Amneal Pharmaceuticals LLC v. Food and Drug Administration

UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF COLUMBIA

AMNEAL PHARMACEUTICALS LLC,

Plaintijj”,
v.
:l()OD AND DRUG ADMINISTRATION, et Civil Action NO. 17_180 (RDM)
Defendcmts,
and

LUPIN PHARMACEUTICALS, INC., et al.,

lntervenor-Defendants.

MEMORANDUM OPINION

The Federal Food, Drug, and Cosmetic Act (“FDCA”), as amended by the Drug Price
Competition and Patent Restoration Act of 1984 (“Hatch-Waxman Act”) and the Medicare
Prescription Drug Improvement and Modernization Act of 2003 (“Medicare Modemization
Act”), grants 180 days of market exclusivity to the first generic manufacturer to file an
abbreviated new drug application (“ANDA”) that challenges a patent covering the brand-name
version of the drug. But that right is not absolute and is subject to various statutorily defined
‘.‘forfeiture events.” As relevant here, an applicant Who fails “to obtain tentative approval of [its
ANDA] Within 30 months after the date on Which the [ANDA] is filed” forfeits its right to the
180 days of market exclusivity, “unless th[at] failure is caused by a change in or a review of the

requirements for approval of the application imposed after the date on Which the application is

nled.” 21 U.s.C. § 355@)(5)(D)(i)(1v).

The plaintiff in this case, Amneal Pharmaceuticals LLC (“Amneal”), was the first
manufacturer to file an ANDA to market a generic version of Namenda XR (memantine
hydrochloride extended release capsules), and its ANDA challenged a patent held by the
manufacturer of the brand-name version of the dmg. Amneal did not, however, obtain tentative
approval for its ANDA within the 30-month window. As a result, Amneal’s eligibility for the
180 days of generic market exclusivity turns on whether its failure to obtain timely approval was
“caused by a change in or a review of the requirements for approval of the” ANDA. ln
proceedings before the Food and Drug Administration (“FDA”), Amneal argued (1) that the
delay was caused by the FDA’s demand for data from a commercial-scale batch of the drug and
(2) that this demand constituted a change in a requirement for approval of the ANDA. The FDA
disagreed on both counts. Amneal, in response, brought the present action pursuant to the
Administrative Procedure Act (“APA”), 5 U.S.C. § 701 et seq., challenging the FDA’s decision.

The case is now before the Court on Amneal’s motions for summary judgment, Dkt. 25,
and for a preliminary injunction or, in the altemative, a temporary restraining order, Dkt. 62, and
the cross-motions for summary judgment filed by the FDA, Dkt. 34, and intervenor-defendant
Lupin Pharmaceuticals (“Lupin”), Dkt. 30. Although Amneal raises a host of other issues, the
heart of the dispute is whether the FDA’s request that Amneal supplement its ANDA with data
from a commercial-scale batch, as opposed to data from the pilot-scale batches that Amneal
originally submitted, constituted a “change in . . . the requirements for approval of” Amneal’s
ANDA_or, more precisely, whether the FDA’s decision that the request did not was contrary to
law or otherwise unreasonable As explained below, the Court concludes that the FDA acted
within its authority and reasonably in rejecting Amneal’s claim to market exclusivity. The Court

will, accordingly, deny Amneal’s motion for summary judgment, Dkt. 25, and will grant the

FDA’s and Lupin’s cross-motions for summary judgment, Dkt. 30; Dkt. 34. Moreover, having
resolved the case on the merits, the Court will deny Amneal’s motion for a preliminary
injunction or, the altemative, a temporary restraining order, Dkt. 62, as moot.
I. BACKGROUND

A. Statutory and Regulatory Background

To obtain approval to market a new drug, the innovator-manufacturer must submit a new
drug application (“NDA”) to the FDA that contains extensive information and data, including
“full reports of investigations which have been made to show” that the new drug is Safe and
effective, “a full statement of the composition” of the new drug, “a hall description of the
methods used in, and the facilities and controls used for, the manufacture, processing, and
packaging” of the new drug, and proposed labeling for the new drug. 21 U.S.C. § 355(b)(1). In
addition, the NDA must include “the patent number and the expiration date of any patent which
claims the drug for which the applicant submitted the [NDA] or which claims a method of using
[the new] drug and with respect to which a claim of patent infringement could reasonably be
asserted if a person not licensed by the owner engaged in the manufacture, use, or sale of the
drug.” Id. The FDA then lists this patent information in the “Orange Book: Approved Drug
Products with Therapeutic Equivalence Evaluations,” or, as it is commonly known, simply the
“Orange Book.”l See also Mylan Labs. Ltd. v. FDA, 910 F. Supp. 2d 299, 301 (D.D.C. 2012).

ln 1984, Congress enacted the Drug Price Competition and Patent Term Restoration Act,
Pub. L. No. 98-417, 98 Stat. 1585 (1984), popularly known as the “Hatch-Waxman Act.” The
Act sought, among other things, to encourage the development of generic drugs to increase

market competition and to lower prices for consumers. See Meaa' Johnson Pharm. Grp., Mead

' Available at https://www.accessdata.fda.gov/scripts/cder/ob/index.cfm.

Johnson & Co. v. Bowen, 838 F.2d 1332, 1333 (D.C. Cir. 1988). To that end, the Hatch-
Waxman Act streamlined the process for bringing new generic drugs to market by creating the
abbreviated new drug application_or “ANDA”_process, under which a manufacturer can
“piggyback[] on the original manufacturer’s evidence of safety and efficacy.” T eva Pharm.,
USA, lnc. v. Leavitt, 548 F.3d 103, 104 (D.C. Cir. 2008). To obtain FDA approval for a generic
drug, the ANDA applicant must demonstrate (1) that it is “bioequivalent” to the brand-name
drug (also referred to as the “listed drug”); (2) that the prescribed conditions of use have been
approved for the listed drug, 21 U.S.C § 355(j)(2)(A); (3) that the generic drug satisfies certain
chemistry and labeling requirements; (4) and that the proposed manufacturing and packaging
processes and controls are adequate to “preserve its identity, strength, quality, and purity,” id. §
355(j)(4)(A).

Although the Hatch-Waxman Act streamlined the process for bringing generic drugs to
market, obtaining FDA approval for an ANDA remains a prolonged task. The process begins
with the submission of an application, which must comply with FDA requirements for receipt. If
an application is not “substantially complete,” 21 C.F.R. § 314.101(b)(l), the FDA will issue a
“refuse-to-receive decision” in which it rejects the ANDA Without evaluating the substance of
the application, id. § 314.101(b)(3). In response, an applicant may resubmit the ANDA with the
required data or may withdraw the ANDA. Id. Only after the FDA has determined that the
ANDA meets the receipt requirements does the agency begin to evaluate whether the applicant’s
product is bioequivalent, manufactured in an appropriate manner, and properly labeled. During
this substantive review, there is “inevitably and invariably back and fort ” between the agency

and the applicant, see Dkt. 69 at 8 (Tr. 8:7), in which the agency generally requests additional

data or information to determine whether the ANDA meets the requirements for approval, see 21
C.F.R. § 314.127. This process can take years to complete.

An ANDA must also include one of four certifications with respect to each of the Orange
Book patents that claims the listed drug. 21 U.S.C. § 355(]')(2)(A)(vii)(I)-(IV). The four
certifications are identified by the paragraph in which they appear: A “paragraph I” certification
attests that, to the best of the applicant’s knowledge, no such patent information has been filed.
Id. § 355(j)(2)(A)(vii)(I). A “paragraph ll” certification attests that any such patent has expired.
Id. § 355(]`)(2)(A)(vii)(ll). A “paragraph III” certification identifies the date on which any such
patent will expire. Ia'. § 355@)(2)(A)(vii)(lll). And, most importantly for present purposes, a
“paragraph IV” certification asserts that any such patent is invalid or will not be infringed by the
generic drug. Id. § 355(j)(2)(A)(vii)(lV). “If an ANDA applicant makes one of the first two
certifications, [the] FDA may approve the ANDA immediately,” and, if the “applicant makes a
paragraph III certification, [the] FDA may grant tentative approval of the ANDA, to be made
effective on the date the patent expires.” Mylan Labs., 910 F. Supp. 2d at 301. The
consequences of making a paragraph IV certification, however, are different in kind. Most
significantly, a paragraph IV certification is treated as an act of patent infringement 35 U.S.C. §
271(e)(2)(A). If the patent holder brings an infringement action within 45 days of receiving
notice of the certification, moreover, the FDA’s approval of the ANDA will not take effect for a
period of 30 months, unless otherwise ordered by the district court in which the patent litigation
is pending Id. § 355(]`)(5)(B)(iii).

