Sandoz Inc. v. Xavier Becerra

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued November 7, 2022 Decided January 10, 2023

No. 22-5202

SANDOZ INC.,
APPELLANT

v.

XAVIER BECERRA, IN HIS OFFICIAL CAPACITY AS SECRETARY
OF HEALTH AND HUMAN SERVICES, ET AL.,
APPELLEES

Appeal from the United States District Court
for the District of Columbia
(No. 1:21-cv-00600)

William M. Jay argued the cause for appellant. With him
on the briefs were Brian T. Burgess and Gerard J. Cedrone.
Michael D. Shumsky entered an appearance.

Steven H. Hazel, Attorney, U.S. Department of Justice,
argued the cause for appellees. With him on the brief were
Brian M. Boynton, Principal Deputy Assistant Attorney
General, Daniel Tenny, Attorney, Samuel Bagenstos, General
Counsel, U.S. Department of Health and Human Services, and
Leslie R. Cohen, Associate Chief Counsel, U.S. Food and Drug
Administration.
 2
Before: MILLETT, PILLARD and RAO, Circuit Judges.

Opinion for the Court filed by Circuit Judge RAO.

RAO, Circuit Judge: Under the Hatch-Waxman Act, a drug
may receive “new chemical entity exclusivity” if no active
ingredient in the drug was previously “approved.” The drug
Aubagio was awarded this exclusivity because the Food &
Drug Administration (“FDA”) determined that Aubagio’s only
active ingredient, teriflunomide, had never previously been
approved. This case concerns a challenge to Aubagio’s
exclusivity period, which Sandoz Inc. raises to secure a solo
period of marketing exclusivity for its generic equivalent.
Sandoz maintains that teriflunomide was previously
“approved” as an impurity in the drug Arava. In the alternative,
Sandoz argues that teriflunomide was in fact approved as an
active ingredient in Arava. The district court granted summary
judgment for the FDA, agreeing with the agency that Aubagio
was entitled to exclusivity because teriflunomide had never
previously been approved.

We affirm the judgment of the district court. Although
Sandoz did not exhaust its statutory argument before the FDA,
in the absence of a statutory or regulatory exhaustion
requirement, we find it appropriate to decide Sandoz’s
challenge. When the FDA approves a new drug, it does not also
“approve” known impurities in that drug for the purpose of new
chemical entity exclusivity. And the record is clear the FDA
did not approve teriflunomide as an active ingredient when it
approved Arava. Aubagio was therefore entitled to new
chemical entity exclusivity, and Sandoz cannot benefit from a
solo exclusivity period for its generic equivalent.
 3
I.

A.

Before a new drug may be sold or marketed in the United
States, it must be approved by the FDA. Typically, drug
sponsors submit a “new drug application,” or “NDA,” to the
agency under section 505(b) of the Federal Food, Drug, and
Cosmetic Act (“FDCA”). See FDCA, Pub. L. No. 75-717,
§ 505(b), 52 Stat. 1040, 1052 (1938) (codified as amended at
21 U.S.C. § 355(b)). A new drug applicant must submit reports
of trials showing the drug is both safe and effective. See 21
U.S.C. § 355(b)(1).

In the Hatch-Waxman Act, Congress amended the FDCA
to provide a more streamlined path for the approval of generic
drugs. Drug Price Competition and Patent Term Restoration
Act of 1984, Pub. L. No. 98-417, tit. I, § 101, 98 Stat. 1585,
1585 (codified as amended at 21 U.S.C. § 355(j)). Generic
drugs are those that “contain[] the same active ingredients but
not necessarily the same excipients as a so-called ‘pioneer
drug’ that is marketed under a brand name.” United States v.
Generix Drug Corp., 460 U.S. 453, 454–55 (1983). Instead of
following the rigorous NDA process, generic manufacturers
may file an “abbreviated new drug application,” or “ANDA.”
In the abbreviated process, a manufacturer may bypass the
safety and efficacy demonstrations required for a new drug by
showing its generic drug contains the same active ingredient as
a previously approved drug and is equivalent in other respects.
See 21 U.S.C. § 355(j)(2)(A).

The Hatch-Waxman Act made it easier to get approval for
generics, but also preserved incentives for the research and
innovation necessary for the development of new drugs. See
Abbott Labs. v. Young, 920 F.2d 984, 985 (D.C. Cir. 1990).
Manufacturers who developed novel drugs would enjoy
 4
specified terms of marketing exclusivity. The most generous
such term—and the one at issue in this case—is known as “new
chemical entity exclusivity,” which applies when “an
application submitted under subsection (b) of this section for a
drug, no active ingredient … of which has been approved in
any other application under subsection (b) of this section, is
approved.” 1 21 U.S.C. § 355(j)(5)(F)(ii). Applications
“submitted under subsection (b) of this section” are NDAs. In
other words, new chemical entity exclusivity applies when a
new drug is approved and no active ingredient in that drug was
previously “approved” in a different drug.

