHOs, because they are the primary source of IP-TFA, and IP-
TFA was the focus of the intake assessment. As stated in our tentative 
determination (78 FR 67169 at 67172), the IOM's recommendation is that 
trans fat consumption should be kept as low as possible while consuming 
a nutritionally adequate diet, recognizing that trans fat occurs 
naturally in meat and dairy products from ruminant animals and that 
naturally-occurring trans fat is unavoidable in ordinary, non-vegan 
diets without significant dietary adjustments that may introduce 
undesirable effects. Therefore, our intake assessment focused only on 
trans fat from the use of PHOs, the primary dietary source of IP-TFA, 
in which trans fat is produced intentionally and is an integral 
component.
    (Comment 20) One comment urged FDA to reevaluate the intake of 
trans fat using the most recent National Health and Nutrition 
Examination Survey (NHANES) data. The comment suggested that the intake 
of trans fat would be lower if the more recent NHANES data were used 
because the mandatory labeling rule for trans fat became effective on 
January 1, 2006.
    (Response) While the 2003-2006 NHANES food consumption data were 
used in the 2010 and 2012 intake assessments, the levels of trans fat 
in the food products were determined based on products that were 
available in the market from 2009 to 2012, therefore capturing trans 
fat reductions due to product reformulation as a result of the 
regulation in Sec.  101.9(c)(2)(ii) (effective in 2006) requiring 
declaration of the trans fat content of food in the nutrition label. 
The consumption of products in the food categories in which PHOs are 
used would not be expected to change significantly over a few years 
because for the most part, foods tend to be commonly consumed with 
little or no change in consumption patterns over short periods of time. 
Further, we compared the typical intake of trans fat using the 2003-
2006 and 2003-2008 NHANES food consumption data and found that there 
were no significant differences in the intakes (Ref. 16).
    (Comment 21) Several comments suggested that using a value of 0.4 g 
trans fat per serving for foods that declared 0 g trans fat on the 
label, but contained a PHO was an overestimation of intake. One comment 
stated that this assumption represents 40% of the estimated daily 
intake of 1.0 g/p/d.
    (Response) FDA disagrees with the comments. For most of the food 
products that declared 0 g trans fat on the label, but contained a PHO, 
a level based on analytical data was used. A value of 0.4 g trans fat/
serving was used for only 2 percent of all of the food codes included 
in the intake assessment (Ref. 16). The value of 0.4 g is the amount of 
trans fat estimated to be in in the food(s) that corresponds to a given 
food code that was used in the intake assessment, and does not 
represent a percentage of total estimated intake. As a result, we do 
not expect that using a lower value would significantly affect the 
overall estimated intake of trans fat from the use of PHOs. The use of 
0.4 g trans fat/serving was reserved for those cases where no other 
information was available (i.e., analytical data or an appropriate 
surrogate). Furthermore, while numerically 0.4 g is 40 percent of 1.0 
g, it is not appropriate to compare these two parameters. Many factors 
(i.e., the amount of the particular food consumed, the percent of the 
population consuming the given food, and the level of trans fat in the 
particular food) were used to derive the overall estimated trans fat 
intake.
    (Comment 22) One comment suggested that American Oil Chemists 
Society (AOCS) methods should be used for the intake assessment instead 
of the AOAC method 996.06 since the AOAC method is outdated and has not 
undergone validation.
    (Response) FDA disagrees. This AOAC method is widely used by 
industry and other international organizations as a method for 
determining the trans fat content in food products. Therefore, we 
considered the AOAC method to be appropriate for analyzing food samples 
for the purposes of our intake assessment. Our choice of the AOAC 
method is not intended to imply that industry must use this method to 
analyze food products.
    (Comment 23) Two comments indicated that a new intake assessment 
should be performed using modeling to explore potential unintended 
consequences of decreasing the trans fat intake given the possible 
replacements for trans fat (e.g., saturated fat, carbohydrate) and 
their impact on CHD risk.
    (Response) The safety of other substances that are possible 
replacements for PHOs is outside the scope of this order. However, 
although we have not updated the intake

[[Page 34659]]

assessment since 2012, we have used this intake assessment to calculate 
the expected impact of this order on CHD events, taking into account 
possible replacements for PHOs (see section IV.B for detailed 
discussion).
    (Comment 24) One comment noted that FDA did not examine the use of 
each PHO and the probable consumption of each use.
    (Response) FDA disagrees that we need to examine the intake of each 
PHO individually; the intent of the intake estimate was to evaluate the 
overall intake of trans fat from the use of all PHOs for the purposes 
described previously in this section. Estimating trans fat intake from 
individual PHOs would be an impractical undertaking, and was not 
necessary for the purposes of the intake assessment.
    (Comment 25) Two comments stated that intake should be evaluated 
based on the presumption that all products with PHOs as an ingredient 
contain trans fat at a specified level (e.g., 0.2 g/serving or per 
reference amount customarily consumed). These comments suggested that 
such an assessment could provide support for an alternative approach 
such as setting an allowable level of trans fat in foods.
    (Response) Because we have concluded that PHOs are no longer GRAS, 
evaluating intake for alternative approaches, such as setting an 
allowable level of trans fat in foods, is not planned at this time.

B. Safety

    In the Federal Register of November 17, 1999 (64 FR 62746), we 
issued a proposed rule entitled ``Food Labeling: Trans Fatty Acids in 
Nutrition Labeling, Nutrient Content Claims, and Health Claims.'' The 
proposed rule would require that trans fat content be provided in 
nutrition labeling, and concluded that dietary trans fats have adverse 
effects on blood cholesterol measures that are predictive of CHD risk, 
specifically low-density lipoprotein cholesterol (LDL-C) levels (64 FR 
62746 at 62754). In the Federal Register of July 11, 2003 (68 FR 
41434), we issued a final rule (the July 2003 final rule) amending the 
labeling regulations to require declaration of trans fat content of 
food in the nutrition label of conventional foods and dietary 
supplements (68 FR 41434). In the July 2003 final rule, we cited 
authoritative reports that recommended limiting intake of trans fat to 
reduce CHD risk (68 FR 41434 at 41442).
    In the November 2013 notice containing our tentative determination 
that PHOs are no longer GRAS for any use in human food, we summarized 
findings reported in the literature since 2003, when we had last 
reviewed the adverse effects of dietary trans fat in support of the 
July 2003 final rule (68 FR 41434 at 41442 through 41449). We noted 
that since 2003, both controlled feeding trials and prospective 
observational studies published on trans fat consumption have 
consistently confirmed the adverse health effects of trans fat 
consumption on risk factor biomarkers (e.g., serum lipoproteins 
including LDL-C) and increased risk of CHD (78 FR 67169 at 67172). We 
describe these two types of studies (controlled feeding trials and 
prospective observational studies) in further detail later in this 
section. We also cited a variety of different kinds of studies and 
review articles showing that, in addition to an increased risk of CHD, 
trans fat consumption (and, accordingly, consumption of food products 
containing PHOs) has also been connected to a number of other adverse 
health effects (id.). These effects included worsening insulin 
resistance, increasing diabetes risk, and adverse effects on fetuses 
and breastfeeding infants, such as impaired growth.
    Since publication of the November 2013 notice, we re-reviewed key 
literature and expert panel reports published since the 1990s on the 
relationship between trans fat consumption and CHD risk (Ref. 18). Our 
review focused on the two main lines of scientific evidence linking 
trans fat intakes and CHD: (1) The effect of trans fat intake on blood 
lipids in controlled feeding trials, a type of randomized clinical 
trial; and (2) observational (epidemiological) studies of trans fat 
intake and CHD risk in populations. Additionally, we reviewed the 
conclusions of recent U.S. and international expert panels on the 
health effects of trans fat. As summarized in our review memorandum 
(Ref. 18), the scientific evidence, including combined analyses of 
multiple studies (meta-analyses), supports a progressive and linear 
cause and effect relationship between trans fatty acid intake and 
adverse effects on blood lipids that predict CHD risk, including LDL-C, 
high-density lipoprotein cholesterol (HDL-C) and ratios such as total 
cholesterol (total-C)/HDL-C and LDL-C/HDL-C. The observational 
(epidemiological) studies demonstrating increased CHD risk associated 
with trans fat intake do not prove cause and effect, but the results 
are consistent with and supportive of the evidence from controlled 
feeding trials of the adverse effect of trans fatty acid intake on 
blood lipids that predict CHD risk. The consistency of the evidence 
from two different study methodologies provides strong support for the 
conclusion that trans fatty acid intake has a progressive and linear 
effect that increases the risk of CHD.
    Risk factors are variables that correlate with incidence of a 
disease or condition. Risk factors include social and environmental 
factors in addition to biological factors. A biomarker is a 
characteristic that can be objectively measured and indicates 
physiological processes. A risk biomarker or risk factor biomarker is a 
biomarker that indicates a risk factor for a disease. In other words, 
it is a biomarker that indicates a component of an individual's level 
of risk for developing a disease or level of risk for developing 
complications of a disease (Ref. 19). LDL-C, HDL-C, total-C/HDL-C ratio 
and LDL-C/HDL-C ratio are all currently considered to be risk 
biomarkers for CHD (Refs. 19, 20, 21, and 22). LDL-C is a risk factor 
biomarker that is also a surrogate endpoint for CHD; a ``surrogate'' is 
a validated predictor of CHD and can substitute for actual disease 
occurrence in a clinical trial (Refs. 19, 20, and 21). HDL-C, total-C/
HDL-C and LDL-C/HDL-C are recognized as major risk factor biomarkers 
that, although they are not validated surrogate endpoints, are 
predictive of CHD risk (Refs. 19 and 22).
    Effect of trans fat intake on blood lipids in controlled feeding 
trials. In controlled feeding trials, a type of randomized clinical 
trial, trans fatty acid intake increased LDL-C (``bad'' cholesterol), 
decreased HDL-C (``good'' cholesterol) and increased ratios of total-C/
HDL-C and LDL-C/HDL-C compared with the same amount of energy intake 
(calories) from cis-unsaturated fatty acids. Increases in LDL-C, total-
C/HDL-C and LDL-C/HDL-C and decreases in HDL-C are adverse changes with 
respect to CHD risk. These adverse effects of trans fat intake on blood 
lipids are based on controlled feeding trials, a study design that is 
able to reveal cause and effect relationships between changes in trans 
fat intake and changes in blood lipids. In addition, increases in CHD 
risk with increases in LDL-C also demonstrate cause and effect. As 
described in our review memorandum (Ref. 18), combined analyses (meta-
analyses) of multiple controlled feeding trials demonstrate a 
progressive and linear relationship between trans fatty acid intake and 
adverse effects on blood lipids including LDL-C, HDL-C, total-C/HDL-C 
and LDL-C/HDL-C. The meta-analyses describe consistent quantitative 
relationships between trans