Would-be generic drug manufacturers thus risk costly patent infringement litigation if
they seek FDA approval before the brand-name drug’s patent has expired or is invalidated_a

risk that Congress feared would discourage market competition and delay bringing lower-priced

generic drugs to market. See T eva Pharm. USA, Inc. v. Sebelius, 595 F.3d 1303, 1305 (D.C. Cir.
2010). To “compensate [generic] manufacturers for research and development costs as well as
the risk of litigation from patent holders,” T eva Pharm., 548 F.3d at 104, Congress enacted an
incentive for generic drug manufacturers to submit ANDAs and, if necessary, to engage in patent
litigation: the “first applicant” to file an ANDA containing a paragraph IV certification is eligible
for 180 days of market exclusivity, “during which the FDA may not approve for sale any
competing generic version of the drug at issue.” T eva Pharm., 595 F.3d at 1305; see also 21
U.S.C. 355(]`)(5)(B)(iv); Mylan Labs., 910 F. Supp. 2d at 302.

The first-filer’s entitlement to the exclusivity period is not absolute, however, and may be
forfeited under certain conditions. As relevant here, the exclusivity period is forfeited if the
“first applicant fails to obtain tentative approval of [its ANDA] within 30 months after the date
on which the [ANDA] is filed.” 21 U.S.C. § 355(j)(5)(D)(i)(IV). Congress added this forfeiture
rule to the statute as part of the Medicare Modemization Act of 2003, Pub. L. No. 108-173, 117
Stat. 2066 (2003), to “ensure that the 180-day exclusivity period enjoyed by the first generic to
challenge a patent [could not] be used as a bottleneck to prevent additional generic competition.”
Hi-Tech Pharmacal Co. v. U.S. Food & DrugAa'min., 587 F. Supp. 2d 1, 4 (D.D.C. 2008)
(quoting 149 Cong. Rec. S15746 (daily ed. Nov. 24, 2003) (statement of Sen. Schumer)).

Congress created one exception to the 30-month forfeiture rule: the 180-day market
exclusivity is not forfeited if the applicant’s failure to obtain tentative approval within 30 months
was “caused by a change in or a review of the requirements for approval of the application
imposed after the date on Which the application is filed.” 21 U.S.C. § 355(]`)(5)(D)(i)(IV). This
exception was included to ensure that ANDA applicants are not penalized “for delays in tentative

approval caused by changes in approval requirements beyond [the] ANDA applicant’s control.”

Mylan, 910 F. Supp. 2d at 311. If an applicant fails to obtain tentative approval within 30
months and the failure is not excused under the exception, however, the FDA may approve other
ANDAs for the same generic drug product. See id.

B. Factual Background

On June 10, 2013, Amneal submitted an ANDA seeking approval to market a generic
version of Namenda XR~a medication intended to treat moderate to severe dementia of the
Alzheimer’s type. AR 153, 160.2 Namenda XR is a memantine hydrochloride capsule designed
to facilitate an “extended release” of the drug. AR 1059. The FDA had previously approved an
NDA submitted by Forest Laboratories LLC (“Forest”) to market Namenda XR in four dosage
levels: 7 mg, 14 mg, 21 mg, and 28 mg. AR 160. Amneal’s ANDA included all four dosage
levels; it was the first filed; and it included a paragraph IV certification for at least one of the
patents listed in the Orange Book. AR 1041, 1043-44.

Amneal’s ANDA proposed to achieve Namenda XR’s extended-release effect by
manufacturing capsules containing tiny “beads” of the active ingredient that would dissolve into
the body at different rates. AR 182. Unlike the brand-name version of Namenda XR, which
includes a single type of bead, Dkt. 33 at 13 & n.5, each capsule of Amneal’s proposed product
would contain both immediate release beads (“IRs”) and extended release beads (“ERs”) -
_, AR 185. Consistent with the FDA’s Guidance on ANDA
Submissions_Refuse-to-Receive Standards, AR 240, 252, Amneal submitted data from three

“pilot batches” of 150,000 capsules for each of the four dosage strengths in its ANDA, AR 198-

2 The seven-part Administrative Record (“AR”) can be found at Dkt. 46 (pages 1_140), Dkt. 46-
1 (pages 141-239), Dkt. 46-2 (pages 240-406), Dkt. 46-3 (pages 418-544), Dkt. 46-4 (pages
545-637), Dkt. 46-5 (pages 638_753), and Dkt. 46-6 (pages 754-1063). For ease of reference,
the Court will cite to the AR page numbers.

99. These pilot batches produced the capsules by combining _

to create the extended release effect. AR 437. For purposes of commercial sales, however,
Amneal proposed to produce much larger batches of 1,000,000 capsules and, in doing so, to
blend up to six different lots of ER beads with two different lots of IR beads. AR 199, 437.

On November 5, 2013, the FDA sent Amneal a letter refusing to receive Amneal’s
ANDA because it was “not sufficiently complete to merit a critical technical review.” AR 63.
Specifically, the FDA explained that Amneal had “failed to produce accelerated stability data
which encompasses at least 84 days in the accelerated stability chamber for the 30 and 500
counts for all strengths.” Id. Amneal responded a month later by clarifying that the required
data had, in fact, been included-it only appeared insufficient because a “typographical error”
indicated the wrong date on which the stability testing began. AR 67. After confirming that the
date in the ANDA was indeed a typographical error, the FDA rescinded its “Refuse to Receive”
letter and treated the date of Amneal’s original submission as the date of the application. AR
109. Although the record does not reflect when other generic manufacturers filed their ANDAS,
there is no dispute that, using Amneal’s original date of submission, Amneal qualified as the
“first applicant” to submit an ANDA for Namenda XR containing a paragraph IV certification,
making it eligible for 180 days of generic exclusivity if approved within 30 months, i.e., by
December 10, 2015. AR1016.

On September 10, 2014, the FDA sent Amneal a letter identifying a number of
substantive deficiencies with Amneal’s ANDA. See AR 23 0-39. As relevant here, the FDA

wrote:

The proposed combination of multiple ER and IR lots for commercial production,
combined with the deficiencies observed in the information submitted for the
exhibit lots, requires the review of the resulting data from the manufacture of a
commercial size lot before application approval. In this regard, please submit all

the required [chemistry, manufacturing, and controls] information for the

production of a 7 mg and 28 mg Memantine HCl lot of the size intended for

commercial distribution
AR 233. As the FDA further explained in its chemistry review, this deficiency was based on
several factors. In particular: (l) Amneal’s product required a specific ratio of IR beads to ER
beads in each capsule _ to ensure the desired rate of release, AR 189, 222-23; (2) the
pilot batches used in the ANDA involved mixing _ of lR beads,
while Amneal proposed to market a product produced in commercial batches made by mixing
“six different ER batches with two IR lots,” AR 199; (3) Amneal’s pilot batches had deficiencies
such as overfilled capsules, AR 224; and (4) the “in-process controls” that Amneal proposed to
use for the commercial manufacture of the drug were “regular controls used for simpler
encapsulated drugs[,] which [might] not be sufficient to . . . monitor the quality of” a product
made by combining “different types of . . . beads,” AR 203. Overall, the FDA reviewers
concluded that, “[d]ue to the complexity of the proposed process, combined with the deficiencies
noted in the review,” and because “mixing multiple lots of ER and IR beads could induce
additional deviations not observed during the manufacture of the [pilot] lots that may require
additional control,” Amneal “should submit a scale up batch before approval.” AR 199; see also
AR 233 (requesting that Amneal “please submit all the required . . . information for” commercial
lot sizes).