This case implicates a four-year term of new chemical
entity exclusivity because Sandoz filed an ANDA that
contained a “paragraph IV certification,” which challenges a
patent of a previously approved drug. 2 The first generic
manufacturer to file such an application enjoys a 180-day

1
Since the events at issue in this case, the FDCA has been amended
to use the phrase “active moiety” rather than “active ingredient.” See
Act to Amend the Federal Food, Drug, and Cosmetic Act, Pub. L.
No. 117-9, § 1(a)(1)(B), 135 Stat. 256, 256 (2021). The parties agree
the distinction is not material in this case. This opinion refers to
“active ingredient,” the statutory term at the time of the agency
decisions under review, and to the version of section 505 in force on
September 12, 2012, the date the FDA awarded new chemical entity
exclusivity to Aubagio.
2
If a manufacturer wishes to market a generic drug before one of the
patents on the underlying drug expires, it must certify “that such
patent is invalid or will not be infringed by the manufacture, use, or
sale of the new drug for which the application is submitted.” 21
U.S.C. § 355(j)(2)(A)(vii)(IV). This is the “paragraph IV
certification” involved here. A new chemical entity will receive five
years of exclusivity against generic competitors that do not challenge
any of the patents associated with the drug. See id. § 355(j)(5)(F)(ii).
 5
period of exclusivity. 3 See 21 U.S.C. §§ 355(j)(5)(B)(iv)(I),
(II)(bb). Under the new chemical entity regime, ANDAs may
not be submitted for generics before the end of the four-year
exclusivity period. Id. § 355(j)(5)(F)(ii). Since multiple
applicants often submit ANDAs on exactly that date, the 180-
day period of marketing exclusivity will often be shared among
generic manufacturers who all qualify as “first applicant[s].”
See id. § 355(j)(5)(B)(iv).

B.

Sandoz, a generic manufacturer, sought approval for a
generic corresponding to the drug Aubagio. Aubagio has
teriflunomide as its sole active ingredient and is used to treat
patients with relapsing multiple sclerosis. When the FDA
approved Aubagio on September 12, 2012, it determined that
it had not previously approved teriflunomide in any other drug
and that Aubagio was eligible for new chemical entity
exclusivity. Accordingly, generic manufacturers could not
submit ANDAs corresponding to Aubagio until September 12,
2016. But Sandoz sought to challenge the exclusivity period
and so moved sooner. It submitted a letter to the FDA on
August 31, 2016, arguing that the agency had previously
approved teriflunomide and that Aubagio was therefore
ineligible for new chemical entity exclusivity. 4

3
Because a paragraph IV certification qualifies as an act of patent
infringement, see 35 U.S.C. § 271(e)(2)(A), this exclusivity
provision serves to “compensate manufacturers for research and
development costs as well as the risk of litigation from patent
holders,” Teva Pharms., USA, Inc. v. Leavitt, 548 F.3d 103, 104
(D.C. Cir. 2008).
4
Under Sandoz’s view, Aubagio may have been eligible for a lesser
three-year term of exclusivity that applies when a new drug uses the
 6
Sandoz’s letter focused on a rheumatoid arthritis drug
known as Arava, which the FDA approved in 1998. Arava uses
leflunomide, which the FDA identified as the sole active
ingredient in the drug. When Arava is manufactured and stored,
however, some of the leflunomide molecules break down into
teriflunomide, a similar compound. The FDA characterized
teriflunomide as an “impurity” and allowed Arava to contain
up to 3.5 percent teriflunomide. In its letter, Sandoz argued that
the small quantities of teriflunomide that build up in Arava
contribute to the functioning of the drug, and that teriflunomide
was “physically present as a bioavailable and
physiologically/pharmacologically active component” of
Arava. Sandoz contended that the FDA had “approved”
teriflunomide when it approved Arava and therefore that the
FDA should rescind new chemical entity exclusivity for
Aubagio, which used teriflunomide as its active ingredient.

While the FDA considered Sandoz’s request to rescind
Aubagio’s exclusivity, Sandoz submitted two ANDAs for its
generic teriflunomide product, each with a paragraph IV
certification: one on September 7, 2016, in advance of the four-
year deadline, and one on September 12, 2016, the day the
deadline expired.

Whether Sandoz would enjoy sole exclusivity, or would
share exclusivity with other manufacturers, hinged on the
FDA’s determination of whether Aubagio was properly
classified a new chemical entity. Because Sandoz was the only
generic manufacturer to file a substantially complete ANDA
prior to the four-year deadline, it would qualify as the only first
applicant should the FDA rescind Aubagio’s four-year
exclusivity term. Sandoz would thus enjoy 180 days of sole

active ingredient in a previously approved drug but repurposes it for
a new use. See 21 U.S.C. § 355(j)(5)(F)(iii).
 7
exclusivity. Conversely, if the FDA rejected Sandoz’s request,
the first ANDA would be premature and Sandoz would share
first applicant status with some twenty generic applicants who
all filed on September 12, 2016, obviously limiting the value
of the exclusivity period. 5

In June 2018, the FDA rejected Sandoz’s request to
rescind Aubagio’s new chemical entity exclusivity. Under the
agency’s longstanding interpretation of the exclusivity
provisions, an ingredient is “approved” in a new drug
application only if it was an active ingredient in that drug. The
agency concluded it had recognized teriflunomide simply as an
impurity in Arava, not as an active ingredient. As the FDA
explained, leflunomide was the sole active ingredient in Arava.
The safety and efficacy studies had focused on leflunomide,
and nothing in Arava’s application indicated that the presence
of teriflunomide as a degradant was designed to serve a
therapeutic function. Because Aubagio was properly granted a
four-year period of exclusivity, the FDA rejected Sandoz’s first
ANDA as premature.