[[Page 34660]]

fat intake and blood lipids and show no evidence of a threshold below 
which trans fatty acids do not adversely affect blood lipids.
    Observational (epidemiological) studies of trans fat intake and CHD 
risk in populations. Epidemiology is the study of the distribution and 
causes of disease in human populations. Analytic epidemiology studies 
are those designed to test hypotheses regarding whether or not a 
particular exposure is associated with causing or preventing a specific 
disease outcome. In prospective observational (cohort) studies, 
subjects are classified according to presence or absence of a 
particular factor (such as usual dietary intake of trans fat) and 
followed for a period of time to identify disease outcomes (such as 
heart attack or death from CHD). Strengths of the prospective 
observational study design are that the time sequence of exposure and 
disease is clearly shown; exposures are identified at the outset of the 
study; and measurement of exposure is not affected by later disease 
status. Results of four major prospective studies, some with one or 
more updates during the followup period, consistently show higher trans 
fat intake associated with increased CHD risk. The association is 
positive and progressive, with no indication of a threshold. A 2009 
meta-analysis of the major prospective studies, based on almost 5,000 
CHD events in almost 140,000 subjects, found that each additional 2 
percent of energy intake from trans fat increased CHD risk by 23 
percent compared with the same energy intake from carbohydrate.
    Conclusions of recent U.S. and international expert panels on the 
health effects of trans fat. As described in our review memorandum 
(Ref. 18), international and U.S. expert panels, using additional 
scientific evidence available since 2002, have continued to recognize 
the positive linear trend between LDL-C and trans fat intake and the 
consistent association of trans fat intake and CHD risk in prospective 
observational studies. The panels have concluded that trans fats are 
not essential nutrients in the diet, and have recommended that 
consumption be kept as low as possible. Recommendations to avoid 
industrial trans fat intake have come from panels with both clinical 
and public health focus. Moreover, international and U.S. panels have 
expressed concern regarding population mean intakes of industrial trans 
fat intakes of 1 percent of energy and lower, recognizing that 
subgroups may be consuming relatively high levels.
    Since publication of the November 2013 notice, we also conducted a 
systematic search of the peer-reviewed literature published since 2008 
and summarized the findings (Ref. 23). The major human health endpoints 
evaluated for associations with trans fat intake reported in the 
literature included CHD, all-cause mortality, cardiovascular disease 
and stroke. Other human health endpoints addressed in our search 
included various types of cancer, metabolic syndrome and diabetes, and 
adverse effects on fertility, pregnancy outcome, cognitive function, 
and mental health. The literature search identified meta-analyses of 
published data; quantitative estimations to predict effects of 
replacing TFA in commercial products; cross-sectional, case-control and 
prospective observational cohort studies; and randomized controlled 
trials, including controlled feeding trials. Regarding cardiovascular 
diseases, the results of the literature search (Ref. 23) are consistent 
with findings discussed in our November 2013 notice (78 FR 67169 at 
67172). Findings associated with higher TFA intakes included increased 
risk of CHD, adverse effects on biomarkers associated with CHD, and 
increased subclinical atherosclerosis. Some recent prospective 
observational studies also found associations between increased trans 
fat intake and increased risk of stroke, which was a new finding (Refs. 
18 and 23). Further understanding of the apparent association between 
increased trans fat intake and increased risk of stroke requires 
additional research, such as whether the association may differ by age, 
sex, aspirin use, geographic region and other risk factors (Refs. 18, 
23, and 24). For the association of trans fat intake with other human 
health effects, such as various types of cancer, metabolic syndrome and 
diabetes, and adverse effects on fertility, pregnancy outcome, 
cognitive function and mental health, the literature reports remained 
limited or inconclusive.
    Since publication of the November 2013 notice, we also conducted a 
quantitative estimate of the potential health benefits expected to 
result from removal of IP-TFA from PHOs from the food supply (Ref. 25). 
We did this to analyze the expected public health benefit of removing 
PHOs from the food supply. We used four methods for estimating changes 
in CHD risk likely to result from replacement of IP-TFA: Method 1, 
based on effects of TFA on LDL-C, a validated surrogate endpoint 
biomarker for CHD, as shown through controlled feeding trials; Method 
2, based on effects of TFA on LDL-C plus HDL-C, a major CHD risk factor 
biomarker, as shown through controlled feeding trials; Method 3, based 
on effects of TFA on total-C/HDL-C plus a combination of emerging CHD 
risk factor biomarkers (lipoprotein(a), apolipoproteinB/
apolipoproteinA1 and C-reactive protein), as shown through controlled 
feeding trials; and Method 4, based on association of TFA with CHD risk 
as shown through prospective observational studies. Methods 1 and 2 
were also used by FDA in analyzing the 1999 and 2003 labeling 
regulations (64 FR 62746 at 62768 and 68 FR 41434 at 41479) and Methods 
3 and 4 were based on published methods (Ref. 26). We estimated the 
change in CHD risk using each of these four methods as applied to two 
different sets of scenarios for replacement of IP-TFA, as follows.
    In general, fats and oils in foods have carbon chains of various 
lengths, with the carbon atoms in these chains connected by single or 
double bonds. If the carbon chain contains no double bonds, the fatty 
acid is called saturated. If the carbon chain contains a single double 
bond, the fatty acid is called monounsaturated, and if the carbon chain 
contains two or more double bonds, the fatty acid is called 
polyunsaturated. Most naturally-occurring dietary unsaturated fatty 
acids have double bonds in a ``cis'' configuration, that is, the two 
hydrogen atoms attached to two carbons are on the same side of the 
molecule at the double bond. Thus, the major chemical forms of fatty 
acids in foods are saturated fatty acids (SFAs), cis-monounsaturated 
fatty acids (cis-MUFAs) and cis-polyunsaturated fatty acids (cis-
PUFAs). (By comparison, in a ``trans'' configuration, the hydrogen 
atoms attached to the carbon atoms at a double bond are not on the same 
side of the double bond). (See definitions in 64 FR 62746 at 62748 to 
62749 (November 17, 1999).)
    One set of scenarios focuses solely on IP-TFA and the estimated 
change in CHD risk by hypothetically replacing IP-TFA with each of the 
major chemical forms of macronutrient fatty acids in foods--i.e., SFAs, 
cis-MUFAs or cis-PUFAs. The other set of scenarios focuses not only on 
IP-TFA but also on the other fatty acids contained in PHOs. This 
hypothetical set of scenarios illustrates the estimated change in CHD 
risk with replacing PHOs in the marketplace that contain 20 percent, 35 
percent, or 45 percent IP-TFA, with other likely replacement fats and 
oils. Therefore, this scenario accounts for not only the replacement of 
IP-TFA with macronutrient fatty acids but also the replacement of the 
overall fatty acid components (or profiles) of the PHOs with the fatty 
acid components (or

[[Page 34661]]

profiles) found in the various replacement fats and oils.
    In the first set of scenarios, we assumed that the current mean 
intake of 0.5 percent of total daily calories (energy) from IP-TFA 
among U.S. adults was replaced by the same percent of energy from three 
types of macronutrient fatty acids, cis-mono- or polyunsaturated fatty 
acids and saturated fatty acids) (cis-MUFAs, cis-PUFAs, and SFAs). As 
measures of risk reduction, we calculated estimated percent changes in 
CHD risk and estimated reduction in annual total cases of CHD, 
including CHD-related deaths. We based changes in CHD cases and deaths 
on a baseline of 915,000 annual new and recurrent fatal and non-fatal 
cases of CHD in U.S. adults, with a 41 percent fatality rate (Ref. 27).
    Results showed an estimated reduction in CHD with replacement of 
IP-TFA with each of the fatty acids (cis-MUFA, PUFA, or SFA), using 
each of the four estimation methods. The estimated decrease in CHD 
ranged from 0.1 percent to 6.0 percent. This corresponded to prevention 
of 1,180 to 7,510 annual CHD cases, including 490 to 3,120 deaths, in 
Method 1 (0.1 percent to 0.8 percent decrease in CHD risk based on LDL-
C), 9,230 to 15,560 cases, including 3,830 to 6,460 deaths, in Method 2 
(1.0 percent to 1.7 percent decrease in CHD risk based on LDL-C and 
HDL-C), and 18,660 to 54,900 cases, including 7,740 to 22,770 deaths, 
in Method 3 (2.0 percent to 2.5 percent decrease in CHD risk using a 
combination of biomarkers) and Method 4 (4.2 percent to 6.0 percent 
decrease in CHD risk using observed CHD outcomes). Method 4, based on 
long-term observations of CHD outcomes in prospective studies, produced 
greater reduction estimates in risk than did Methods 1 and 2, which 
were based on short-term changes in blood lipid risk factors in 
controlled feeding trials. This suggests that there may be additional 
mechanisms, besides changes in blood lipids, through which trans fat 
consumption contributes to CHD risk. Thus, the adverse effects from 
trans fat intake may be greater than predicted solely by changes in 
blood lipids. The greater estimated reduction in CHD in Method 3, 
compared with Methods 1 and 2, suggests that the emerging risk factor 
biomarkers in Method 3 may help to identify additional mechanisms 
through which trans fat contributes to CHD risk.
    In the second set of scenarios, we estimated the reduction in risk 
by replacing the same 0.5 percent of energy from IP-TFA, along with the 
other component fatty acids in three different formulations of PHOs, 
with eight alternative fats and oils (soybean oil, canola oil, 
cottonseed oil, high oleic sunflower oil, high oleic soybean oil, palm 
oil, lard, and butter). This approach covers a range of composition of 
replacement fats and oils, from highly saturated (high in SFAs) to 
highly unsaturated (high in cis-MUFAs and/or cis-PUFAs), and is based 
on that reported in 2009 by Mozaffarian and Clarke as part of the World 
Health Organization (WHO) scientific update on trans fatty acids (Refs. 
25 and 26). Among the eight fats and oils, soybean oil and cottonseed 
oil contain the highest amounts of cis-PUFAs. Canola oil, high oleic 
acid sunflower oil, and high oleic acid soybean oil have the highest 
amounts of cis-MUFAs. Butter has the highest amount of SFAs; lard and 
palm oil are also high in SFAs. We used the same four methods to 
estimate risk reduction in this analysis. These calculations take into 
account the fatty acid profiles of the replacement fats and oils and 
the other fatty acids in the PHOs in addition to IP-TFA.
    Overall, the analysis showed that removing 0.5 percent of energy 
from IP-TFA by replacing an example PHO containing 35 percent IP-TFA 
with each of eight alternative fats and oils would reduce CHD risk by 
0.4 percent to 1.5 percent across the respective replacement fats and 
oils using Method 2, 2.3 percent to 3.0 percent using Method 3, and 2.7 
percent to 6.4 percent using Method 4. This would correspond to 
prevention of 3,900 to 58,210 CHD cases including 1,620 to 23,350 CHD 
deaths per year.
    In a few instances, the analysis in the second set of scenarios 
estimated that there would be increased CHD risk when examples of PHOs 
were replaced entirely with fats or oils high in saturated fat (Ref. 
25) using Method 1. This reflects the saturated fatty acids in 
alternative fats and oils replacing the cis-unsaturated fatty acids 
present in the PHO in addition to IP-TFA. Method 1 alone likely 
underestimates the overall change in risk that would result from 
replacing PHOs containing IP-TFA because it analyzes only impacts on 
LDL-C alone and therefore does not account for the demonstrated adverse 
effects of IP-TFA on HDL-C, or the adverse effects of IP-TFA on other 
emerging CHD risk factors. Methods 2, 3, and 4 in the second set of 
scenarios, which consider other known risk factors as well as LDL-C, 
provides a more thorough estimate of risk reduction than considering 
only LDL-C in isolation, and leads us to conclude that there would be 
an expected benefit to public health from PHO replacement even if PHOs 
are replaced by oils high in saturated fat. Consistent with published 
analyses, our results show that estimated changes in CHD risk expected 
to occur with replacement of PHOs depends on the fatty acid profiles of 
both the PHOs and the replacement fats and oils (Refs. 25, 26, and 28). 
We also note that research indicates removal of trans fat over the past 
decade has generally not been accompanied by extensive increases in 
saturated fat (Ref. 29), suggesting that all IP-TFA currently in the 
marketplace would not likely be replaced by oils high in saturated fat.
    Among the strengths of our quantitative analyses is the use of 
established cause and effect relationships between IP-TFA intakes and 
adverse changes in CHD biomarker risk factors, including LDL-C and HDL-
C, derived from high quality, controlled feeding trials. Our 
assessments also relied on a set of emerging risk factors for CHD, 
including total cholesterol to HDL-C ratios, Apo-lipoprotein B to Apo-
lipoprotein A-I ratios, lipoprotein(a) and C-reactive protein changes 
obtained from these same feeding trials. In addition, we relied on 
information from direct observations of CHD outcomes associated with 
frequent usual intake assessments of trans fatty acids and other 
macronutrient fatty acids in meta-analyses of four large cohorts with 
long-term followups. These estimates build on the agency's previous 
quantitative assessment based on short-term changes in LDL-C and HDL-C 
alone (68 FR 41434 at 41466 to 41492).
    We acknowledge that there are always some uncertainties in 
assessing risk. The estimates we used were based on 100 percent 
replacement of IP-TFA by a group of individual types of fatty acids or 
by individual alternative fats and oils, when actual replacement mixes 
of fats and oils might vary and individual diets would reflect a 
combination of replacement fatty acids and replacement fats and oils. 
We assumed a no threshold, linear relationship between changes in IP-
TFA intakes and changes in biomarker risk factors for CHD because 
current scientific evidence indicates that the relationship between 
trans fatty acid intake and LDL-C, HDL-C and the total cholesterol to 
LDL cholesterol ratio is progressive and linear.
    Given these uncertainties, our assessments for the change of CHD 
risk at the current U.S. mean daily intake of 0.5 percent of energy 
derived from IP-TFA are conservative estimates. The results also 
suggest that a small shift to lower CHD risk could prevent large 
numbers of annual cases of CHD and CHD-related deaths. The current U.S.