Complying with the FDA’s request required considerable time and effort on Amneal’s
part. Amneal needed to “acquire commercial-scale quantities of source materials, ramp up
production to manufacture commercial-size lots, conduct comprehensive analytical testing on the
finished product, evaluate the resulting data, and resubmit its ANDA”_a process that took eight

months to complete. AR 654. But, upon review of the additional data, the FDA noted several

issues. Most problematically, Amneal’s proposed lR-to-ER bead ratio “was not followed in the
newly submitted batches” and, while “some latitude in this ratio [was] required[,] . . . the
deviation . . . was not justified.” AR 460.

Amneal submitted a lengthy response to the FDA’s review, which, among other things,
highlighted the differences in processes used in producing the pilot-scale and commercial-scale
batches. AR 43 7-44. As Amneal explained, “[i]n the ANDA exhibit batches, _
_ were used to manufacture the intermediate blend; whereas
for manufacturing of future commercial scale batches, multiple batches of IR pellets and ER
pellets will be used.” AR 437. Acknowledging that combining multiple lots of IR and ER beads
introduced additional “complexity” into the production process, which “may result in calculation
errors,” Amneal proposed an “additional control . . . to ensure [that the] accurate dose of the drug
product is administered to the patients.” AR 444. The FDA, however, concluded that the
additional control was “not acceptable,” and it, accordingly, “recommend[ed]” that Amneal
“scale down the proposed commercial batch size[] to a size with a constant, reproducible capsule
filling weight as well as a constant bead type composition.” AR 483. Amneal accepted the
FDA’s recommendation to scale down the proposed commercial-scale batches_from l,000,000
capsules to 150,000 capsules_thus resolving the issue and clearing the way for approval of its
ANDA. AR 488-89.

In the meantime, however, Amneal experienced a further delay: an “Import Alert” was
issued on October 15, 2015, to the supplier of Amneal’s active pharmaceutical ingredient
(“API”). AR 486, 490, 527. As the FDA has explained, an Import Alert signals that an
“imported product will be detained because it appears to be in violation of the FDCA or its

implementing regulations.” Dkt. 34 at 16 n.10. In light of this development, Amneal sought

10

approval to use a different API supplier, AR 486, 490, 527, and sought to amend its ANDA to
substitute the new supplier in place of the one that was previously identified, AR 495. Amneal’s
troubles continued, however, and, on December 7, 2015, the FDA notified Amneal that its
proposal to change its source of API Was “not acceptable” because Amneal had yet to
demonstrate bioequivalence. AR 740. Amneal provided the missing data demonstrating
bioequivalence using the new source of API on February 25, 2016, but FDA chemistry reviewers
raised additional deficiencies in Amneal’s product, which were not addressed until August 11,
2016 and September 15, 2016. AR 1037. Finally, on September 28, 2016, the FDA tentatively
approved Amneal’s ANDA. Ia'. By that time, however, the 30-month period for approval had
run.

C. FDA’s Determination

On December 2, 2015, Amneal requested that the FDA “confirm . . . that Amneal’s
Memantine XR ANDA will remain eligible for 180-day generic marketing exclusivity regardless
of whether the [a]gency grants tentative approval . . . or final approval to that ANDA on or
before December 10, 2015.” AR 640. In support of that request, Amneal raised a variety of
arguments, but only one is relevant for present purposes. That argument posited, as Amneal
argues in this proceeding, that its failure to obtain approval within 30 months of June 10, 2013_
the date that the FDA treated as the date of filing-_was caused by the FDA’s “post-filing review
of and change in the requirements for approval.” AR 651. Amneal asserted that, prior to the
date on which it filed its ANDA, the FDA “long . . . permitted applicants to submit ANDAs . . .
based on the production of, and presentation of date regarding pilot-size test batches.” Id. lt
was only after the ANDA was filed, Amneal argued, that the FDA changed “the pertinent

approval requirements” and required “that [it] scale up and complete ‘the production of a 7 mg

11

and 28 mg Memantine HCl lot of the size intended for commercial distribution’ and then ‘submit
all the required . . . information’ in order to enable FDA ‘review of the resulting data from the
manufacture of a commercial size lot before application approval.”’ Id. (emphasis omitted).
Finally, Amneal asserted that it was unaware “of any other ANDA for which the [a]gency ha[d]
ever required the production and analysis of . . . data from a commercial-size lot as a condition of
approval,” and, indeed, despite having obtained 95 ANDA approvals, it had “never previously
been required to produce” such data in support of an ANDA. AR 652 (emphasis omitted).

On September 28, 2016, the same day the FDA tentatively approved Amneal’s ANDA,
the agency issued a 17-page letter decision concluding that Amneal had forfeited the 180-day
period of exclusivity. AR 1041~57. In that letter, the FDA explained that its request that
Amneal provide commercial-scale data was not a “change” in “requirements for approval” that
would trigger the exception. AR 1055. The “approval requirement” at issue was not, according
to the FDA, the submission of commercial-scale data; rather, the relevant requirement was that
the applicant “demonstrate that the methods used in, or the facilities and controls used for, the
manufacture, processing, and packing of the drug are adequate to assure and preserve its identity,
strength, quality, and purity.” Ia'. The FDA wrote: “That statutory requirement has not
changed, nor has [the] FDA’s specific requirements for certain complex products like Amneal’s
proposed memantine hydrochloride ER capsules to meet it.” ld.

The FDA’s letter decision further explained that its request for commercial-scale data
was nothing new. To the contrary, although limited in its ability to disclose proprietary
information regarding other applications, AR 1055 n.43, the FDA represented that it has, “[s]ince
at least 2010,” requested commercial-scale data for “certain complex drug products and/or a

complex manufacturing process.” AR 1055. The decision whether that information is

12

“necessary,” however, is case-specific “and is made during the substantive review of the ANDA
by the chemistry reviewers.” Ia'. “Merely having a complex process,” the FDA continued, “is
not, by itself, typically sufficient to warrant a request for the manufacture of commercial batches
pre-approval,” butz

[l]f the applicant does not demonstrate a good understanding of its product and

manufacturing process and have appropriate in-process controls, [the] FDA will

have concerns about the ability of the applicant to successfully produce scaled-up

commercial lots of the drug product upon approval. In that situation, [the] FDA

typically would request that an applicant manufacture a commercial scale batch and
submit information on that batch to the Agency for review prior to approving the

ANDA. lf, as is the case here, the applicant then revises its proposed commercial

batch size to match the size of its exhibit batches, this would address the [a]gency’s

concern, as the applicant would no longer be seeking to produce a scaled-up

commercial batch.
Ia'. Accordingly, the FDA reasoned, because neither statutory requirements nor agency policy
had changed, the exception to the 30-month rule did not apply. AR 1056.

As an alternative basis for its decision, the FDA also concluded that the FDA’s request
for data from a commercial-scale batch, even if a change in a requirement fo`r approval, did not
cause Amneal to miss the 30-month mark. AR 1054-55. It premised that conclusion on two
points: First, the FDA’s request for commercial scale data was resolved before the forfeiture
date and, therefore, could not have “preclude[d] tentative approval or approval at that time.” AR
1055. Second, Amneal did not “demonstrate[] that the [FDA’s] request” that Amneal provide
the commercial-scale data “caused it to fail to obtain tentative approval or approval by [the
relevant date].” Ia'.

D. Procedural History
Four months after receiving the FDA’s adverse determination, Amneal filed this APA

action, alleging that the FDA’s decision was arbitrary and capricious and contrary to law. Dkt. 1.

Two other pharmaceutical manufacturers with tentatively approved ANDAs for generic versions

13

of Namenda XR_Lupin Pharmaceuticals, Inc. and Lupin Limited (collectively, “Lupin”), and
Par Pharrnaceutical, Inc. (“Par”)_moved to intervene in support of the FDA, Dkt. 20; Dkt. 52,
and the Court granted those motions, Minute Order (Mar. 6, 2017); Minute Order (Nov. 8, 2017).
ln addition, Amneal filed a motion for summary judgment, Dkt. 25, and the FDA and Lupin
responded with their own cross-motions for summary judgment, Dkt. 30; Dkt. 34.