In an administrative appeal, Sandoz argued that Arava was
a “combination drug product containing both leflunomide and
teriflunomide” as active ingredients. The FDA denied this
appeal on February 12, 2021, affirming its prior reasoning and
again concluding that leflunomide was the only active
ingredient approved in Arava.

5
While Aubagio’s term of exclusivity has since expired, the validity
of that term remains relevant because no manufacturer has marketed
a generic corresponding to Aubagio and triggered the 180-day
exclusivity provision. Sandoz explains that this is because Aubagio’s
sponsor, Sanofi-Aventis, has granted licenses for generic
manufacturers to enter the U.S. market on March 12, 2023. This
opinion has been prepared on an expedited basis to reflect that date.
 8
C.

Sandoz filed suit under the Administrative Procedure Act
in the U.S. District Court for the District of Columbia, alleging
the FDA’s decision to maintain new chemical entity exclusivity
for Aubagio was “not in accordance with law.” 5 U.S.C.
§ 706(2)(A). 6

Sandoz challenged the FDA’s interpretation of the
exclusivity provisions, arguing the FDA had misinterpreted the
FDCA in determining that an ingredient was “approved” only
if it was the active ingredient in some previously approved
drug. Sandoz emphasized the grammar of the exclusivity
provision, and its reference to the approval of “an application
submitted under subsection (b) of this section for a drug, no
active ingredient … of which has been approved in any other
application under subsection (b) of this section.” 21 U.S.C.
§ 355(j)(5)(F)(ii). According to Sandoz, this phrase requires
the FDA to determine whether the active ingredient in the new
drug is the same as any known substance, whether an active
ingredient or an impurity, in a previously approved drug. Even
if the FDA correctly concluded teriflunomide was only an
impurity in Arava, on Sandoz’s understanding, the FDA
necessarily approved that impurity, and Aubagio is therefore
ineligible for new chemical entity exclusivity. In the
alternative, Sandoz maintained that teriflunomide was present
as an active ingredient in Arava because its presence as a
degradant was therapeutically useful.

The district court granted summary judgment for the
Government. As to Sandoz’s statutory argument, the district
court determined the phrase “no active ingredient … of which

6
Sandoz also argued the FDA’s decision was arbitrary and
capricious, but it does not press that argument on appeal.
 9
has been approved” was most naturally read to mean “approved
as an active ingredient” and not merely as an impurity. The
district court agreed with the FDA that teriflunomide was not
an active ingredient in Arava, and that the agency had not
approved it as such. Sandoz timely appealed the district court’s
grant of summary judgment, which we review de novo. See
Rempfer v. Sharfstein, 583 F.3d 860, 864–65 (D.C. Cir. 2009).

II.

We begin with Sandoz’s statutory argument that when the
FDA approves a new drug it “approve[s]” known impurities in
that drug. Under this reasoning, Aubagio was not entitled to
exclusivity because its active ingredient was teriflunomide,
which already had “been approved” in Arava as an impurity.
Although Sandoz did not explicitly raise this issue before the
FDA, in the absence of any statutory or regulatory exhaustion
requirement, we find it appropriate to decide Sandoz’s
arguments. On the merits, we uphold the FDA’s interpretation
because it is consistent with the text and structure of the FDCA.

A.

The Government contends that Sandoz failed to exhaust its
statutory challenge before the FDA. Although the district court
did not address exhaustion, we address it here, because the
Government is correct that Sandoz did not raise this argument
before the agency.

In its letters to the FDA, Sandoz did not contest the
agency’s well-established interpretation that a new drug that
utilizes an ingredient previously identified only as an impurity
may be eligible for new chemical entity exclusivity. Sandoz
simply argued that teriflunomide had been approved as an
active ingredient, not merely an impurity, in Arava. Shortly
before the FDA issued its final decision on Sandoz’s
 10
administrative appeal, Sandoz shared with the agency a draft
complaint that we are told resembles the one ultimately filed in
the district court. But Sandoz did not explicitly ask the agency
to revisit its statutory conclusion, nor did it request that the
FDA take time to review the arguments raised in the complaint.
In district court, Sandoz argued for the first time that the FDA
had erroneously interpreted the statute. We agree with the
Government that simply sharing a draft complaint at the very
end of a lengthy administrative process—without even
requesting the agency consider the arguments raised in that
complaint—does not “reasonably flag[]” an issue “for the
agency’s consideration.” NTCH, Inc. v. FCC, 841 F.3d 497,
508 (D.C. Cir. 2016) (cleaned up).