[[Page 34662]]

background rates for CHD are already high, with considerable baseline 
variability due to abnormal serum lipid profiles in large percent of 
U.S. adults (33.5 percent have elevated LDL-C) and other risk factors 
for CHD (Ref. 25). More people may be vulnerable to CHD at the current 
mean intake of IP-TFA from PHOs than the risk reduction estimates as 
discussed above.
    In sum, our quantitative estimates demonstrate that large numbers 
of CHD events and deaths may be prevented with the elimination of PHOs. 
We also note that our estimates are in line with published results 
regarding potential effects of replacing PHOs (Refs. 26 and 28). In 
replacing PHOs containing IP-TFA, a more significant reduction in CHD 
risk is estimated by replacement with vegetable oils containing higher 
amounts of cis-unsaturated fatty acids than with those high in 
saturated fatty acids, but we expect a risk reduction even if IP-TFA is 
replaced with fats and oils high in saturated fatty acids, based on our 
conservative risk estimates using combinations of the four peer-
reviewed methods with two different sets of likely scenarios for IP-TFA 
replacement for each method. Additional details of these results, and 
results for replacement of example PHOs containing 20 percent IP-TFA 
and 45 percent IP-TFA, are provided in our review memorandum (Ref. 25).
    We have also analyzed the comments we received regarding the 
scientific basis for our tentative determination in the November 2013 
notice. Comments regarding the safety of PHOs that were opposed to our 
tentative determination were generally related to one of four subject 
areas: (1) Dose-response relationship of trans fat intake and adverse 
health effects in human studies and whether there is a threshold below 
which intake of trans fats is generally recognized as safe; (2) 
reliance on expert panel reports and recommendations; (3) health 
benefits and clinical significance of replacements for PHOs; and (4) 
alternative approaches. Comments regarding the safety of PHOs that were 
in support of our determination raised concerns about other adverse 
health effects besides effects on LDL-C, such as adverse effects on 
other risk factors for CHD (e.g., HDL-C, total-C/HDL-C ratio, LDL-C/
HDL-C ratio, and other lipid and non-lipid biomarkers), inflammatory 
effects, harm to subpopulations, and increased diabetes risk.
1. Dose-Response and Evidence of a Threshold Level
    (Comment 26) A number of comments stated that the studies relied 
upon by FDA were not designed to address the impact of lowering TFA 
intake below 1% of energy. The comments asserted that although the 
expert panel reports state that there is no threshold intake level for 
IP-TFA that would not increase an individual's risk of CHD or adverse 
effects on risk factors for CHD, a review of the supporting 
documentation accompanying the reports does not support this statement; 
rather, the comments noted that panel reports indicate that due to the 
paucity of evidence in the 0 to 4% energy range, no evidence-based 
conclusions could be made.
    (Response) FDA disagrees; the published research described in our 
review memorandum (Ref. 18) includes six regression analyses of 
controlled feeding trials summarizing the dose-response relationship of 
IP-TFA on blood cholesterol levels, published from 1995 to 2010. In 
addition, a 2010 meta-analysis included 23 trans fat feeding trials and 
28 TFA levels, including a low-dose level of 0.4 percent of energy (or 
less than the current mean intake) (Ref. 30). Across these regression 
analyses, the reported effect of TFA on LDL-C, a validated surrogate 
biomarker that serves as a direct causal link to CHD, was very 
consistent and the analyses showed a linear dose-response, with an 
increase in LDL-C of about 0.038 to 0.049 millimoles per liter (mmol/L) 
for each 1 percent of energy intake from replacement of cis-
monounsaturated fat with trans fat (Table 3 in Ref. 18). The regression 
analyses also showed a consistent linear dose response for HDL-C, with 
a decrease of about 0.008 to 0.013 mmol/L for each 1 percent of energy 
from replacement of cis-monounsaturated fat with trans fat (Table 3 in 
Ref. 18). Therefore, we conclude that the available data show that even 
at low intake levels (e.g., below 3 percent energy) there is no 
identifiable threshold, rather the available data support a conclusion 
that IP-TFA causes a linear increase in blood levels of LDL-C, a 
validated surrogate biomarker of CHD risk and a linear decrease in 
blood levels of HDL-C, a major risk biomarker for CHD. If interested 
parties are or become aware of information and data supporting 
establishment of a threshold, such information and data could be 
submitted to FDA as part of a food additive petition(s) proposing safe 
conditions of use for PHOs.
    (Comment 27) Many comments disagreed with our conclusion that there 
is a linear relationship between TFA intake and LDL-C at low TFA intake 
levels. Some comments stated that we did not establish causality 
between low doses of TFA (less than 1% of caloric energy) and increased 
CHD risk. Other comments stated that the review of available data shows 
that low levels of TFA intake (3% of energy or less) have no effect on 
serum LDL-C and total-C levels. Some comments criticized FDA's reliance 
on the Ascherio et al. 1999 paper (Ref. 31) and raised issues with this 
paper and the linear extrapolation used by the researchers. One comment 
suggested that using a different dose-response model is a more 
appropriate approach to determine the relationship between PHOs and 
LDL-C and HDL-C, rather than defaulting to a linear function, due to 
the quantity and type of data available at low intake levels. One 
comment stated that, in general, linear regression is an inappropriate 
tool to determine a safe or unsafe level of a dietary substance and 
questioned the use of low-dose linear extrapolation in this instance.
    (Response) FDA disagrees with these comments. Given that effects of 
trans fat on LDL-C have been demonstrated at doses as low as 0.4 
percent and 2.8 percent of caloric energy (Table 2 in Ref. 18), FDA 
disagrees that there is no evidence of an adverse effect from trans fat 
intake below 3 percent of energy. In addition, results of regression 
analyses published from 1995 to 2010, including Ascherio et al. 1999 
(Refs. 26, 30, 31, 32, 33, and 34), are very consistent regarding the 
effect of TFA on serum lipids, thus indicating that the relationship 
between TFA intake and CHD risk is progressive and linear with no 
evidence of a threshold at which effects would not be expected to 
occur. Furthermore, we are not aware of any published study that 
supports an abrupt reduction in the adverse effects of TFA across the 
relatively narrow intake range of 0 percent to 3 percent of energy nor 
are we aware of any published scientific reports that provide a dose-
response model that might reveal a different relationship for TFA 
intake and CHD risk that is generally accepted by qualified experts. 
FDA is aware of an unpublished meta-regression analysis, including 
consideration of the low-intake range (Ref. 35), suggesting that the 
data on dietary trans fat intake and changes in LDL-C may fit a dose-
response curve that is non-linear. However, this analysis is neither 
published (generally available) nor does it demonstrate a consensus of 
expert opinion that the use of PHOs at low levels in food is safe as 
required for general recognition of safety.\2\
---------------------------------------------------------------------------

    \2\ FDA also reviewed and considered an unpublished report of 
this analysis and its executive summary, which were submitted to FDA 
with the request that they be kept confidential. FDA is including 
these documents in the administrative record for this matter but is 
not placing them in the public docket because they are confidential.