While the events described above were unfolding, another generic manufacturer, Teva
Pharmaceuticals USA (“Teva”), filed its own ANDA seeking approval to market a generic
version of Namenda XR, and Teva’s ANDA included a paragraph IV certification asserting that
six of the associated patents were invalid or would not be infringed by its product. Forest Labs.,
Inc. v. Teva Pharm. USA, lnc., Nos. 14-121, 14-200, 14-508, 14-686, 14-1058, 14-1271, 2016
WL 54910 (D. Del. Jan. 5, 2016). Forest then filed an infiingement action against Teva in the
U.S. District Court for the District of Delaware, which ultimately entered a judgment of
invalidity on the ground of indefiniteness. Id. On December ll, 2017, the Federal Circuit
affirmed that decision. Forest Labs, Inc. v. T eva Pharm. USA, Inc., Nos. 2016-2550, 2016-2553,
2017 WL 6311688 (Fed. Cir. Dec. 11, 2017). Although the Federal Circuit’s mandate has yet to
issue, and a petition for rehearing is pending, a final affirmance in the T eva case would open the
door for generic manufacturers with approved ANDAs to market their products. Dkt. 63 at 11_
12. Given that prospect, and the prospect that other holders of approved ANDAS might
imminently enter the market, Amneal filed a motion for a preliminary injunction or, in the
alternative, for a temporary restraining order on December 14, 2017. Dkt. 62.

The following day, the Court held a status conference to address how best to proceed in
light of these developments To promote efficiency, the Court proposed-and the parties

agreed_that the Court would issue a decision on the fully briefed cross-motions for summary

14

judgment on an expedited basis, and, at the request of the FDA_and with no objection from
Amneal_the Court agreed to stay the time for Defendants to respond to Amneal’s motion for a
preliminary injunction or, in the altemative, a temporary restraining order. Dkt. 67 at 7 (Tr.
7:12-14). On January 10, 2018, Court heard oral argument on the pending cross-motions for
summary judgment See Dkt. 69.
II. LEGAL STANDARD

Under Federal Rule of Civil Procedure 56, summary judgment is ordinarily available if
the movant demonstrates “that there is no genuine dispute as to any material fact and” that, based
on the uncontested facts, “the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P.
56(a). In the unique context of a case brought under the APA, however, the district court “sit[s]
as an appellate tribunal,” Marshall Cty. Health Care Auth. v. Shalala, 988 F.2d 1221, 1222-23
(D.C. Cir. 1993), to decide “as a matter of law [whether] the agency action is supported by the
administrative record and is otherwise consistent with the APA standard of review,” Coal. for
Common Sense in Gov ’t Procurement v. Unitea' States, 821 F. Supp. 2d 275, 280 (D.D.C. 2011);
see also Citizens to Preserve Overton Park, lnc. v. Volpe, 401 U.S. 402, 415 (1971); Sw. Merch.
Corp. v. NLRB, 53 F.3d 1334, 1341 (D.C. Cir. 1995). In short, it is the role of the administrative
agency to “resolve factual issues” and “to arrive at a decision that is supported by the
administrative record,” while it is the role of the district court “to determine whether or not as a
matter of law the evidence in the administrative record permitted the agency to make the
decision it did.” Hi-Tech Pharmacal, 587 F. Supp. 2d at 18.

III. ANALYSIS
All parties agree that this case is governed by the two-step framework established in

Chevron U.S.A. v. Natural Resources Defense Council, 467 U.S. 837 (1984), as well as the

15

APA’s arbitrary and capricious standard, 5 U.S.C. § 706(2)(A). Under Chevron’s first step, the
Court must consider “whether Congress has directly spoken to the precise question at issue.”
467 U.S. at 842. If so, the Court must “give effect to the unambiguously expressed intent of
Congress.” Id. at 843. But, if the Court concludes that Congress has left an ambiguity or “gap”
to fill on the “precise question at issue,” the Court proceeds to the second step of Chevron. 467
U.S. at 842-43. Under the second step, the Court asks whether the agency’s construction of the
statute is a “permissible” one. Ia'. at 843. If it is, the Court must defer to that construction.

The APA also precludes agency action that is “arbitrary, capricious, an abuse of
discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). An agency’s
decision, accordingly, must be the product of “reasoned decisionmaking” Motor Vehicle Mfrs.
Ass ’n of U.S. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 52 (1983). An agency action will
normally be set aside as “arbitrary and capricious” if the agency “has relied on factors which
Congress has not intended it to consider, entirely failed to consider an important aspect of the
problem, offered an explanation for its decision that runs counter to the evidence before the
agency, or is so implausible that it could not be ascribed to a difference in view or the product of
agency expertise.” Ia'. at 43.

As this Court has previously observed, “[t]hese standards overlap and are, at times,
intertwined.” Amarin Pharm. Ireland Lta’. v. Food & Drug Admin., 106 F. Supp. 3d 196, 206
(D.D.C. 2015). Because Chevron’s second step asks Whether an agency’s interpretation is
“arbitrary or capricious in substance,” Judulang v. Hola’er, 565 U.S. 42, 52 n.7 (2011) (citing
Mayo Found. for Med. Educ. & Research v. Um`tea' States, 562 U.S. 44, 53 (2011)), “the analysis
required pursuant to Chevron [s]tep [t]wo, and that required under the arbitrary and capricious

standard enunciated in State Farm” overlap, EchoStar Satellite L.L.C. v. FCC, 704 F.3d 992,

16

1001-02 (D.C. Cir. 2013) (citation omitted). “Ultimately, under either standard of review, the
relevant question is whether the FDA’s decision represents the result of a reasonable exercise of
its authority.” Amarin Pharm., 106 F. Supp. 3d at 206. This means that, under either standard,
judicial review is “fundamentally deferential_especially with respect to ‘matters relating to [an
agency’s] areas of technical expertise.”’ Fox v. Clinton, 684 F.3d 67, 75 (D.C. Cir. 2012)
(citation omitted). lt does not mean, however, that courts must “simply accept whatever
conclusion an agency proffers.” Tripolz` Rocketry Ass ’n v. Bureau of Alcohol, Tobacco, Firearms
& Explosives, 437 F.3d 75, 77 (D.C. Cir. 2006). ln short, it is the Court’s role to decide whether
the agency acted “within the scope of its lawfi.il authority,” and whether it engaged in “reasoned
decisionmaking,” id., but not to second guess an agency’s reasonable exercise of the authority
that Congress gave it.

In determining that Amneal forfeited the 180-day period of generic exclusivity, the FDA
concluded (1) that Amneal failed “to obtain tentative approval of’ its ANDA “within 30 months
after the date on which the application [was] filed,” see 21 U.S.C. § 355(j)(5)(D)(i)(IV), and (2)
that its failure to do so was not “caused by a change in or a review of the requirements for
approval of the application imposed after the date on which the application [was] filed,” see id.;
see also AR 1055-56. Amneal does not dispute the first conclusion, but it contends that the
second is flawed on multiple levels. Most significantly, Amneal disagrees with the FDA’s
conclusion that the agency’s “demand for commercial-scale data” did not constitute a “change in
a requirement for approval of the application.” Dkt. 25 at 29. lt also argues that the FDA erred
in failing “even [to] consider whether the delay in receipt [of Amneal’s ANDA due to the
typographical error] resulted from a change in or review of approval requirements,” id., and

erred in failing to treat “all of the activity regarding whether the [a]gency would receive the

17

application” as a “‘change in or [a] review of . . . a ‘requirement[] for approval of the
application,”’ id. at 42 (quoting 21 U.S.C. § 355(j)(5)(D)(i)(IV)). Finally, Amneal devotes the
lion’s share of its brief to arguing that the FDA erred in concluding in the altemative, that even
if either of these regulatory actions constituted a change in a requirement for approval of the
ANDA, neither action “caused” Amneal to miss the 30-month deadline. Dkt. 25 at 23-38.

As explained below, the Court holds that the FDA did not violate the plain terms of the
statute or act unreasonably in deciding that neither of the regulatory steps identified by Amneal
constituted “a change in or review of [a] requirement[] for approval of” Amneal’s ANDA. The
Court, in contrast, does have significant doubts about the FDA’s conclusion that the delay
occasioned by its request for the commercial-scale data was not a cause of Amneal’s failure to
meet the deadline. But, even if Amneal were correct about that portion of the FDA’s decision, it
would not save Amneal’s claim; Amneal can prevail only if it can show that the FDA erred in
declining to treat one or both of the identified regulatory steps as a change in or review of a
requirement for approval and that the FDA’s causation analysis was flawed. Because Amneal
fails to clear the first of these hurdles, the Court need not decide whether it has cleared the
second.