In this case, however, the Government identifies no
applicable statutory or regulatory exhaustion requirement, and
we are not aware of one. As the Supreme Court has explained,
“requirements of administrative issue exhaustion are largely
creatures of statute,” and the Court’s “cases addressing issue
exhaustion reflect this fact.” Sims v. Apfel, 530 U.S. 103, 107–
08 (2000). In some circumstances, courts may apply a
“judicially imposed issue-exhaustion requirement.” Id. at 108.
But in the absence of a legal exhaustion requirement, the
“administrative-waiver doctrine does not represent an ironclad
rule.” Advocs. for Highway & Auto Safety v. Fed. Motor
Carrier Safety Admin., 429 F.3d 1136, 1148 (D.C. Cir. 2005).

This court has recognized that as a prudential matter we
may “exercise our discretion to address” an issue not exhausted
before the agency. R.R. Yardmasters of Am. v. Harris, 721 F.2d
1332, 1337 (D.C. Cir. 1983); see also Okla. Dep’t of Env’t
Quality v. EPA, 740 F.3d 185, 192 (D.C. Cir. 2014)
(considering on the merits an issue not raised before the
Environmental Protection Agency). When determining
whether to exercise such discretion, we must heed the Supreme
 11
Court’s caution that courts should not apply judicial issue
exhaustion without considering the particulars of an
administrative scheme. Prudential issue exhaustion arises out
of “an analogy to the rule that appellate courts will not consider
arguments not raised before trial courts,” and we must not
“reflexively assimilat[e] the relation of administrative bodies
and the courts to the relationship between lower and upper
courts.” Sims, 530 U.S. at 108–09, 110 (cleaned up). Recently,
in Carr v. Saul, the Supreme Court reiterated that whether a
court should impose an issue exhaustion requirement “depends
on the degree to which the analogy to normal adversarial
litigation applies in a particular administrative proceeding.”
141 S. Ct. 1352, 1358 (2021) (quoting Sims, 530 U.S. at 109).

In Sims and Carr, the Supreme Court declined to impose
exhaustion in the context of highly structured procedures
adopted by the Social Security Administration involving
review before administrative adjudicators. See Carr, 141 S. Ct.
at 1356 (involving a hearing before an administrative law
judge); Sims, 530 U.S. at 105 (involving review of an
administrative law judge’s decision by the Social Security
Appeals Council); 20 C.F.R § 404.900 (explaining the six-step
administrative review process at issue in both cases). The Court
declined to impose issue exhaustion because it concluded that
these procedures were not similar enough to adjudication in
lower courts. More generally, the Court has noted that
determining whether to require administrative exhaustion is
“intensely practical” and turns on “both the nature of the claim
presented and the characteristics of the particular
administrative procedure provided.” McCarthy v. Madigan,
503 U.S. 140, 146 (1992) (cleaned up) (discussing these
principles in the context of administrative remedies).

Applying these considerations, we see no reason to treat
Sandoz’s statutory argument as forfeited. First, the proceedings
 12
here were far from adversarial. Sandoz submitted letters to the
FDA challenging the new chemical entity exclusivity that had
been awarded years earlier to Arava. While Sanofi-Aventis,
Aubagio’s sponsor, also sent a letter to the FDA defending its
exclusivity, the letters were simply considered under the
FDA’s internal review procedures, which are governed by
nonbinding guidance documents. Formal Dispute Resolution:
Sponsor Appeals Above the Division Level at 1–2,
https://www.fda.gov/media/126910/download (Nov. 2017)
(providing procedures for internal agency appeals that “do not
establish legally enforceable responsibilities”). The FDA’s
review involved no adjudicator, but rather determinations by
counsel in the Office of Regulatory Policy and by the Director
for the Office of Generic Drugs.

The FDA’s informal process for reviewing such letters
before agency officials cannot be analogized to “normal
adversarial litigation,” involving “a litigant opposing the
claimant.” Carr, 141 S. Ct. at 1358, 1359 (cleaned up). The
FDA’s procedures are entirely distinct from those cases in
which the Supreme Court has found issue exhaustion
appropriate, all of which involved a quasi-judicial process
conducted between adverse parties before a neutral examiner. 7

7
For instance, Hormel v. Helvering concerned a proceeding before
the Board of Tax Appeals, where the petitioner and the
Commissioner of Internal Revenue presented separate arguments to
the Board. See 312 U.S. 552, 553–54 (1941); see also Blair v.
Oesterlein Mach. Co., 275 U.S. 220, 225 (1927) (same).
Unemployment Compensation Commission of Alaska v. Aragon
involved a proceeding before a hearing examiner who similarly took
testimony from adverse parties. See 329 U.S. 143, 147–48 (1946);
Aragon v. Unemployment Comp. Comm’n of Alaska, 149 F.2d 447,
450–52 (9th Cir. 1945). And United States v. L.A. Tucker Truck
Lines, Inc. concerned a proceeding before the Interstate Commerce
Commission, where the hearing examiner was tasked solely with
 13
The FDA’s review is also far less formal than the multi-step
regulatory process in the Social Security context, which the
Court found insufficiently adversarial in Sims and Carr. In
short, the proceedings before the FDA do not support the
“‘analogy to judicial proceedings’ that undergirds judicially
created issue-exhaustion requirements.” Carr, 141 S. Ct. at
1360 (quoting Sims, 530 U.S. at 112 (plurality opinion)).