---------------------------------------------------------------------------

[[Page 34663]]

    Further, we did not rely solely on the Ascherio et al. 1999 paper 
regarding the effect of IP-TFA intake on serum LDL-C and other lipid 
biomarkers. Over time, the number of studies covered by the published 
regression analyses or meta-analyses increased from 5 studies and 6 TFA 
levels in 1995 (Ref. 32) through 8 studies and 12 TFA levels in 1999 
(Ref. 31) to 23 studies and 28 TFA levels in 2010 (Ref. 30). Across 
these studies, the reported magnitude of the effect of IP-TFA on LDL-C 
and HDL-C levels is very consistent. Furthermore, FDA notes that the 
2009 National Research Council report, Science and Decisions: Advancing 
Risk Assessment (Ref. 36), describes conceptual models in which low-
dose linearity with no threshold can arise. Absent evidence of a 
threshold intake level for TFA that does not increase an individual's 
risk of CHD or adverse effects on risk factors for CHD, FDA concludes 
that a linear low-dose extrapolation is appropriate for assessing the 
dose-response relationship between TFA intake and risk of CHD (as 
evidenced by effects on LDL-C, a validated surrogate biomarker for CHD, 
and HDL-C, a risk biomarker (Ref. 18)).
    Our conclusion that there is a linear relationship (also known as a 
proportional effect, or proportionality) between trans fat intake and 
CHD risk is consistent with the body of evidence from controlled 
feeding studies on the proportionality of fatty acid intake and blood 
lipids, beginning with landmark studies in the 1950s and 1960s (Refs. 
18, 37, 38, 39, and 40). Meta-analyses in the 1990s and early 2000s 
showed that the proportionality in the earlier landmark studies 
extended not only to total cholesterol but to LDL-C, HDL-C, total-C/
HDL-C ratio and LDL-C/HDL-C ratio (Refs. 33, 41, and 42). Authors of a 
1992 meta-analysis noted, ``a simple linear model in which diets are 
characterized solely by their contents of saturated, monounsaturated 
and polyunsaturated fatty acids goes a long way toward predicting group 
mean changes in serum lipid and lipoprotein levels'' (Ref. 42). Results 
of an early controlled feeding trial of trans fat intake and LDL-C and 
HDL-C were questioned because of the high trans fat intake (Ref. 43). 
However, when combined with a subsequent study at a lower dose, 
preliminary data from these two studies suggested that the effect of 
trans fat intake on LDL-C and HDL-C is proportional (Ref. 18). 
Subsequent meta-analyses discussed previously supported the linear 
proportionality of the data, and the quantitative relationships of 
dose-response are very consistent across the analyses (Ref. 18). The 
proportional relationship of trans fat intake and blood lipids has also 
been repeatedly affirmed by a series of expert panels (Ref. 18). 
Therefore, we conclude that the totality of the data supports the 
proportionality of changes in trans fat intake and changes in blood 
lipids (and therefore, CHD risk) and supports the use of a linear 
regression model to describe this relationship.
    (Comment 28) Some comments objected to the approach of ``forcing'' 
the regression line of the dose-response curve through zero (the 
origin), as done by Ascherio et al. 1999 (Ref. 31) and believed this 
was not appropriate.
    (Response) FDA disagrees. Whether or not to fix the intercept at 
zero depends on the meaning of the data, the research question to be 
answered, and the particular study design. (We further discuss the 
methodology for the meta-analyses in our review memorandum (Ref. 18)). 
In feeding studies where the total energy intake remains the same for 
both control and treatment groups, the zero intercept means that, with 
zero intake of trans fat, there is no effect of trans fat on (that is, 
no change in) the LDL-C, the LDL-C/HDL-C ratio, or other serum lipid 
biomarker being studied. This is the one data point that is known to be 
true by virtue of the study design, and many analyses using this 
approach have been published in peer-reviewed literature (Refs. 30, 31, 
32, 44, and 45). In these analyses, the authors calculated the 
differences in serum lipid levels between the trans fat diet and the 
control diet for each controlled feeding trial, with adjustment for 
differences in intake of the other fatty acids between the two diets, 
using published dose-response coefficients (Refs. 33 and 42). The serum 
lipid and trans fat intake differences for each study were included in 
a linear regression model and expressed with respect to a specific 
replacement macronutrient (such as cis-monounsaturated fatty acids or 
carbohydrate). Therefore, we conclude that it is logical and 
appropriate to fit (not ``force'') the regression lines through zero 
because a zero change in trans fat intake results in zero change in 
blood lipids attributable to trans fat intake.
    (Comment 29) Some comments criticizing our scientific review stated 
that prospective observational (epidemiological) studies which we 
relied on were not designed to demonstrate a cause and effect 
relationship between a substance and a disease, and are subject to 
various forms of bias.
    (Response) Although observational studies with long-term followup 
do not prove cause and effect, the results are consistent with and 
supportive of the conclusions from the controlled feeding trial 
evidence discussed previously in this section (which does demonstrate 
cause and effect). The consistency of the evidence from two different 
study methodologies is strong support for the conclusion that trans 
fatty acid intake has a progressive and linear effect that increases 
the risk of CHD. Our review memorandum (Ref. 18) provides a summary of 
the scientific evidence from the observational studies on the 
association of TFA intake and actual CHD outcomes in large populations 
and addresses in detail the study designs and adjustments for 
confounding variables. There are four major prospective observational 
studies (Refs. 46, 47, 48, 49, 50, 51, and 52), some with one or more 
updates during the followup period (e.g., the Nurses' Health Study had 
followups at 8, 14, and 20 years), that are further discussed in detail 
in one of our review memoranda (Ref. 18). These are prospective 
(cohort) studies, which is the strongest study design for observational 
studies, and the results consistently show that higher trans fat intake 
is associated with increased CHD risk. In several studies, not only was 
the association of the highest versus lowest level (category) of trans 
fat intake with greater CHD risk statistically significant, but also 
there was a significant test for linear trend, indicating a positive 
and progressive association of trans fat intake with CHD risk (or CHD 
deaths) across levels (low, intermediate, or high categories) of intake 
(Refs. 46, 48, 49, 50, and 51). In addition to the analysis of trans 
fat intake grouped in several levels or categories, in certain studies, 
numerical trans fat intake, as a continuous variable, was significantly 
associated with CHD risk, again indicating a positive and progressive 
association of increased trans fat intake with increased CHD risk 
across the range of observed intake (Refs. 49 and 51).
    There are also a number of meta-analyses of the major prospective 
studies (Refs. 26, 51, 52, 53, 54, and 55). In a 2009 meta-analysis, 
based on almost 5,000 CHD events in almost 140,000 subjects, each 
additional 2 percent of energy intake from trans fat increased CHD risk 
by 23 percent compared with the same energy intake from carbohydrate 
(Ref. 52). The magnitude of the increase in CHD risk associated with 
trans fat intake among

[[Page 34664]]

meta-analyses has remained consistent over time, including the studies 
with additional updates during the followup periods. Further, the 
prospective studies measure actual CHD occurrence in large groups of 
people over long time periods, and describe all CHD risk associated 
with trans fat intake, regardless of the mechanism of action by which 
trans fat intake may be associated with CHD (i.e., these studies do not 
rely on biomarkers or risk factors but instead measure actual 
occurrence of disease). The magnitude of the observed CHD risk from TFA 
intake is greater in the prospective observational studies than from 
the controlled feeding studies.
    We also reviewed related observational studies of TFA intake and 
cardiovascular disease health outcomes that considered all causes of 
mortality and cardiovascular disease endpoints other than CHD, as well 
as studies that used blood and tissue levels as biomarkers of TFA 
intake instead of dietary questionnaires, and retrospective case 
control studies (Ref. 18). The results from these studies generally 
showed trans fat intake or biomarkers associated with adverse health 
outcomes. The consistent findings of adverse health effects of trans 
fat from these studies with different methodologies strengthen our 
conclusions based on the evidence from the major prospective 
observational studies and controlled feeding studies summarized 
previously.
    (Comment 30) Several comments cited a 2011 publication by FDA 
authors (Ref. 56) as evidence of PHO safety and evidence that a 
threshold can be determined below which there is general recognition of 
safety. The comments argued that these authors reviewed data from 
clinical trials to assess the relationship between trans fat intake and 
LDL-C and total-C and that their regression analysis showed no 
association between trans fat consumption and either LDL-C or total-C 
levels. Also, the comments stated that the authors do not ``force'' the 
regression line through zero unlike in the Ascherio et al. 1999 paper, 
relied upon by FDA in the tentative determination.
    (Response) FDA disagrees. We note that the authors of this paper 
stated that their regression analysis of TFA intake and LDL-C 
``supports the IOM's conclusion that any intake level of trans fat 
above 0 percent of energy increased LDL cholesterol concentration.'' 
This paper did not identify a threshold level at which LDL-C began to 
increase. The analysis in the paper was limited to validated surrogate 
endpoint biomarkers of CHD, total cholesterol and LDL-C, and did not 
consider other CHD risk factor biomarkers such as HDL-C, or total-C/
HDL-C or LDL-C/HDL-C ratios. The paper focused on methodology for 
attempting to identify a tolerable upper intake level for trans fat. 
The appropriateness of fitting the intercept through zero in a 
regression analysis depends on the meaning of the data, the research 
question to be answered, and the particular study design, and is 
discussed further in our response to Comment 28.
    In addition to the feeding trial data discussed in the 2011 
publication, the authors of the 2011 paper presented data from 
prospective observational studies showing that, compared with the 
lowest trans fat intake level, there was a statistically significant 
increase in CHD risk at some levels of trans fat intake, but not at 
others. Based on this, they stated that, at least theoretically, ``a 
threshold level could be identified for trans and saturated fat,'' but 
they were not actually able to identify any specific threshold level. 
We note that other data from prospective studies that were not 
discussed in this paper support the conclusion that there is a direct 
and progressive relationship between TFA intake and CHD risk, and no 
threshold has been identified. Several studies showed a positive trend 
for higher CHD risk with higher intake categories of TFA that was 
statistically significant (Refs. 46, 48, 49, 50, and 51) and certain 
studies also analyzed numerical TFA intake without using categories 
(that is, as a continuous variable) and found a significant positive 
linear association of TFA intake with CHD risk across the range of 
usual TFA intake levels of participants in the studies (Refs. 49 and 
51). These results, not discussed in the paper, are inconsistent with 
the existence of a threshold. Therefore, we conclude that there is no 
currently identifiable threshold below which there is general 
recognition that PHOs may be safely used in human food. However, if 
there are data and information that demonstrates to a reasonable 
certainty that no harm will result from a specific use of a PHO in 
food, that information could be submitted as part of a food additive 
petition to FDA seeking issuance of a regulation to prescribe 
conditions under which the additive may be safely used in food.
    (Comment 31) Some comments stated that FDA made conclusions that 
any incremental increase in trans fat intake increases the risk of CHD 
based on endpoints that are not considered validated surrogate 
biomarkers for CHD, such as LDL-C/HDL-C ratio in the Ascherio et al. 
1999 paper (Ref. 31).
    (Response) We used LDL-C, a validated surrogate endpoint biomarker 
for CHD (Ref. 21), as the primary endpoint for evaluating the adverse 
effects of IP-TFA intake from PHOs. As discussed previously in this 
section, validated surrogate endpoint biomarkers are those that have 
been shown to be valid predictors of disease risk and may therefore be 
used in place of clinical measurement of the incidence of disease 
(Refs. 19 and 20). In addition, we considered the adverse effects of 
trans fat intake on other risk factor biomarkers, including HDL-C and 
the LDL-C/HDL-C and total-C/HDL-C ratios. In fact, these other risk 
factor biomarkers indicate additional adverse effects of IP-TFA, beyond 
the primary adverse effect of raising LDL-C. Although these other risk 
factor biomarkers are not validated surrogate endpoint biomarkers for 
CHD, they raise significant questions about the safety of PHOs and are 
therefore relevant to our determination that PHOs are not GRAS. For 
example, HDL-C levels have been shown to be a useful predictor of CHD 
risk (Refs. 22 and 57). Because it has not been shown that drug therapy 
to raise HDL-C decreases CHD in clinical trials, HDL-C is not 
considered a validated surrogate endpoint for CHD (Ref. 19). We did not 
primarily rely on the relationship between trans fat intake and adverse 
effects on HDL-C and CHD risk, we recognize that a relationship is 
known to exist and therefore considered it in our analysis. We 
discussed this issue in detail in the July 2003 final rule (68 FR at 
41434 at 41448 through 41449).
    Recent studies have affirmed HDL-C and total-C/HDL-C ratio as risk 
factors that predict CHD (Ref. 18). In a large, pooled meta-analysis of 
prospective observational studies, including 3,020 CHD deaths during 
1.5 million person-years of followup, each 1.33 unit decrease in the 
total-C/HDL-C ratio was associated with a 38 percent decrease in risk 
of CHD death (Ref. 22). Each 0.33 mmol/L decrease in HDL-C was 
associated with a 61percent higher risk of CHD death. The authors 
concluded: ``HDL cholesterol added greatly to the predictive ability of 
total cholesterol.'' They stated: ``Higher HDL cholesterol and lower 
non-HDL cholesterol levels were approximately independently associated 
with lower IHD [CHD] mortality, so the ratio of total/HDL cholesterol 
was substantially more informative about IHD mortality than either, and 
was more than twice as informative as total cholesterol'' (Ref. 22).
    (Comment 32) One comment stated that safety evaluation of 
macronutrients,