A. Request for Commercial-Scale Data

The principal dispute between the parties turns on whether the FDA’s request that
Amneal supplement its ANDA by submitting data from a commercial-scale batch constituted “a
change in . . . the requirements for approval of” Amneal’s ANDA. Although one might think
that it would be relatively straightforward to identify “the requirements for approval of the
application” to market a generic version of a drug, all agree that more is required than simply

flipping through the statute, regulations, and policies. To be sure, a change in the statute,

18

regulations, or policies would generally be sufficient to trigger the proviso. But the FD_A
acknowledges that other steps can alter “the requirements for approval of ’ ANDAS. An ANDA
applicant, for example, must demonstrate that the labeling it proposes to use is the same as the
labeling for the listed drug, and, accordingly, if the label of the listed drug changes, the FDA will
typically require that ANDA applicants submit new, proposed labels. Dkt. 33 at 25. FDA
concedes that such a scenario represents a change in “the requirements for approval.” Similarly,
when the FDA changes or reviews the types of study designs that it recommends to establish
bioequivalence, that process may delay ANDA applicants who rely on that guidance Ia’. Again,
FDA concedes that such a change would trigger the proviso.

What all of these circumstances have in common, according to the FDA, is that the rule,
guidance, or practice at issue applies generally to other similarly-situated ANDA applicants.
That is, according to the FDA, “requirements for approval” are those steps or efforts that the
“FDA would demand from any and all applicants submitting an ANDA for the drug at issue.”
Ia’. at 22. The FDA does not, in contrast, construe the phrase “requirements for approval” to
encompass occasions when the FDA requests additional data or information from a specific
applicant to address a specific deficiency noted in the course of the agency’s scientific review of
the ANDA. Id.

Amneal, for its part, takes a broader view of what it means to change a “requirement[] for
approval of” an ANDA. It agrees that the types of changes of general applicability identified by
the FDA meet the terms of the proviso. But it would also include what it characterizes as
“changes applied to particular applications.” Dkt. 25 at 43. Amneal concedes, as it must, that
this does not mean that every request for additional data or information that emerges as part of

the FDA review process constitutes a change in the relevant requirements; all agree that the

19

review process is an iterative one “in which there is inevitably and invariably back and forth”
requiring applicants to provide additional information. Dkt. 69 at 8 (Tr. 816-7). What separates
this case from those garden-variety requests for additional information, according to Amneal, is
that the FDA’s request in this case for data from a commercial-scale batch was different in kind
from the request for pilot-scale batch data that the FDA required at the time Amneal filed its
ANDA. Dkt. 25 at 40. The FDA may not have changed “what” Amneal was required to
demonstrate_-that the production methods “are adequate to assure and preserve [the drug’s]
identity, strength, quality, and purity,” AR 1055; see also 21 U.S.C. § 355(b)(1)(D),
(j)(2)(A)(vi)-but it changed “how” Amneal was required to show that it satisfied that
standard_by using a commercial-scale batch as opposed to pilot-scale batches. Dkt. 38 at 11;
Dkt. 69 at 15 (Tr. 15:16-22), id. at 26 (Tr. 26:11-15).

Understanding what Amneal means by this, however, requires consideration of what, if
anything, actually changed from when Amneal filed its ANDA and when the FDA approved it.
Amneal asserts that, before it submitted its ANDA, the FDA had “stated [that] it [would]
generally receive and approve applications containing only pilot-scale data,” Dkt. 25 at 40
(emphasis added), and that “[i]t is undisputed that [the] FDA’s public policies required only the
submission of pilot-scale data,” Dkt. 38 at 12. In support of this contention, Amneal cites to five
FDA guidance documents. See Dkt. 25 at 17 & n.2; Dkt. 38 at 12; Dkt. 68 at 32 (Tr. 32:8-9).
Although each of those documents supports Amneal’s contention that the FDA generally

requires only pilot-batch data for purposes of receipt of an ANDA as “substantially complete”

20

under 21 C.F.R. § 314.101(b), none of them supports Amneal’s further contention that the FDA
has stated that pilot-batch data is sufficient to support approval of an ANDA.3

The Court does not doubt that the type of information required for the FDA to receive an
ANDA will often mirror the type of information required to approve it. But, the receipt of a
“substantially complete” application only begins the process, which inevitably requires the
submission of additional information See Dkt. 69 at 8 (Tr. 8:7). As the FDA put it in its
decision letter, the “Agency’s requirements in 21 C.F.R. [§] 314.101 as to whether an ANDA is
sufficiently complete to permit substantive review, and thus may be received by the Agency, are
distinct from the question as to whether the application meets the requirements for approval set
forth in 21 C.F.R. § 314.127.” AR 1055 n.44. And, the approval “requirements” make clear that
the FDA will “refuse to approve an ANDA” if “[t]he methods used in, . . . and controls used for,
the manufacture, processing, and packaging of the drug product are inadequate to ensure and [to]

preserve its identity, strength, quality, and purity.” 21 C.F.R. § 314.127(a)(1) (emphasis added).

3 See ANDA Submissions - Refuse-to-Receive Standards l (May 2015), AR 243 (stating that
the guidance “highlights deficiencies that may cause FDA to refuse to receive an ANDA”
(emphasis added)); Guidance on the Packaging of Test Batches 1, 5 (Sept. 2012), AR 44, 48
[hereinafter Guidance: Test Batches] (noting that “[t]he minimum amount to be packaged is
100,000 units” and explaining that “[i]t is critical that all testing be conducted on samples that
represent the entire batch and mimic the product which will be marketed post-approval”
(emphasis added)); QlA(R2) Stability Testing of New Drug Substances and Products 8 (Nov.
2003), https://www.fda.gov/downloads/drugs/guidances/ucm073369.pdf [hereinafter Guidance:
Stability Testing] (noting that “two of the three batches should be at least pilot scale batches”
and also that “[t]he manufacturing process used for primary batches should simulate that to be
applied to production batches”); ANDAS: Stability Testing of Drug Substances and Products,
Questions and Answers 3, 8 (May 2014),
https://www.fda.gov/downloads/drugs/guidances/ucm366082.pdf (indicating what data “should
[be] submit[ted],” with the application); ANDAs: Stability Testing of Drug Substances and
Products 2 (June 2013), https://www.fda.gov/downloads/drugs/guidances/ucm320590.pdf
(explaining what data “should be provided” with the application).

21

The same guidance documents that Amneal cites, moreover, inform ANDA applicants
that “[i]t is critical that all testing be conducted on samples that represent the entire batch and
mimic the product which will be marketed post-approval.” Guidance: Test Batches at 1
(emphasis added); see also Guidance: Stability Testing at 8 (“The manufacturing process used
for primary batches should simulate that to be applied to production batches and should provide
product of the same quality and meeting the same specification as that intended for marketing”);
id. at 18 (defining “pilot scale batch” as “[a] batch of drug product manufactured by a procedure
fully representative of and simulating that to be applied to a full production scale batch”). That
admonition is apt here because Amneal proposed using a significantly different manufacturing
process for “the product which [would] be marketed post-approval” than it used for the pilot
batches that it submitted with its ANDA. Amneal’s pilot batches, as explained above, were
produced by combining - of lR (that is, immediate release) beads with - of ER (that
is, extended release) beads, yielding 150,000 capsules. For commercial purposes, however,
Amneal proposed producing batches of 1,000,000 capsules by combining up to two lots of IR
beads with up to six lots of ER beads. It was this difference in production techniques that, at
least in substantial part, prompted the FDA to request data from a commercial-scale batch to

address concerns about the _ required to achieve the desired rate of

release of the drug

Against this backdrop, Amneal’s proposed reading of the statute must be clarified as
follows: A change in the requirements for approval occurs when the FDA allows for the receipt
of an ANDA that contains a certain type of information-here, data from pilot-scale batches__
and, then, in the course of its technical review of the application, the FDA concludes that a

different type of information-_here, data from a commercial-scale batch_is necessary for it to

22

approve the ANDA. To this, Amneal adds that the change here was a significant one, requiring
for example, substantially larger quantities of the active ingredient, AR 654, and that “across a
portfolio of 95 approved ANDAS, Amneal ha[d] never previously been asked to produce a
commercial-scale lot as a condition of approval,” AR 1056. With this clarification in mind, the
Court turns to the Chevron framework.