Second, as to the nature of Sandoz’s claim, it is a question
of law on which the agency’s views have been fully articulated
before the district court and in the briefs and arguments before
this court. Moreover, the resolution of this dispute does not
“require the development of a factual record.” Yardmasters,
721 F.2d at 1338.

In the circumstances here, Sandoz’s failure to present the
issue to the agency does not bar adequate consideration by this
court. Review of Sandoz’s legal arguments is appropriate, and
we decline to impose an exhaustion requirement.

B.

Considering the text and structure of the FDCA, we hold
that when the FDA approves a new drug, known impurities in
that drug have not “been approved” within the meaning of the
statute. Therefore, a new drug that employs a known impurity
as its active ingredient is eligible for new chemical entity
exclusivity. 8 Aubagio employs a known impurity—

“decid[ing] justly between contestants in an adversary proceeding.”
344 U.S. 33, 36 (1952). These factual scenarios could not be further
afield from the letter correspondence at issue here.
8
We do not address the circumstance in which a new drug uses as its
active ingredient an “inactive ingredient” identified in a previously
approved drug. FDA regulations and guidance documents provide
distinct definitions for inactive ingredients and impurities, as well as
 14
teriflunomide—as the active ingredient in a new drug to treat
multiple sclerosis. We agree with the FDA that because
teriflunomide was only an impurity and therefore not
previously “approved” in Arava, Aubagio was entitled to new
chemical entity exclusivity.

1.

We begin with the text of the statute. The FDA and Sandoz
disagree about whether FDA approval of a drug means the
agency “has approved” known impurities in that drug, such that
a later new drug that uses an impurity as its active ingredient
can benefit from new chemical entity exclusivity.

New chemical entity exclusivity is granted “for a drug, no
active ingredient … of which has been approved in any other”
new drug application. 21 U.S.C. § 355(j)(5)(F)(ii). Reading
this provision in context, it is plain the FDA does not “approve”
a known impurity when it approves a drug that contains that
impurity. To “approve” means “[t]o sanction officially.”
Approve, BLACK’S LAW DICTIONARY (6th ed. 1990); see also
Approve, v.1, OXFORD ENGLISH DICTIONARY (2d ed. 1989)
(meaning “[t]o confirm authoritatively; to sanction”); Approve,
MERRIAM-WEBSTER’S NINTH NEW COLLEGIATE DICTIONARY
(1983) (meaning “to accept as satisfactory” or “to give formal
or official sanction to”).

The FDA did not approve or “sanction officially”
teriflunomide when it recognized it as an impurity in Arava.
When the FDA receives a new drug application, the question
before the agency is whether the proposed drug meets a variety

different regulatory treatment. The issues presented and the parties’
briefing concern only impurities. This opinion does not address the
application of new chemical entity exclusivity to drugs that make use
of a previously recognized inactive ingredient.
 15
of criteria—including whether pharmacological tests reliably
demonstrate the drug’s efficacy and safety and whether the
manufacturing process will preserve the drug’s purity. See 21
U.S.C. § 355(d). The FDA makes a single decision whether to
approve or reject the drug.

While the FDA approves the drug as a whole, assessment
and study of the active ingredient is central to the new drug
approval process. See, e.g., 21 U.S.C. §§ 355(b)(1), j(4),
(j)(5)(F)(ii), (j)(5)(F)(iii); see also Pharmanex v. Shalala, 221
F.3d 1151, 1156 (10th Cir. 2000) (noting that “approval of
active ingredients is integral to the overall new drug approval
process”). Active ingredients are those “intended to furnish
pharmacological activity or other direct effect in the diagnosis,
cure, mitigation, treatment, or prevention of disease.” 21
C.F.R. § 314.3(b). The approval process focuses primarily on
assessing the operation and safety of such ingredients. The
active ingredient must fulfill a specific pharmacological
purpose as demonstrated by clinical trials. Id.; 21 U.S.C.
§ 355(d). An active ingredient is approved in a new drug only
at a specified quantity and for a specified use, parameters that
must be set out in the drug’s label. See 21 U.S.C. §§ 352(e)-(f).
For instance, the Arava label states the drug contains either 10,
20, or 100 milligrams of leflunomide, to be taken orally to treat
active rheumatoid arthritis. The agency’s review of the active
ingredient is affirmative and thorough, designed to ensure that
upon approval the drug will serve its intended function safely.