[[Page 34665]]

such as PHOs, is very complex and requires a far more robust assessment 
of the totality of technical and scientific evidence. The comment 
criticized FDA for relying on ``an isolated physiological endpoint such 
as serum lipoproteins'' as predictive of CHD, and states that this 
methodology is not appropriate for a GRAS assessment.
    (Response) FDA disagrees; the results of feeding trials showing 
changes in LDL-C, a validated surrogate endpoint biomarker for CHD, and 
other risk factor biomarkers, are supported by the results of 
observational studies showing actual CHD disease outcomes (heart 
attacks and deaths) associated with TFA intake in large populations. 
The consistency of the evidence from two different study methodologies 
is strong support for the conclusion that trans fatty acid intake has a 
progressive and linear effect that increases the risk of CHD. Such 
health effects are appropriate for FDA to consider when assessing the 
safety of food ingredients.
2. Expert Panel Reviews and Recommendations
    The November 2013 notice discussed expert panel conclusions and 
recommendations, including the 2002/2005 IOM reports. The conclusions 
and recommendations of this report have since been affirmed by a series 
of U.S. and international expert panels. The recent expert panels have 
continued to recognize the progressive linear relationship between LDL-
C (increase) and HDL-C (decrease) and trans fat intake, and have 
concluded that trans fats are not essential nutrients in the diet and 
consumption should be kept as low as possible. We have compiled a 
detailed summary of the expert panel reports in a review memorandum 
(Ref. 18).
    (Comment 33) Some comments stated that FDA should convene an expert 
panel to specifically address whether evidence exists to indicate the 
effect of TFA on LDL-C is linear at low intakes (below 3% energy). 
Other comments stated that there is consensus among qualified experts 
that TFA intake should be less than 1% of energy, and cited expert 
panel reviews as evidence. Similar comments stated that PHOs are safe 
at current intake levels, and TFA intake is already below levels 
recommended by nutrition experts.
    (Response) We decline to convene another expert panel in light of 
the substantial evidence available on the adverse effects of consuming 
trans fat. FDA notes that a 2013 National Institutes of Health, 
National Heart, Lung, and Blood Institute (NIH/NHLBI) expert panel 
conducted a systematic evidence review and concluded with moderate 
confidence that, for every 1 percent of energy from TFA replaced by 
mono- or polyunsaturated fatty acids (MUFA or PUFA), LDL-C decreases by 
an estimated 1.5 milligrams per deciliter (mg/dL) and 2.0 mg/dL, 
respectively (Ref. 58). The panel also concluded that replacement of 
TFA with saturated fatty acids (SFA), MUFA, or PUFA increases HDL-C by 
an estimated 0.5, 0.4 and 0.5 mg/dL, respectively. This panel's 
conclusions were not limited to a specific TFA dose range and did not 
indicate any threshold TFA intake. The conclusions were based on 
previously published linear regression analyses (Refs. 26 and 33).
    We also disagree that, based on generally available information, 
there is a consensus among qualified experts that trans fats are safe 
at some level, and we note that recommendations from expert panels 
either: (1) Do not state a recommended level (Ref. 13); or (2) 
recommend consideration of further reduction in IP-TFA intake, below 
current levels (Refs. 59, 60, 61, and 62). Since 2002, many expert 
panels have considered the adverse effects associated with trans fat 
consumption. Table 1 provides a list of organizations that have 
published reports on trans fat and indicates whether they have 
conducted an evidence review and/or made formal intake recommendations 
regarding trans fat consumption. The conclusions and recommendations 
made by these organizations further demonstrate a lack of consensus 
regarding the safety of PHOs, as the primary dietary source of IP-TFA.

                     Table 1--List of Organizations That Have Published Reports on Trans Fat
----------------------------------------------------------------------------------------------------------------
                                                                                                Formal trans fat
            Organization                    Report title           Year      Evidence  review        intake
                                                                             and  conclusions    recommendation
----------------------------------------------------------------------------------------------------------------
IOM.................................  Dietary Reference          2002/2005                 X                  X
                                       Intakes for Energy and
                                       Macronutrients (Ref.
                                       7).
European Food Safety Authority,       Opinion on the presence         2004                 X   .................
 Scientific Panel on Dietetic          of trans fatty acids
 Products, Nutrition and Allergies.    in foods and the
                                       effect on human health
                                       of the consumption of
                                       trans fatty acids
                                       (Ref. 63).
FDA Food Advisory Committee,          Subcommittee Meeting,           2004                 X   .................
 Nutrition Subcommittee.               Summary Minutes (Ref.
                                       14).
Dietary Guidelines Advisory           Report of the 2005 DGAC         2005                 X   .................
 Committee (DGAC).                     (Ref. 64).
U.S. Dept. of Health and Human        Dietary Guidelines for          2005  .................                 X
 Services, U.S. Dept. of Agriculture   Americans (Ref. 12).
 (DHHS/USDA).
World Health Organization (WHO).....  Scientific Update on            2009                 X                  X
                                       Trans Fatty Acids
                                       (Ref. 60).
Food and Agriculture Organization,    Background Papers for           2009                 X   .................
 World Health Organization (FAO,       Expert Consultation on
 WHO).                                 Fats and Fatty Acids
                                       in Human Nutrition
                                       (Ref. 59).
FAO, WHO............................  Expert Consultation on          2010                 X                  X
                                       Fats and Fatty Acids
                                       in Human Nutrition
                                       (Ref. 61).
DGAC................................  Report of the 2010 DGAC         2010                 X   .................
                                       (Ref. 65).
DHHS/USDA...........................  Dietary Guidelines for          2010  .................                 X
                                       Americans (Ref. 13).
NHLBI...............................  Evidence Report on              2013                 X   .................
                                       Lifestyles
                                       Interventions to
                                       Reduce Cardiovascular
                                       Risk (Ref. 58).
American College of Cardiology,       Guideline on Lifestyle     2013/2014  .................                 X
 American Heart Association.           Management to Reduce
                                       Cardiovascular Risk
                                       (Ref. 62).
----------------------------------------------------------------------------------------------------------------


[[Page 34666]]

3. Safety of Replacements for IP-TFA in PHOs
    (Comment 34) Several comments questioned whether further reductions 
in TFA intake will be clinically significant and subsequently affect 
public health.
    (Response) Since publication of the November 2013 notice, we have 
quantitatively analyzed the public health significance of removing PHOs 
from the food supply (Ref. 25), and the results show that removing PHOs 
from human food would have an expected positive impact on public 
health. We note that further reductions in IP-TFA intake below current 
levels may result in small reductions in LDL-C and small improvements 
in other biomarkers that may not seem clinically significant for an 
individual; however, when considered across the U.S. population, small 
reductions in CHD risk would be expected to prevent large numbers of 
heart attacks and deaths, as illustrated in FDA estimates (Ref. 25). 
Moreover, the 2013 Guideline on Lifestyle to Reduce Cardiovascular Risk 
from the American College of Cardiology and the American Heart 
Association (Ref. 62) strongly recommends that clinicians advise adults 
who would benefit from LDL-C reduction to reduce their percentage of 
calories from trans fat (the report notes that the majority of U.S. 
adults have one or more risk factors involving abnormal lipids, high 
blood pressure or pre-high blood pressure; 33.5 percent of adults have 
elevated LDL-C). Therefore, further reduction in IP-TFA intake below 
current levels is expected to be clinically significant and to prevent 
a large number of heart attacks and deaths in the United States.
    (Comment 35) Some comments stated that the safety implications of 
replacing TFA with other nutrients (e.g., saturated fat, unsaturated 
fat, carbohydrates) have yet to be determined.
    (Response) We recognize that removing PHOs from the food supply 
will result in replacing the IP-TFA from PHOs with other 
macronutrients, most likely other fatty acids, but disagree that the 
safety implications of these changes have not been considered. The 
adverse effect of TFA on LDL-C and other blood lipids and non-lipids 
when replacing other macronutrients (such as carbohydrate, saturated 
fat and cis-unsaturated fat) was extensively demonstrated in controlled 
feeding trials and summarized in regression analyses (Refs. 18, 26, 30, 
31, 32, 33, 44, and 45). In prospective observational studies, 
reduction in CHD risk was also associated with replacement of TFA with 
other macronutrients (Refs. 18 and 49). These analyses, as well as FDA 
estimates discussed previously in section IV, demonstrate that 
replacement of TFA with other macronutrients is expected to result in 
decreased CHD risk.
    We also recognize that replacement of PHOs will result in fatty 
acids from other fats and oils replacing not only IP-TFA but also the 
other fatty acids in the PHOs, but disagree that the safety 
implications of these changes have not been considered. One recent 
study estimated the change in CHD risk from changes in blood lipids due 
to replacing soybean oil PHOs with application specific oils (Ref. 28). 
Results showed that each of the TFA replacement strategies modeled 
changed the fatty acid intake profile in a manner predicted to decrease 
CHD risk, with differences in the projected decreased risk due to 
different replacement oils. Another recent study estimated the effect 
of the replacement of three example PHOs with seven replacement fats 
and oils, based on changes in blood lipids and non-lipids and other 
risk factor biomarkers from controlled feeding trials and on changes in 
CHD risk from prospective observational studies (Ref. 26). Results 
showed that replacement of PHOs with other fats and oils would 
substantially lower CHD risk (Ref. 26). Both studies estimated a 
greater reduction in CHD risk with replacement of PHOs with vegetable 
oils containing higher amounts of cis-unsaturated fatty acids than with 
those high in saturated fat (Refs. 26 and 28). FDA also notes that 
replacement of PHOs containing IP-TFA with other fats and oils over the 
past decade has not been accompanied by extensive increases in 
saturated fat (Ref. 29), which could have diminished the impact of 
removing trans fat.
    The safety implications of replacing IP-TFAs in PHOs with other 
macronutrients and replacing PHOs containing IP-TFAs with other fats 
and oils have been addressed in published studies (Refs. 18, 26, 28, 
30, 31, 32, 33, 44, 45, and 49) and are also addressed in our 
quantitative estimate of decrease in CHD risk with replacement of IP-
TFA, summarized previously in section IV.B (Ref. 25).
4. Alternative Approaches and Evidence for Safety
    In the tentative determination, we requested data to support other 
possible approaches to address the use of PHOs in food, such as setting 
a specification for trans fat levels in food (78 FR 67169 at 67174).
    (Comment 36) Several comments proposed that we should limit the 
percentage of trans fat in finished foods or oils, or set a threshold 
in foods for the maximum grams (g) of trans fat per serving. Some 
comments suggested various specification levels ranging from 0.2 to 0.5 
g trans fat per serving or as a percentage of total fat in foods or 
oils. Another comment urged FDA to establish a reasonable level for 
trans fat in food to specifically account for minor uses of PHOs as 
processing aids.
    Some comments urged us to declare that certain uses of PHOs in 
foods are GRAS, or to issue interim food additive regulations for 
specific low level uses. Examples of such uses provided by comments 
included emulsifiers, encapsulates for flavor agents and color 
additives, pan release agents, anti-caking agents, gum bases, and use 
in frostings, fillings, and coatings. The use of PHOs in chewing gum 
was specifically noted in some comments as deserving special 
consideration due to the claim that there is no meaningful PHO intake 
from this use. Several comments suggested we issue interim food 
additive regulations that would allow certain uses of PHOs in food, 
pending completion of studies evaluating the health effects of low 
level consumption of trans fat that reflect current intake levels. 
Furthermore, one comment advised that if we decide to treat certain 
low-level uses of PHOs as food additives, then the GRAS status for 
these uses should not be revoked until a food additive approval is 
issued.
    In contrast, we also received numerous comments opposed to 
establishing limits of trans fat in foods. Most of these comments noted 
that scientific evidence has shown that no amount of trans fat in food 
is safe and therefore, supported our tentative determination. One 
comment noted that trans fat threshold limits in food would be too 
difficult to monitor and enforce, and therefore, should not be 
established.
    (Response) Regarding the proposals for alternate approaches 
suggesting a threshold for trans fat in food or oils or suggesting that 
FDA declare some uses of PHOs as GRAS, no comments provided evidence 
that any uses of PHOs meet the GRAS standard, or evidence that would 
establish a safe threshold exposure level. Further, although the intake 
from such minor uses may be low, adequate data (e.g., specific 
conditions of use, use level, trans fat content of the PHOs used) were 
not provided so that intake from these uses could be estimated. 
Therefore we are not setting a threshold for trans fat. If industry or 
other interested individuals believe that safe conditions of use for 
PHOs can be demonstrated, it or they may submit a food additive