1. Chevron Step One

Under Chevron step one, the Court must consider “whether Congress has directly spoken
to the precise question at issue.” 467 U.S. at 842. If so, the Court “must give effect to the
unambiguously expressed intent of Congress.” Id. at 843. ln making this determination, the
Court applies the “traditional tools of statutory construction,” id. at 843 n.9, including looking to
“the text, structure, and the overall statutory scheme, as well as the problem Congress sought to
solve,” Financial Planning Assoc. v. SEC, 482 F.3d 481, 487 (D.C. Cir. 2007), in order “to
determine whether Congress has ‘unambiguously foreclosed the agency’s statutory
interpretation.”’ Vill. of Barrington, lll. v. Surface Transp. Bd., 636 F.3d 650, 659 (D.C. Cir.
2011) (quoting Catawba Cty., N.C. v. EPA, 571 F.3d 20, 35 (D.C. Cir. 2009)).

The Court starts, as it must, with the text of the statute, See Sebelius v. Cloer, 569 U.S.
369, 376 (2013). Section 355@)(5)(D)(ii) provides that “[t]he 180-day exclusivity period . . .
shall be forfeited by a first applicant if a forfeiture event occurs with respect to that first
applicant.” 21 U.S.C. § 355(]`)(5)(D)(ii). The statute then specifies a series of “forfeiture
events,” only one of which is relevant in this case. That “forfeiture event” occurs if:

The first applicant fails to obtain tentative approval of the application within 30

months after the date on which the application is filed, unless the failure is caused

by a change in or a review of the requirements for approval of the application
imposed after the date on which the application is jiled.

23

Id. § 355(]`)(5)(D)(i)(lV) (emphasis added). For present purposes, then, the question is whether
this statutory text speaks directly to the question whether the FDA’s request for data from a
commercial-scale batch constituted “a change in . . . the requirements for approval of” Amneal’s
ANDA. Amneal contends that its interpretation is compelled, thus foreclosing the FDA’s
competing interpretation See Vill. of Barrington, 636 F.3d at 659. The Court disagrees.

Although the relevant statutory terms are not defined, they are not technical or
uncommon terms. A “requirement” is “a condition which must be complied with,” XIII Oxford
English Dictionary 682 (2d ed. 1989), and a “change” is the “substitution of one thing for
another,” Ill Oxford English Dictionary at 15. Understanding the meaning of those words,
however, does little to resolve the parties’ dispute. As noted above, all agree that not every
request for additional information-_even if the FDA will not approve the ANDA without that
information_constitutes a “change” in the governing “requirements.” What makes this case
different, in Amneal’s view, is that the FDA required one type of data as a condition of receipt of
its ANDA_data from pilot-scale batches_and, then, required a dijferent type of data as a
condition of approval of the ANDA_data from a commercial-scale batch. “In other words,”
according to Amneal, “at the time of the ANDA’s submission [the] FDA [did] not require
submission of commercial-scale data and only [made] that determination later in the review
process. That’s a change by any definition.” Dkt. 25 at 40.

ln support of its contention that this “change” falls within the plain terms of the statute,
Amneal argues that the statute “is not limited to broader . . . changes” in the statute or in FDA
policy, but also includes “changes applied to particular applications.” Id. at 43 (emphasis
added). The central question under Amneal’s view of the law, accordingly, is whether the

applicant faced a different set of hurdles before and “after the date on which [its] application was

24

filed.” 21 U.S.C. § 355(]`)(5)(D)(i)(IV). Evidence of the statutory focus on the hurdles faced by
the applicant in the context of the FDA’s review of its particular ANDA is found, according to
Amneal, in the statute’s references to the requirements for approval “of the application” that are
imposed after “the application” is filed. Dkt. 25 at 43-44. If the FDA were right, Amneal adds,
and the focus was instead on changes in generally applicable rules and policies, the words “of the
application” would be rendered meaningless, thus violating the canon against surplusage Id. at
44.

The Court, of course, agrees that statutes should be construed, where possible, to avoid
surplusage See Indep. Ins. Agents of Am., Inc. v. Hawke, 211 F.3d 638, 644-45 (D.C. Cir.
2000). It is unpersuaded, however, that the FDA’s reading of the statute runs afoul of that canon
Under the canon against surplusage, the Court must ask whether, on the FDA’s reading the
statute, the phrase “of the application” could be dropped from that statute with no consequence
lt cannot. To the contrary, under both the FDA’s and Amneal’s proposed constructions of the
statute, those words perform the same, essential function: they are the object of the phrase “the
requirements for approval,” and, without them, the reader would be left to ask, “approval of
what?” The phrase “of the application,” in short, is needed to complete the sentence

Amneal’s construction of the statute, in contrast, either reads the phrase “approval of” out
of the statute or, at a minimum, gives that phrase a strained meaning Amneal presumes that the
requirements for the FDA to receive an ANDA mirror the requirements for approval. But the
FDA has explained-and Amneal has not identified any evidence to the contrary_that the two
requirements, even if closely related, are distinct. AR 1055 n.44. Amneal’s textual argument,
however, conflates the two, and, in doing so, it sidesteps the critical question whether the data

from the pilot-scale batches that it submitted with its ANDA was ever sufficient (before or after

25

the ANDA was filed) to support approval of the application, Understood in this light, nothing in
the statutory text supports Amneal’s reading much less unambiguously forecloses the FDA’s
approach. See Vill. of Barrington, 636 F.3d at 659.

Nor is the Court convinced that the distinction that Amneal draws between “what” the
applicant must demonstrate and “how” it must do so has any bearing on the meaning of the
statute The FDA, for its part, does not dispute that either type of change may constitute a
“change in . . . the requirements for approval of” the ANDA. As noted above, however, the
question is not whether the requirements for receipt of the ANDA were different from the
requirements for approval of the ANDA; the question is whether the FDA changed the applicable
“requirements for approva ” after the ANDA was filed. Whether cast as a change in what
Amneal was required to demonstrate or how it was required to do so, Amneal cannot prevail in
the absence of evidence that the FDA changed the requirements for approval of Amneal’s
ANDA.

Amneal also argues that its construction of the statute is necessary to achieve the
statutory goal of “ granting exclusivity to an applicant whose delay was caused by change in the
review requirements ‘of the application”’ Dkt. 25 at 44-45. According to Amneal, “[i]f the
FDA is correct, and a substantial change in expectations like a post-filing demand for
commercial-scale data is not a change sufficient to trigger the statute, then the aims of the statute
will be thwarted.” Id. at 45. “An ANDA sponsor hoping to ensure 180-day exclusivity,”
Amneal further asserts, would have significant reason under the FDA’s reading of the statute to
wait to submit its ANDA “until it has fully scaled up to commercial production and obtains all of
the requisite commercial-scale data.” Id. And that, according to Amneal, would then delay

generic entry into the market, contrary to the aims of the Hatch-Waxman Act. Id.

26

This argument might more appropriately be raised under Chevron step two, but it is, in
any event, unconvincing Amneal is correct that Congress enacted the Hatch-Waxman Act to
promote generic competition, T eva Pharm., 595 F.3d at 1304, and that Congress included the
180-day exclusivity period to “reward . . . generic[] [manufacturers] that stick out their necks (at
the potential cost of a patent infringement suit) by claiming that patent law does not extend the
brand maker’s monopoly as long as the brand maker has asserted,” id. at 1318. But by including
the 30-month forfeiture rule in the statute, Congress conditioned that “reward” on prompt ANDA
approval to “ensure that the 180-day exclusivity period enjoyed by the first generic to challenge
a patent cannot be used as a bottleneck to prevent additional generic competition.” Hi-Tech
Pharmacal Co., 587 F. Supp. 2d at 4 (quoting 149 Cong. Rec. S15746 (daily ed. Nov. 24, 2003)
(statement of Sen. Schumer)); see also Mylan Labs., 910 F. Supp. 2d at 311.