By contrast, the FDCA new drug approval process
includes no reference to impurities at all. Instead, impurities are
a category created by the agency, relating to the statutory
requirement that new drug applicants list the “components” of
the proposed drug. 21 U.S.C. § 355(b)(1). FDA guidance
documents define an “impurity” as “[a]ny component” of the
drug that is neither the active ingredient nor an “excipient.”
 16
Excipients are “inactive ingredients that are routinely and
purposefully added … to enhance the performance of the active
ingredient.” Glaxo Grp. Ltd. v. Apotex, Inc., 376 F.3d 1339,
1347 (Fed. Cir. 2004). Unlike active ingredients and excipients,
impurities are not deliberately added to the drug but may “arise
during the synthesis, purification, and storage of a new drug
substance,” when, for instance, chemicals in the drug react or
degrade.

Impurities are reviewed only to assess whether the
manufacturing process is “inadequate to preserve” the drug’s
“identity, strength, quality, and purity.” 21 U.S.C. § 355(d).
The FDA focuses on whether the impurity undermines the
safety or efficacy of the drug but does not otherwise directly
study the pharmacological effects of an impurity. The agency
may tolerate low levels of impurities when approving a drug.
But the limited review does not sanction or approve the
impurities, some which may be benign or even beneficial and
others which may in fact be toxic or carcinogenic. See Sandoz
Inc. v. Becerra, 2022 WL 2904262, at *6 (D.D.C. July 22,
2022); Draft Guidance for Industry on Genotoxic and
Carcinogenic Impurities in Drug Substances and Products:
Recommended Approaches; Availability, 73 Fed. Reg. 76,361,
76,362 (Dec. 16, 2008) (draft guidance designed to provide
“approaches for handling impurities with known genotoxic or
carcinogenic potential,” including “reduction of the impurity
level to an acceptable threshold”).

In approving a new drug, the FDA does not specify the
precise amount of a given impurity that the drug must contain,
only a ceiling. This makes sense because, as was the case for
Arava, impurities often arise unintentionally in manufacture
and storage, and the quantity of a given impurity may vary
“from batch to batch” or “at different times over the shelf life”
of the drug. If the impurity could be eliminated altogether from
 17
the manufacturing process, so much the better. Whatever the
variable quantity of an impurity, the FDA insists only that it
does not reach a level that undermines the safety or efficacy of
the drug. The statute and regulatory scheme make clear that
ingredients identified as impurities are not rigorously studied
for their “direct effect” in the treatment and cure of disease.
Rather, the FDA ensures that any impurities arising from the
drug’s production and manufacture do not indirectly
compromise the overall safety and efficacy of the drug and its
active ingredient.

The FDA’s identification of a permissible upper level of
an impurity in no way suggests the impurity “has been
approved” within the plain meaning of the FDCA. Therefore,
we conclude that when the FDA recognized teriflunomide as
an impurity in Arava, it did not approve that ingredient. The
FDA properly granted new chemical entity exclusivity to
Aubagio, which utilized teriflunomide as an “active ingredient”
in a new drug.

This conclusion is reinforced by the “language and design
of the statute as a whole.” K Mart Corp. v. Cartier, Inc., 486
U.S. 281, 291 (1988). To begin with, active ingredients are also
central to the FDCA’s approval process for generic drugs. An
ANDA may be submitted only when the “active ingredient” of
the proposed generic drug “is the same as that of the listed
drug.” 21 U.S.C. § 355(j)(2)(A)(ii). Generic drugs also must be
identical to the listed drug in other ways—for instance, by
having the same dosage and identical labeling. See id.
§§ 355(j)(2)(A)(iii), (v). Yet there is no requirement that a
generic drug contain the same impurities as the listed drug.

Similarly, the FDA closely regulates the labeling of drugs,
but not with respect to impurities. A drug’s label must disclose
each active ingredient and its exact quantity, as well as a list of
 18
each inactive ingredient, but the label need not say anything at
all about impurities. See id. §§ 352(e)(1)(A)(ii)–(iii) (requiring
the name and quantity of each active ingredient and the name
of each inactive ingredient to be listed, but imposing no
requirement as to impurities); 21 C.F.R. § 201.66(c)
(enumerating the labeling requirements for over-the-counter
drugs and making no reference to impurities); id. § 201.100(b)
(same for prescription drugs). These provisions confirm the
indirect and largely negative consideration of impurities in the
FDA’s approval and labeling requirements.

Granting new chemical entity exclusivity to Aubagio is
also consistent with Congress’s careful calibration of the terms
of exclusivity. The FDCA grants a longer term of exclusivity
for drugs developed with a novel active ingredient and a shorter
period for drugs that utilize an already approved active
ingredient for a novel use. Compare 21 U.S.C.
§ 355(j)(5)(F)(ii) (permitting new chemical entity exclusivity
for four or five years when the active ingredient in the drug has
never been approved), with id. § 355(j)(5)(F)(iii) (permitting
three years of exclusivity when the active ingredient in a new
drug was previously approved but the sponsor makes use of
“new clinical investigations” “essential to the approval of the
application”).