[[Page 34667]]

petition or food contact notification to FDA for review.
    Interim food additive regulations are appropriate only when there 
is a reasonable certainty that a substance is not harmful. See 21 CFR 
180.1(a). As discussed throughout this section, the available 
scientific evidence raises substantial concerns about the safety of 
PHOs. Based on the currently available data and information, FDA cannot 
conclude that there is a reasonable certainty that PHOs are not 
harmful, nor did any comments provide information that would allow FDA 
to establish conditions of safe use at this time. Therefore, an interim 
food additive regulation would not be appropriate.
    (Comment 37) Several comments suggested various changes to our 
labeling regulations to encourage industry to reformulate products to 
contain less trans fat and help consumers reduce trans fat intake. In 
addition, one comment stated that a 0 g trans fat declaration should 
not be allowed on a label if a PHO is in the ingredient list. Some 
comments indicated that a statement recommending that consumers limit 
their intake of trans fat should be added to the Nutrition Facts Panel. 
A few comments suggested we set a Daily Value for trans fat and 
consider establishing disclosure or disqualifying levels of trans fat 
for nutrient content and health claims. Many comments noted that the 
risk of developing CHD is dependent on many factors, and therefore, the 
association between intake of macronutrients, such as PHOs, and adverse 
health outcomes is best addressed through nutrition labeling and 
consumer education.
    (Response) FDA disagrees that labeling is the best approach to 
address the use of PHOs because FDA has determined that PHOs are not 
GRAS for any use in human food and therefore are food additives subject 
to the requirement of premarket approval under section 409 of the FD&C 
Act. Although we recognize that the requirement to label trans fat 
content led to significant reduction in trans fat levels in products, 
further changes to labeling are outside the scope of this 
determination, which relates to ingredient safety.
    (Comment 38) Some comments suggested that we should work with 
industry to encourage voluntary reductions in PHO use and to foster the 
development of innovative hydrogenation technologies that produce PHOs 
containing low levels of trans fat.
    (Response) FDA disagrees that a voluntary program is the best way 
to remove PHOs from the food supply, given our conclusion on the GRAS 
status of PHOs. FDA has determined that PHOs are not GRAS for any use 
in human food. FDA agrees, however, that we should work with the food 
industry to review new regulatory submissions or data as new 
technologies and/or ingredients are developed that may serve as 
alternatives to PHOs, and we will continue to do so.

V. Citizen Petitions

    As discussed in the tentative determination (78 FR 67169 at 67173), 
we received two citizen petitions regarding the safety of PHOs. In 
2004, the Center for Science in the Public Interest (CSPI) submitted a 
citizen petition (``CSPI citizen petition'' which can be found under 
Docket No. FDA-2004-P-0279) requesting that we revoke the GRAS status 
of PHOs, and consequently declare that PHOs are food additives. The 
petition also asked us to revoke the safe conditions of use for 
partially hydrogenated products that are currently considered food 
additives,\3\ to prohibit the use of partially hydrogenated vegetable 
oils that are prior sanctioned, and to initiate a program to encourage 
manufacturers and restaurants to switch to more healthy oils (CSPI 
citizen petition at pp. 3 through 5, 29 through 30). The CSPI citizen 
petition excluded trans fat that occurs naturally in meat from ruminant 
animals and dairy fats, and that forms during the production of non-
hydrogenated oils (Id. at pp. 2 through 3). It also did not include 
FHOs, which contain negligible amounts of trans fat, and PHOs that may 
be produced by new technologies that result in negligible amounts of 
trans fat in the final product (Id. at p. 3). The CSPI citizen petition 
stated that trans fat promotes CHD by increasing LDL-C and also by 
lowering HDL-C, and therefore has greater adverse effects on serum 
lipids (and possibly CHD) than saturated fats (Id., at pp. 15 through 
18). The CSPI citizen petition also stated that, beyond its adverse 
effects on serum lipids, trans fat may promote heart disease in 
additional ways. Based on these findings, CSPI asserted that PHOs can 
no longer be considered GRAS.
---------------------------------------------------------------------------

    \3\ The petition from CSPI provided, as an example, partially 
hydrogenated methyl ester of rosin, which is approved as a food 
additive for use as a synthetic flavoring substance (32 FR 7946, 
June 2, 1967; 21 CFR 172.515) and as a masticatory substance in 
chewing gum base (29 FR 13894, October 8, 1964; 21 CFR 172.615). 
Partially hydrogenated methyl ester of rosin is not a PHO as 
discussed in section II; accordingly, this this substance is outside 
the scope of this order.
---------------------------------------------------------------------------

    In 2009, Dr. Fred Kummerow submitted a citizen petition (``Kummerow 
citizen petition,'' which can be found at Docket No. FDA-2009-P-0382) 
requesting that we ban partially hydrogenated fat from the American 
diet. The Kummerow citizen petition cited studies linking intake of IP-
TFA to the prevalence of CHD in the United States. The Kummerow citizen 
petition also asserted that trans fat may be passed to infants via 
breast milk and that the daily intake of trans fat related to the 
health of children has been ignored since children do not exhibit overt 
heart disease (Id. at p. 6). The Kummerow citizen petition further 
stated that inflammation in the arteries is believed to be a risk 
factor in CHD and studies have shown that trans fatty acids elicit an 
inflammatory response (Id.).
    This order constitutes a response, in part, to the citizen 
petitions. As discussed above in section III.C (response to Comment 
10), we plan to amend the regulations regarding LEAR and menhaden PHOs 
in a future action, and we will consider taking future action regarding 
related regulations. As discussed in section III.B, we intend to 
address any claims of prior sanction for specific uses of PHO in a 
future action.

VI. Environmental Impact

    We have carefully considered the potential environmental effects of 
this action. We have determined, under 21 CFR 25.32(m), that this 
action ``is of a type that does not individually or cumulatively have a 
significant effect on the human environment'' such that neither an 
environmental assessment nor an environmental impact statement is 
required.
    FDA received some comments on the tentative determination relating 
to potential environmental impacts of removing PHOs from the human food 
supply. We considered these comments in determining whether 
extraordinary circumstances existed under 21 CFR 25.21. Our discussion 
is contained in a review memorandum (Ref. 66).

VII. Economic Analysis

    This notice is not a rulemaking. It is a declaratory order under 5 
U.S.C. 554(e) to terminate a controversy or remove uncertainty. We have 
prepared a memorandum updating our previous estimate published in the 
November 2013 notice, using information available to us as well as 
information we received during the comment period. We estimated the 20-
year costs and benefits of removing PHOs from the U.S. human food 
supply, an outcome that could result from this order (Ref. 17). We 
estimated the costs of all significant effects of the removal, 
including

[[Page 34668]]

packaged food reformulation and relabeling, increased costs for 
substitute ingredients, and consumer, restaurant, and bakery recipe 
changes. We monetized the expected health gains from the removal of 
PHOs from the food supply using information presented in FDA's safety 
assessment (Ref. 17) and the peer-reviewed literature, and added this 
to expected medical expenditure savings to determine the expected 
benefits of this order.
    We estimate the net present value (NPV) (over 20 years; Table 2) of 
quantified costs of this action to be $6.2 billion, with a 90 percent 
confidence interval of $2.8 billion to $11 billion. We estimate the net 
present value of 20 years of benefits to be $140 billion, with a 90 
percent confidence interval of $11 billion to $440 billion. Expected 
NPV of 20 years of net benefits (benefits reduced by quantified costs) 
are $130 billion, with a 90 percent confidence interval of $5 billion 
to $430 billion.