Even assuming that Amneal is right that its reading would provide additional incentives
for generic manufacturers promptly to file ANDAs challenging the patents of the listed drugs_
and that proposition is far from certain_it would do so by encouraging applicants to file as soon
as they can satisfy the threshold requirements for receipt of the ANDA under 21 C.F.R. §
314.101, regardless of whether they are able to satisfy the requirements for approval. That, then,
would risk creating the type of bottleneck that Congress enacted the 30-month rule to prevent.
And, at the same time, it would engender increased uncertainty_and, perhaps, litigation_about
whether each new request for information in the FDA review process is sufficiently “different in
kind” from the information required as a condition of receipt of the ANDA that the request
constitutes “a change in . . . the requirements for approval.” The statutory scheme, in short,
reflects a balance between creating an incentive for a generic manufacturer to be the first to

challenge the branded manufacturer’s patents and preventing bottlenecks that could delay

27

additional generic competition Amneal’s argument, however, focuses on one side of this
balance, to the exclusion of the other.

The Court, accordingly, concludes that nothing in the text, structure or purposes of the
statute forecloses the FDA’s construction of the statute and that, indeed, the FDA’s construction
better coheres with the terms of the statute than the reading that Amneal posits.

2. Chevron Step T wo and Arbitrary and Capricious Review

Amneal’s claim also fails under Chevron’s second step and the APA’s arbitrary and
capricious standard. “At Chevron step two,” the Court “defer[s] to the agency’s permissible
interpretation, but only if the agency has offered a reasoned explanation for why it chose that
interpretation.” Vill. of Barrington, 636 F.3d at 660. Like the Chevron step two analysis,
moreover, the APA’s “arbitrary and capricious” inquiry asks “whether an agency’s actions under
a statute are unreasonable” Am. Fed ’n of Gov ’t Emps., AFL-CIO, Local 46 v. Nicholson, 475
F.3d 341, 355 (D.C. Cir. 2007) (intemal quotation marks omitted); see also, e.g., Agape Church,
Inc. v. FCC, 738 F.3d 397, 410 (D.C. Cir. 2013) (“The analysis . . . under Chevron [s]tep [t]wo
and arbitrary and capricious review is often the same, because under Chevron step two, the court
asks whether an agency interpretation is arbitrary or capricious in substance.” (intemal quotation
marks and brackets omitted)). The arbitrary and capricious standard goes a step further,
however, and requires consideration, more generally, of whether the agency action is “supported
by ‘reasoned decisionmaking”’ Tripoli Rocketry, 437 F.3d at 77 (citations omitted). The “result
must be logical and rational,” id., as well as “adequately explained” and “coheren[t],” Fox, 684
F.3d at 75 (citations and internal quotation marks omitted). The Court must evaluate the
agency’s decision, moreover, solely “on the basis articulated by the agency itself.” State Farm,

463 U.S. at 50. Although lack of “clarity” alone is not a sufficient basis to strike down an

28

agency decision, Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. 281, 2854
86 (1974), neither counsel nor the Court may amend or supply the agency’s rationale, see Rijj’i`n
v. Surface Transp. Bd., 592 F.3d 195, 198 (D.C. Cir. 2010).

Amneal contends that the FDA’s decision fails Chevron step two and the arbitrary and
capricious standard on two grounds: First, it argues that the decision was premised, at least in
part, on “secret, nonpublic policies.” Dkt. 25 at 45 . Second, it contends that the decision “is

inexplicably inconsistent with FDA’s prior precedent.” Id. The Court is not convinced by either

argument.

a. Nonpublic Policy

Amneal first argues that the FDA relied on a “secret” or “nonpublic policy” in its
forfeiture decision FDA’s forfeiture decision asserts that “[s]ince at least 2010, the [a]gency has
requested that applicants with certain complex drug products and/or a complex manufacturing
process manufacture commercial scale batches to support approval of the ANDA by
demonstrating that the applicant will be able to successfully scale-up its manufacturing process
upon approval.” AR 1055. The decision acknowledges, however, that this precedent was not
publicly available because the “FDA is limited in disclosing information concerning other
applications under [its] disclosure regulations.” Id. at 1055 n.43 (citing 21 C.F.R. § 314.430).

The decision continues:

Amneal . . . represents that, across a portfolio of 95 approved ANDAS, Amneal has
never previously been asked to produce a commercial-scale lot as a condition of
approval. As described above, FDA’s practice has been to ask for pre-approval
production of commercial-scale batches in very specific circumstances: when a
drug is complex or has a complex manufacturing process, and the evidence suggests
that the applicant does not have a good understanding of the product and
manufacturing process and lacks appropriate in-process controls. Thus, it is not
surprising that FDA’s general policies on manufacturing test batches do not
describe these specific circumstances, nor it is surprising that Amneal believes it
has not been previously asked to provide a commercial-scale test batch to support

29

approval Amneal’s unfamiliarity with the case-specific requirement does not

mean it reflects a “change in or review of the requirements for approval” under

section 505(]`)(5)(D)(i)(lV) of the [FDCA].

AR 1056. Relying on the fact that the FDA’s prior requests for commercial-scale data were not
publicly available, Amneal argues that it was the victim of “unfair surprise,” Dkt. 25 at 46-47
(quoting Christopher v. SmithKline Beecham Corp., 567 U.S. 142, 156 (2012)). That is,
according to Amneal, it was “arbitrary and capricious for the [a]gency to . . . find a forfeiture of
exclusivity where [Amneal’s] failure to obtain approval was caused by application of a secret
policy that could not have been anticipated, and which was disclosed after the application had
already been filed.” Id. at 47. This contention suffers from three flaws.

First, Amneal overstates the premise of its argument; the FDA did not adopt a “secret
policy” that it merely revealed after Amneal filed its ANDA. To the contrary, the forfeiture
decision makes clear that the FDA’s prior requests for data from commercial-scale batches were
“application specific” requests “made during the substantive review of the ANDA by the
chemistry reviewers,” based on the applicant’s failure to “demonstrate a good understanding of
its product and manufacturing process” and failure to “have appropriate in-process controls.”
AR 1055. Thus, rather than constituting a “secret policy,” the FDA’s prior requests merely
represented case-by-case evaluations of whether data from a commercial-scale batch was needed
to address a case-specific deficiency.

Second, as previously noted, pre-existing FDA guidance put ANDA applicants on notice
that it was “critical that all testing be conducted on samples that represent the entire batch and
mimic the product which will be marketed post-approval.” Guidance: Test Batches at 1
(emphasis added); AR 44. Amneal, accordingly, had “fair warning” Christopher, 567 U.S. at

156, that the FDA’s chemical reviewers would conclude that the production data that Amneal

30

submitted with its application-involving the mixing of - of IR beads with - of ER
beads_was insufficient to demonstrate that the process that Amneal proposed to use for
commercial purposes_involving the mixing of two lots of IR beads with six lots of ER beads~
would assure the “identity, strength, quality, and purity” of the product, 21 U.S.C. §
3550)(4)(A);-

Amneal does not dispute that its initial submission was deficient, but contends that the
cure that the FDA demanded (data from a commercial-scale batch) came as a surprise Dkt. 69
at 16 (Tr. 16:13-19). lt argues, for example, that instead of requiring data from a commercial-
scale batch, the FDA might have asked that Amneal submit “a pilot batch that better
mimic[ked]” the product that Amneal proposed to market Id. at 28 (Tr. 281 10). That issue was
not raised before the FDA and therefore is not properly before the Court See Nuclear Energy
Inst., Inc. v. EPA, 373 F.3d 1251, 1297 (D.C. Cir. 2004) (per curiam). And, even if the Court
could reach the issue, it would owe substantial deference to the scientific judgment of the FDA
regarding what type of testing is necessary to “mimic the product which will be marketed post-
approval” and whether less onerous or time-consuming approaches were available See Marsh v.
Or. Nat. Res. Council, 490 U.S. 360, 377 (1989) (“Because analysis of the relevant documents
requires a high level of technical expertise, we must defer to the informed discretion of the
responsible agencies.” (intemal quotation marks and citation omitted)); Serono Labs., 158 F.3d
at 1320 (explaining that when “[t]he FDA’s determination” rests on its “evaluations of scientific
data within its area of expertise, [it] is entitled to a high level of deference from th[e] court”
(intemal citations and quotation marks omitted)).