Because Aubagio was the first drug to utilize
teriflunomide as an active ingredient, the FDA properly
granted four years of exclusivity. As already discussed, the
evaluation of a new drug’s active ingredient requires rigorous
testing and safety reviews before the drug may receive
approval. When a previously approved active ingredient is
repurposed to a new use, the applicant and the agency will have
had the benefit of prior clinical trials, a previous new drug
approval process squarely focused on resolving whether the
ingredient is efficacious and safe, and previous experience in
 19
manufacturing the ingredient at a definite quantity. A
manufacturer developing an impurity as an active ingredient
for the first time will not have the advantage of such rigorous
testing and study. After all, impurities are reviewed only for the
specific purpose of determining whether their presence in a
limited quantity will compromise the drug.

As the Government explains, “[i]n practice, few
substances are entirely unknown,” and major innovations often
occur when substances are first given therapeutic use, even if
the substance was previously recognized by the FDA.
Therefore, identifying that an impurity serves a
pharmacological benefit as an active ingredient will require
additional research and testing to satisfy the FDA’s rigorous
approval standards. See id. § 355(b)(1) (requiring “full reports
of investigations” to show the drug is safe and effective).

Consistent with the statutory and regulatory scheme, the
development of an impurity into an active ingredient is much
more like identifying a novel active ingredient than like
undertaking clinical investigations to find a new use for a
previously approved active ingredient. Aubagio employed a
known impurity as its active ingredient and so the longer
exclusivity period was appropriate.

2.

Furthermore, we note Congress has clarified the scope of
new chemical entity exclusivity in a manner that reaffirms that
the FDA’s approval process is concerned with active
ingredients and not impurities. In section 505(u) of the FDCA,
Congress extended new chemical entity exclusivity to certain
drugs that use part of the active ingredient in a previously
approved drug. Food and Drug Administration Amendments
Act of 2007, Pub. L. No. 110-85, tit. XI, § 1113, 121 Stat. 823,
976–77 (codified at 21 U.S.C. § 355(u)). More precisely, a
 20
single enantiomer of a racemic mixture may be eligible for
exclusivity. A racemic mixture is an equal combination of two
enantiomers—molecules that are oriented as mirror images of
each other, much like a person’s left and right hands. See
generally Mark A. Lemley, Expecting the Unexpected, 92
NOTRE DAME L. REV. 1369, 1377 (2017). Section 505(u)
ensures that a drug making novel use of a single enantiomer is
eligible for new chemical entity exclusivity, even if the racemic
mixture was previously approved. Applicants for such a drug
may “elect to have the single enantiomer not be considered the
same active ingredient as that contained in the approved
racemic drug” for purposes of new chemical entity exclusivity.
21 U.S.C. § 355(u)(1). This reflects the fact that a single
enantiomer may be therapeutically useful in novel ways.
Lemley, 92 NOTRE DAME L. REV. at 1377–78.

Section 505(u), however, is operative only if the racemic
mixture was previously approved as an “active ingredient,” not
if the racemic mixture was only a known impurity. See 21
U.S.C. § 355(u)(1) (providing that the enantiomer may “not be
considered the same active ingredient as that contained” in the
previously approved racemic drug) (emphasis added). This
confirms that FDA approval of a racemic drug is equivalent to
approval of the active ingredient in that racemic drug.

Sandoz argues that a “racemic drug” must be a drug that
has a racemic mixture as an active ingredient, so section 505(u)
simply is not implicated when a racemic mixture is approved
as an impurity. But Sandoz’s interpretation of section 505(u),
combined with its view that exclusivity is barred for a known
impurity, leads to a puzzling outcome. Consider a new drug
that employs an enantiomer as its active ingredient. Under
Sandoz’s understanding, if the racemic mixture of which the
enantiomer is a part was previously recognized only as an
impurity, the new drug would not receive exclusivity. On the
 21
other hand, if the racemic mixture was previously approved as
an active ingredient, the new drug using the enantiomer would
be eligible for exclusivity under section 505(u).

If we accepted Sandoz’s interpretation, there would be no
exclusivity for researching and developing an enantiomer
previously recognized as a mere impurity, yet there would be
exclusivity for repurposing an enantiomer that had already
been rigorously studied and approved as an active ingredient.
This outcome cannot be squared with the text and structure of
the FDCA, in which Congress aligned longer periods of
exclusivity with greater innovation and research.

3.

Finally, we are unpersuaded by Sandoz’s contrary
interpretation, which would significantly constrain the scope of
new chemical entity exclusivity in a manner inconsistent with
the statutory scheme. Sandoz maintains that when the FDA
approves a new drug, it approves all the ingredients in that drug
that are known and disclosed to the FDA, including any
impurities. This argument relies heavily on the grammar of the
relevant statutory language. Recall that new chemical entity
exclusivity is available when “an application submitted under
subsection (b) of this section for a drug, no active ingredient …
of which has been approved in any other application under
subsection (b) of this section, is approved.” 21 U.S.C.
§ 355(j)(5)(F)(ii). Because the relative pronoun “which”
modifies “drug,” Sandoz argues the only question is whether
the active ingredient in the later drug was “approved” as any
ingredient in a previously approved drug. Sandoz would deem
any known impurities in the drug, along with the active
ingredients, to have been approved by the FDA.