                            Table 2--Costs and Benefits of PHO Removal, USD Billions
----------------------------------------------------------------------------------------------------------------
                  20-Year net present value of                     Low  Estimate       Mean       High  Estimate
----------------------------------------------------------------------------------------------------------------
Costs *.........................................................            $2.8            $6.2             $11
Benefits........................................................              11             140             440
Net Benefits *..................................................               5             130             430
----------------------------------------------------------------------------------------------------------------
* This does not include some unquantified costs, see the economic estimate memo (Ref. 17) for discussion.

VIII. Compliance Date and Related Comments With FDA Responses

    We received numerous comments about the time needed to reformulate 
products to remove PHOs should FDA make a final determination that PHOs 
are not GRAS. We also received comments about challenges to 
reformulation, specific product types that will be difficult to 
reformulate, and effects on small businesses.
    (Comment 39) The comments recommended compliance dates ranging from 
immediate to over 10 years. Several comments stated that fried foods 
should have less time (i.e., 6 months) to phase out the use of PHOs. 
One comment stated that if the use of low levels of PHOs were to remain 
permissible by virtue of being GRAS or through food additive approval, 
then the estimated time to reformulate would be 5 years; however, if 
FDA does not authorize low level uses of PHOs, the timeline would need 
to be 10 years. In general, the food industry urged FDA to provide 
sufficient time for all companies to secure a supply of alternatives 
and transition to new formulations. Some comments stated that FDA 
should coordinate the compliance date with updates to the Nutrition 
Facts Panel.
    Some comments stated that domestically grown oilseed crops must be 
planted about 18 months prior to their expected usage in order for the 
crop to be grown, harvested, stored, crushed, oil extracted, processed, 
refined, delivered, and used in foods. One comment stated that the oil 
industry will need a minimum of 3 years to fully commercialize the 
various oils capable of replacing PHOs in food. A number of comments 
stated that it could take several additional years to reformulate after 
the development of the new oils.
    Several comments expressed concern about adequate availability of 
alternative oils, especially palm oil. One comment stated that the food 
industry would prefer to replace PHOs with domestically produced 
vegetable oils (e.g., high-oleic soybean oil) rather than palm oil, but 
time is needed to commercialize these options. Some comments stated 
that sudden demand for palm oil would pose challenges for obtaining 
sustainably-sourced palm oil, as the current market would likely not be 
able to meet the demand.
    Other comments indicated that the time needed for removal of PHOs 
is dependent on the product category. A number of comments indicated 
that the baking industry will have difficulty replacing the solid 
shortenings used in bakery products. Other comments indicated 
difficulties in the categories of cakes and frostings, fillings for 
candies, chewing gum, snack bars, and as a component of what the 
comments termed minor use ingredients, such as for use in coatings, 
anti-caking agents, encapsulates, emulsifiers, release agents, flavors, 
and colors.
    Several comments indicated that other challenges to PHO removal 
include the need for new transportation infrastructure (e.g., 
terminals, rail cars, barges, and storage facilities), packaging 
changes, and disruption of international trade.
    A number of comments noted challenges faced by small businesses, 
such as access to alternative oils, inability to compete for supply, 
fewer resources to commit to research and development, and effect of 
ingredient costs on growth of the business. Some comments noted that 
small businesses represent a relatively small contribution to overall 
IP-TFA intake. One comment recommended that we allow small businesses 
an additional 2 years beyond the rest of industry. Another comment 
stated that small businesses would need at least 5 years due to their 
limitations in research and development expertise, inability to command 
supply of scarce ingredients, and economic pressures of labeling 
changes. A related comment requested that FDA take into consideration 
the magnitude of private label products impacted. Other comments stated 
that small businesses should not be given special consideration or 
longer times for implementation.
    (Response) Based on our experience and on the changes we have 
already seen in the market, we believe that 3 years is sufficient time 
for submission and review and, if applicable requirements are met, 
approval of food additive petitions for uses of PHOs for which industry 
or other interested individuals believe that safe conditions of use may 
be prescribed. For this reason, we are establishing a compliance date 
for this order of June 18, 2018. We recognize that the use of PHOs in 
the food supply is already declining and expect this to continue even 
prior to the compliance date. Regarding the use of ``low levels'' of 
PHOs, no comments provided a basis upon which we can currently conclude 
that any use of PHO is GRAS (discussed in section IV). We recognize the 
challenges faced by small businesses, however, considering our 
determination that PHOs are not GRAS for any use in human food, we 
conclude that providing 3 years for submission and review of food 
additive petitions and/or food contact notifications is reasonable, and 
will have the additional benefit of allowing small businesses time to 
address these challenges. We understand the difficulties faced by small 
businesses due to limited research and development resources and

[[Page 34669]]

potential challenges to gain timely access to suitable alternatives.
    The compliance date will have the additional benefit of minimizing 
market disruptions by providing industry sufficient time to identify 
suitable replacement ingredients for PHOs, to exhaust existing product 
inventories, and to reformulate and modify labeling of affected 
products. Three years also provides time for the growing, harvesting, 
and processing of new varieties of edible oilseeds to meet the expected 
demands for alternative oil products and to address the supply chain 
issues associated with transition to new oils.
    (Comment 40) Several comments stated that how FDA defines PHOs and 
FHOs will affect reformulation efforts and the time needed to 
reformulate. These comments suggested it was unclear from the tentative 
determination whether FHOs would be subject to this final 
determination.
    (Response) As discussed in section II, we have defined PHOs, the 
subjects of this order, as fats and oils that have been hydrogenated, 
but not to complete or near complete saturation, and with an IV greater 
than 4 as determined by an appropriate method. We have also defined 
FHOs as those fats and oils that have been hydrogenated to complete or 
near complete saturation, and with an IV of 4 or less, as determined by 
an appropriate method. Thus, FHOs are outside the scope of this order 
and there is no need to allow additional time for reformulation of 
products containing FHO.

IX. Conclusion and Order

    As discussed in this document, for a substance to be GRAS, there 
must be consensus among qualified experts based on generally available 
information that the substance is safe under the intended conditions of 
use. In accordance with the process set forth in FDA's regulations in 
Sec.  170.38, FDA has determined that there is no longer a consensus 
that PHOs, the primary source of industrially-produced trans fat, are 
generally recognized as safe for use in human food, based on current 
scientific evidence discussed in section IV.B regarding the health 
risks associated with consumption of trans fat. FDA considers this 
order a partial response to the citizen petitions from CSPI and Dr. 
Kummerow.

X. References

    The following references have been placed on display in the 
Division of Dockets Management (see ADDRESSES) and may be seen by 
interested persons between 9 a.m. and 4 p.m., Monday through Friday, 
and are available electronically at http://www.regulations.gov. (FDA 
has verified the Web site addresses in this reference section, but we 
are not responsible for any subsequent changes to the Web sites after 
this document publishes in the (Federal Register.)

1. Tarrago-Trani, M., K. M. Philips, L. E. Lemar, et al.,``New and 
Existing Oils and Fats Used in Products With Reduced Trans-Fatty 
Acid Content,'' Journal of the American Dietetic Association, 
106:867-877, 2006.
2. Kodali, D. R. and G. R. List, Eds., Trans Fats Alternatives, AOCS 
Press, Champaign, IL, p. 34-35, 2005.
3. USDA National Nutrition Database for Standard Reference, Release 
23, 2010; http://www.ars.usda.gov/Services/docs.htm?docid=8964.
4. Kodali, D. R. and G. R. List, Eds., Trans Fats Alternatives, AOCS 
Press, Champaign, IL, pp. 4, 2005.
5. Memorandum from J. Park to M. Honigfort, August 10, 2005.
6. Memorandum from J. Park to M. Honigfort, August 19, 2010.
7. IOM/NAS, ``Dietary Reference Intakes for Energy Carbohydrate, 
Fat, Fatty Acids, Cholesterol, and Amino Acids (Macronutrients),'' 
chapters 8 and 11, National Academies Press, Washington DC, 2002/
2005; http://www.nap.edu.
8. American Heart Association, http://www.heart.org/HEARTORG/GettingHealthy/FatsAndOils/Fats101/Trans-Fats_UCM_301120_Article.jsp.
9. Eckel, R.H., S. Borra, A.H. Lichtenstein, et al., ``Understanding 
the Complexity of Trans Fatty Acid Reduction in the American Diet,'' 
Circulation, 115:2220-2235, 2007.
10. Kris-Etherton, P. M., S. Innis, ``Position of the American 
Dietetic Association and Dietitians of Canada: Dietary Fatty 
Acids,'' Journal of the American Dietetic Association, pp. 1599-
1611, 2007.
11. WHO, ``Diet, Nutrition, and the Prevention of Chronic Disease,'' 
Technical Series Report 916, pp. 81-85, Geneva, 2003.
12. USDA and Department of Health and Human Services (HHS), Dietary 
Guidelines for Americans, 2005, 6th ed., pp. 29-34, Washington, DC: 
U.S. Government Printing Office, January 2005.
13. USDA and HHS, Dietary Guidelines for Americans, 2010, 7th ed., 
pp. 24-27, Washington, DC: U.S. Government Printing Office, December 
2010.
14. HHS/FDA/Center for Food Safety and Applied Nutrition Food 
Advisory Committee, Nutrition Subcommittee Meeting, Total Fat and 
Trans Fat, April 27-28, 2004.
15. Dietz, W. H. and K. S. Scanlon, ``Eliminating the Use of 
Partially Hydrogenated Oil in Food Production and Preparation,'' 
Journal of the American Medical Association, 108:143-144, 2012.
16. Memorandum from D. Doell, D. Folmer, and H. Lee to M. Honigfort, 
June 11, 2015.
17. Memorandum from R. Bruns to M. Honigfort, June 11, 2015.
18. Memorandum from J. Park to M. Honigfort, Scientific Update on 
Experimental and Observational Studies of Trans Fat Intake and 
Coronary Heart Disease Risk, June 11, 2015.
19. IOM, ``Evaluation of Biomarkers and Surrogate Endpoints in 
Chronic Disease'', Washington, DC: National Academies Press, 2010.
20. Rasnake, C. M., P. R. Trumbo, and T. M. Heinonen, ``Surrogate 
Endpoints and Emerging Surrogate Endpoints for Risk Reduction of 
Cardiovascular Disease,'' Nutrition Reviews, 66:76-81, 2008.
21. Baigent C., A. Keech, P. M. Kearney, et al., ``Efficacy and 
Safety of Cholesterol-lowering Treatment: Prospective Meta-analysis 
of Data from 90,056 Participants in 14 Randomised Trials of 
Statins,'' Lancet, 366:1267-1278, 2005.
22. Lewington S., G. Whitlock, R. Clarke, et al., ``Blood 
Cholesterol and Vascular Mortality by Age, Sex, and Blood Pressure: 
A Meta-analysis of Individual Data from 61 Prospective Studies with 
55,000 Vascular Deaths,'' Lancet, 370:1829-1839, 2007.
23. Memorandum from J. Park to M. Honigfort, Literature Review, June 
11, 2015.
24. Kiage J. N., P. D. Merrill, S. E. Judd, et al., ``Intake of 
Trans Fat and Incidence of Stroke in the Reasons for Geographic And 
Racial Differences in Stroke (REGARDS) Cohort,'' American Journal of 
Clinical Nutrition, 99:1071-1076, 2014.
25. Memorandum from J. Park to M. Honigfort, Quantitative Estimate 
of Industrial Trans Fat Intake and Coronary Heart Disease Risk, June 
11, 2015.
26. Mozaffarian, D. and R. Clarke, ``Quantitative Effects on 
Cardiovascular Risk Factors and Coronary Heart Disease Risk of 
Replacing Partially Hydrogenated Vegetable Oils With Other Fats and 
Oils,'' European Journal of Clinical Nutrition, 63:S22-S33, 2009.
27. Go A. S., Mozaffarian, D., Roger, V. L., et al., on behalf of 
the American Heart Association Statistics Committee and Stroke 
Statistics Subcommittee, ``Executive Summary: Heart Disease and 
Stroke Statistics--2014 Update: A Report from the American Heart 
Association,'' Circulation, 129:399-410, 2014.
28. Lefevre, M., R. P. Mensink, P. M. Kris-Etherton, et al., 
``Predicted Changes in Fatty Acid Intakes, Plasma Lipids, and 
Cardiovascular Disease Risk Following Replacement of Trans Fatty 
Acid-Containing Soybean Oil with Application-Appropriate 
Alternatives,'' Lipds, 47:951-962, 2012.
29. Mozaffarian, D., M. F. Jacobson, J. S. Greenstein, ``Food 
Reformulations to Reduce Trans Fatty Acids'' [Letter to the editor], 
New England Journal of Medicine, 362:2037-2039, 2010.
30. Brouwer, I. A., A. J. Wanders, M. B. Katan, ``Effect of Animal 
and Industrial Trans Fatty Acids on HDL and LDL Cholesterol Levels 
in Humans--A

[[Page 34670]]

Quantitative Review,'' PLoS One, 5(3):e9434, 2010.
31. Ascherio, A., M. B. Katan, P. L. Zock, et al., ``Trans Fatty 
Acids and Coronary Heart Disease,'' New England Journal of Medicine, 
340:1994-1998, 1999.
32. Zock, P. L., M. B. Katan, and R. P. Mensink, ``Dietary Trans 
Fatty Acids and Lipoprotein Cholesterol'' [Letter to the editor], 
American Journal of Clinical Nutrition, 61:617, 1995.
33. Mensink, R. P., P. L. Zock, A. D. Kester, et al., ``Effects of 
Dietary Fatty Acids and Carbohydrates on the Ratio of Serum Total to 
HDL Cholesterol and on Serum Lipids and Apolipoproteins: A Meta-
Analysis of 60 Controlled Trials,'' American Journal of Clinical 
Nutrition, 77:1146-1155, 2003.
34. Mozaffarian, D., M. B., Katan, A. Asherio, et al., ``Trans Fatty 
Acids and Cardiovascular Disease,'' New England Journal of Medicine, 
354:1601-1613, 2006.
35. Dourson, M. ``Mode of Action and Dose-Response Evaluation of the 
Effect of Partially Hydrogenated Oils on LDL-Cholesterol,'' 
Presented at the SOT FDA Colloquia on Emerging Toxicological Science 
Challenges in Food and Ingredient Safety, November 7, 2014.
36. National Research Council, Science and Decisions: Advancing Risk 
Assessment, National Academies Press, Washington, DC, 2009; http://www.nap.edu).
37. Keys, A., J. T. Anderson, F. Grande, ``Serum Cholesterol 
Response to Changes in the Diet: I. Iodine Value of Dietary Fat 
Versus 2S-P,'' Metabolism, 14:747-758, 1965.
38. Hegsted, D. M., R. B. McGandy, M. L. Myers, et al., 
``Quantitative Effects of Dietary Fat on Serum Cholesterol in Man,'' 
American Journal of Clinical Nutrition, 17:281-295, 1965.
39. Keys, A. and R. W. Parlin, ``Serum Cholesterol Response to 
Changes in Dietary Lipids,'' American Journal of Clinical Nutrition, 
19:175-181, 1966.
40. Page, I. H., E. V. Allen, F. L. Chamberlain, et al., ``Dietary 
Fat and Its Relation to Heart Attacks and Strokes,'' Circulation, 
23:133-136, 1961.
41. Clarke, R., C. Frost, R. Collins, et al. ``Dietary Lipids and 
Blood Cholesterol: Quantitative Meta-Analysis of Metabolic Ward 
Studies,'' BMJ, 314:112-117, 1997.
42. Mensink R. P. and M. B. Katan, ``Effect of Dietary Fatty Acids 
on Serum Lipids and Lipoproteins. A Meta-Analysis of 27 Trials,'' 
Arteriosclerosis, Thrombosis, and Vascular Biology, 12:911-919, 
1992.
43. Reeves, R. M., ``Effect of Dietary Trans Fatty Acids on 
Cholesterol Levels'' [Letter to the editor], New England Journal of 
Medicine, 324:338-339, 1991.
44. Katan M. B., P. L. Zock, R. P. Mensink, ``Trans Fatty Acids and 
their Effects on Lipoproteins in Humans,'' Annual Review of 
Nutrition, 15:473-493, 1995.
45. Zock P. L. and R. P. Mensink, ``Dietary Trans-Fatty Acids and 
Serum Lipoproteins in Humans,'' Current Opinion in Lipidology, 7:34-
37, 1996.
46. Oh, K., F. B. Hu, J. E. Manson, et al., ``Dietary Fat Intake and 
Risk of Coronary Heart Disease in Women: 20 Years of Follow-up of 
the Nurses' Health Study,'' American Journal of Epidemiology, 
161:672-679, 2005.
47. Ascherio A., E. B. Rimm, E. L. Giovannucci, et al., ``Dietary 
Fat and Risk of Coronary Heart Disease in Men: Cohort Follow Up 
Study in the United States,'' BMJ, 313:84-90, 1996.
48. Willett W. C., M. J. Stampfer, J. E. Manson, et al., ``Intake of 
Trans Fatty Acids and Risk of Coronary Heart Disease Among Women,'' 
Lancet, 341:581-585, 1993.
49. Hu, F. B., M. J. Stampfer, J. E. Manson, et al., ``Dietary Fat 
Intake and the Risk of Coronary Heart Disease in Women,'' New 
England Journal of Medicine, 337:1491-1499, 1997.
50. Pietinen, P., A. Ascherio, P. Korhonen, et al., ``Intake of 
Fatty Acids and Risk of Coronary Heart Disease in a Cohort of 
Finnish Men. The Alpha-Tocopherol, Beta-Carotene Cancer Prevention 
Study,'' American Journal of Epidemiology, 145:876-887, 1997.
51. Oomen, C. M., M. C. Ocke, E. J. Feskens, et al., ``Association 
Between Trans Fatty Acid Intake and 10-year Risk of Coronary Heart 
Disease in the Zutphen Elderly Study: A Prospective Population-Based 
Study,'' Lancet, 357:746-751, 2001.
52. Mozaffarian D., A. Aro, W. C. Willett, ``Health Effects of 
Trans-fatty Acids: Experimental and Observational Evidence,'' 
European Journal of Clinical Nutrition, 63:S5-21, 2009.
53. Skeaff C. M. and J. Miller, ``Dietary Fat and Coronary Heart 
Disease: Summary of Evidence from Prospective Cohort and Randomised 
Controlled Trials,'' Annals of Nutrition & Metabolism, 55:173-201, 
2009.
54. Bendsen N. T., R. Christensen, E. M. Bartels, et al., 
``Consumption of Industrial and Ruminant Trans Fatty Acids and Risk 
of Coronary Heart Disease: A Systematic Review and Meta-analysis of 
Cohort Studies,'' European Journal of Clinical Nutrition, 65:773-
783, 2011.
55. Chowdhury R., S. Warnakula, S. Kunutsor, et al., ``Association 
of Dietary, Circulating, and Supplement Fatty Acids with Coronary 
Risk. A Systematic Review and Meta-analysis,'' Annals of Internal 
Medicine, 160:398-406, 2014.
56. Trumbo, P. R. and T. Shimakawa, ``Tolerable Upper Intake Levels 
for Trans Fat, Saturated Fat, and Cholesterol,'' Nutrition Reviews, 
69:270-278, 2011.
57. Willett W. C., ``Dietary Fats and Coronary Heart Disease,'' 
Journal of Internal Medicine, 272:13-24, 2012.
58. National Heart, Lung, and Blood Institute, ``Lifestyle 
Interventions to Reduce Cardiovascular Risk: Systematic Evidence 
Review From the Lifestyle Work Group,'' Bethesda, MD: HHS, National 
Institutes of Health, 2013 (http://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk-reduction/lifestyle).
59. Burlingame B., C. Nishida, R. Uauy, et al., ``Fats and Fatty 
Acids in Human Nutrition: Introduction,'' Annals of Nutrition & 
Metabolism, 55:5-7, 2009.
60. Uauy, R., A. Aro, R. Clarke, et al., ``WHO Scientific Update on 
Trans Fatty Acids: Summary and Conclusions,'' European Journal of 
Clinical Nutrition, 63: S68-S75, 2009.
61. Food and Agricultural Organization of the United Nations (FAO) 
and WHO, Fats and Fatty Acids in Human Nutrition. Report of an 
Expert Consultation. Rome: FAO; 2010.
62. Eckel, R. H., J. M. Jakicic, J. D. Ard, et al. ``2013 AHA/ACC 
Guideline on Lifestyle Management to Reduce Cardiovascular Risk: A 
Report of the American College of Cardiology/American Heart 
Association Task Force on Practice Guidelines,'' Circulation, 
129:S76-S99, 2014.
63. European Food Safety Authority (EFSA), ``Opinion of the 
Scientific Panel on Dietetic Products, Nutrition and Allergies on a 
Request from the Commission Related to the Presence of Trans Fatty 
Acids in Foods and the Effect on Human Health of the Consumption of 
Trans Fatty Acids,'' EFSA Journal, 81:1-49, 2004.
64. Dietary Guidelines Advisory Committee, Report of the Dietary 
Guidelines Advisory Committee on Dietary Guidelines for Americans, 
2005, Washington, DC: HHS, 2005; http://www.health.gov/dietaryguidelines/dga2005/report/default.htm.
65. Dietary Guidelines Advisory Committee, Report of the Dietary 
Guidelines Advisory Committee on Dietary Guidelines for Americans, 
2010, Washington, DC: USDA, Agricultural Research Service, 2010; 
http://origin.www.cnpp.usda.gov/DGAs2010-DGACReport.htm.
66. Memorandum from M. Pfeil to M. Honigfort, June 11, 2015.

    Dated: June 12, 2015.
Leslie Kux,
Associate Commissioner for Policy.
[FR Doc. 2015-14883 Filed 6-16-15; 8:45 am]
 BILLING CODE 4164-01-P