Sandoz’s argument hinges on the rules of grammar. While
such rules may be a useful guide to statutory meaning, they
 22
cannot resolve meaning when divorced from the text and
structure of the statute. A narrow focus on grammar fails to
answer the relevant statutory question here, which is whether
the FDA has “approved” an impurity when it approves a new
drug containing that impurity. Our conclusion that the FDA
does not in fact approve such an impurity is compelled by the
language of the provision, the statutory context, and the scope
of the new drug approval process implemented by the FDA
through its regulations and guidance.

Sandoz suggests the FDA approves “each of the known
components disclosed” in an NDA, but as we have already
explained, the limited role that impurities play in the new drug
approval process makes clear the FDA does not “approve”
known impurities within the meaning of the statute. Sandoz’s
interpretation cannot be squared with the FDCA and the reality
of the agency’s approval process for new drugs.

***

In sum, the FDA does not officially sanction an impurity
when it approves a drug as a whole. When the FDA concludes
a particular impurity is permissible up to certain levels in an
approved drug, it is merely determining that a limited quantity
of the impurity will not bar the approval of a new drug. The
FDA did not “approve” teriflunomide when it approved the
drug Arava, and therefore Aubagio was entitled to new
chemical entity exclusivity.

III.

We also reject Sandoz’s alternative argument that
teriflunomide in fact was an active ingredient in Arava. The
record is clear that the FDA did not approve teriflunomide as
such when it approved Arava. The Arava application referred
to leflunomide as the only active ingredient in the drug. The
 23
Arava approval process and efficacy review focused entirely
on leflunomide. The FDA’s ultimate approval was simply for
“Arava (leflunomide) [t]ablets” with the label “ARAVATM
(leflunomide).” Nothing in the Arava application suggested
that the variable quantity of teriflunomide as a degradant was
“intended” to furnish any “pharmacological activity” or “direct
effect.” Instead, the application focused entirely on the
therapeutic effects of leflunomide.

In light of this record, Sandoz does not suggest the FDA
explicitly approved teriflunomide as an active ingredient.
Instead, Sandoz argues teriflunomide met the regulatory
criteria for an “active ingredient,” because it is a “component
that is intended to furnish pharmacological activity or other
direct effect in the diagnosis, cure, mitigation, treatment, or
prevention of disease.” 21 C.F.R. § 314.3(b). Sandoz contends
that the small quantity of teriflunomide that accumulates in
Arava during manufacturing and storage assists in treating
rheumatoid arthritis and thus falls within this definition. To
support its claim, Sandoz cites to contemporaneous patent
filings, which supposedly suggest Arava’s sponsors believed
the degradation of leflunomide into teriflunomide would assist
Arava’s pharmacological effects. Sandoz maintains the FDA
erred in not considering teriflunomide as an active ingredient.

The FDA, however, was not required to take notice of
patent filings in its review of Arava’s application. In reviewing
new drug applications, the FDA makes decisions “upon the
basis of the information submitted … as part of the application”
or otherwise “before” the agency. 21 U.S.C. § 355(d). While
applications must list patents for which “a claim of patent
infringement could reasonably be asserted,” see id.
§ 355(b)(1), these requirements are designed to make public
the patents associated with a drug, not to further the FDA’s
substantive review of the drug. The FDA’s review of patent
 24
information is “purely ministerial,” and the statute requires the
agency “to publish submitted patent information” but not “to
review the merits of the patent information provided.” Teva
Pharms., USA, Inc. v. Leavitt, 548 F.3d 103, 106 (D.C. Cir.
2008). After all, the FDA has no particular expertise in patent
law. Cf. aaiPharma Inc. v. Thompson, 296 F.3d 227, 241 (4th
Cir. 2002). Moreover, new drug applications must include
certain safety and efficacy information, including an
exhaustive list of “technical sections,” but none of these
requirements reference patents at all. See 21 C.F.R.
§ 314.50(d). And finally, the Arava application did not even
include the patents that Sandoz suggests should have informed
the FDA’s review.

In effect, Sandoz asks this court to modify the FDA’s
Arava approval from 1998 to include teriflunomide as an
additional active ingredient based on patent applications
outside the record, despite the fact that Arava’s application,
clinical trials, approval, and labeling all considered
leflunomide as the only active ingredient. We decline Sandoz’s
invitation to rewrite the FDA’s drug approval decision decades
after the fact. On the record before it, the FDA properly
examined leflunomide as the active ingredient in Arava and
considered teriflunomide only as an impurity. New chemical
entity exclusivity was therefore appropriate for Aubagio, which
was the first drug to use teriflunomide as an active ingredient.

***

Consistent with the statutory and regulatory requirements
for new drug approval, the FDA’s approval of a new drug
includes approval of the drug’s active ingredient, but not its
impurities. When the FDA approved Arava, it recognized
teriflunomide as an impurity and hence did not approve it
within the meaning of the FDCA. Teriflunomide was approved
 25
for the first time as the active ingredient in Aubagio, which
properly received new chemical entity exclusivity. Because
Sandoz’s challenge fails, its generic equivalent will not benefit
from a period of sole marketing exclusivity. The judgment of
the district court is affirmed.

So ordered.