Third, Amneal is not challenging the FDA’s request for data from a commercial-scale

batch, but rather the agency’s forfeiture decision The relevant question, accordingly, is not

31

whether Amneal was surprised by the FDA’s request for the data, but whether the agency’s
forfeiture decision was arbitrary and capricious. That question must be considered in light of the
statutory standard_whether there was “a change in . . . the requirements for approval of”
Amneal’s ANDA-and, for the reasons discussed above, the Court concludes that the FDA
reasonably concluded that there was not ln short, Amneal was required to demonstrate that its
manufacturing processes and controls were adequate to assure the “identity, strength, quality,
and purity” of its product, 21 U.S.C. § 355(j)(4)(A), and it was required to produce testing on
samples that would “mimic the product which [would] be marketed post-approval,” Guidance:
Test Batches at 1; AR 44. The fact that the FDA noted a deficiency and requested data that it
concluded_in its expert judgment_was needed to meet that standard did not cross the Rubicon
of the arbitrary and capricious standard. Amneal’s argument to the contrary stands on its
statutory argument_that the exception focuses on “application-specific requirements,” Dkt. 25
at 44-and fails for the same reason its statutory argument fails.
b. lncor'lsisteiicy

Amneal also argues that the FDA’s interpretation is inconsistent with the agency’s past
decisions. ln support of this contention, Amneal points to (1) a 2010 FDA letter decision
excusing a failure by Nycomed U.S., lnc. to obtain approval within 30 months, id. at 47-48; (2) a
similar decision in the case of Synthon Pharmaceuticals, lnc., id. at 48 n.14; (3) another similar
decision in the case of Sandoz, lnc., id.; and, finally, (4) FDA guidance explaining that “changes
in a [listed] drug’s labeling or formulation that require . . . an applicant to conduct additional

testing” may constitute a change in the requirements for approval, id. at 47.4 Amneal fails to

4 Amneal also points to the FDA’s decision concluding the Teva Pharmaceuticals was excused
for failing to obtain approval for its lrbesartan Tablets USP ANDA within 30 months, Dkt. 25 at

32

identify any evidence, however, that the FDA applied a standard in any of those cases that
differs from the standard the FDA proffers here: the exception to the 30-month requirement
applies to changes in how all similarly-situated ANDA applicants must demonstrate that their
ANDAs warrant approval, and it does not apply to requests for additional information prompted
by deficiencies Specific to the particular ANDAs. Dkt. 33 at 24.

Nycomed, for example, sought to market a 5% lmiquimod Cream and failed to obtain
approval within 30 months. The FDA nonetheless concluded that Nycomed qualified for the
180-day market exclusivity because the delay “was caused by the agency’s ongoing review of
the requirements for approval of lmiquimod Cream, 5%.” Letter from FDA to Nycomed, U.S.,
lnc. at 2 (Feb. 25, 2010).5 As the FDA has explained, that review involved “the study designs
recommended to establish bioequivalence with” the listed drug Dkt. 33 at 25 (citing Draft
Guidance on lmiquimod Cream, 5% (Feb. 2010)).6 As a result, unlike the present case, “the
change in or review of the requirement[] for approval” was not prompted by a deficiency noted
in Nycomed’s application, and it applied to all similarly-situated applicants

The Synthon decision cited by Amneal is distinguishable on similar grounds. As

previously discussed, an ANDA applicant is required to demonstrate that its proposed label is the

48, but the FDA’s decision letter merely asserts that the agency “changed the approval
requirements for Teva’s proposed product” and that, “[a]s a result, Teva was required to perform
additional testing and [to] include an additional drug substance specification prior to approval,”
Letter from FDA to Teva Pharmaceuticals USA at 2 n.l (March 30, 2012),
http://www.accessdata.fda.gov/drugsatfda_docS/appletter/ZO l 2/0771 59s0001tr.pdf. Nothing
contained in this brief discussion of the issue, or in any other materials identified by Amneal,
explains how the Teva case supports Amneal’s claim.

5 Available at https://www.accessdata.fda.gov/drugsatfda_docs/appletter/Z010/
078548s0001tr.pdf.

6 Available at https://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatory
lnformation/Guidances/UCM201988.pdf.

33

same as the proposed label for the listed drug See 21 U.S.C. § 355(]')(2)(A)(v); 21 C.F.R. §§
314.94(a)(8)(iii), 314.127(a)(7). Thus, if the label for the listed drug is amended, ANDA
applicants must submit new proposed labels for their products. Such a change delayed the
approval of Synthon’s ANDA, and, consistent with the FDA’s reading of the statute, the agency
concluded that Synthon had not forfeited its eligibility to the 180 days of exclusivity:

The agency has determined that there was a change in the requirements for approval

of this ANDA. Specifically, the labeling of the [listed drug] changed after

submission of the ANDA. The agency has also determined that this change was

the cause of your not obtaining tentative approval of the ANDA within 30 months

after the date on which it was filed.
Letter from FDA to Synthon Pharmaceuticals Inc. at 2 n.l (March 30, 2012).7

The same is true for the Sandoz decision Like the labeling requirement, changes to the
formulation of the listed drug requires ANDA applicants either to seek approval for their generic
version based on the new formulation or to seek a determination that the old formulation was not
withdrawn for safety reasons. See Guidance for lndustry (Draft) - 180-Day Exclusivity:
Questions and Answers 23 (Jan. 2017)8 [hereinafter Draft Guidance: 180-Day Exclusivity]. ln
the Sandoz decision, the FDA explained that “[w]hen a change in formulation for a listed drug

referenced requires an ANDA applicant to respond[,] . . . we will consider this a ‘change in or

review of the requirements for approval’ within the meaning of [the FDCA].” Letter from FDA

7 Available at http://www.accessdata.fda.gov/drugsatfda_docs/appletter/2010/
090229s0001tr.pdf.

8 Available at http://www.fda.gov/downloads/Drugs/GuidanceCompliance
Regulatorylnformation/Guidances/UCM536725.pdf.

34

to Sandoz at 9 (Sept. 20, 2011).9 Thus, this example also fails to reveal any inconsistency in
administrative practice

Finally, the FDA guidance that Amneal references is to the same effect. As with the
preceding examples, that guidance merely affirms that changes in the listed drug’s labeling or
formulation require all ANDA applicants seeking to market a generic version of the listed drug
to follow suit. See Draft Guidance: 180-Day Exclusivity at 23. For the reasons explained above,
that rule is also consistent with the FDA’s reading of the exception, and nothing it did with
respect to Amneal’s ANDA or request for exclusivity is at odds with the guidance

Accordingly, the FDA’s conclusion that its request for commercial-scale data did not
constitute a “change in . . . requirements for approval” such that Amneal was exempt from the
30-month forfeiture date was both reasoned and consistent with agency precedent
B. Receipt of Amneal’s ANDA

Amneal’s other theory of relief requires only brief discussion Amneal argues that the
FDA’s “consideration of whether to receive Amneal’s application notwithstanding a
typographical error” also constituted a “change in or review of the requirements for approval” of
its ANDA. Dkt. 25 at 41. “lt is a hard and fast rule of administrative law, rooted in simple
faimess, that issued not raised before an agency are waived and will not be considered by a court
on review.” Nuclear Energy Inst., 373 F.3d at 1297. Accordingly, because Amneal did not raise
this issue before the FDA, see AR 63 8_66, it may not do so now.

Moreover, even had Amneal preserved the issue, it would not advance Amneal’s cause

9 Available at https://www.regulations.gov/contentStreamer?documentld=FDA-2010-P-0632-
001 7&attachmentNumber=1 .

35

There is no dispute that established FDA guidance counseled Amneal that ANDAs must contain
84 days of accelerated stability data, nor is there any doubt that Amneal’s ANDA contained a
typographical error that led the FDA reviewers to conclude that the ANDA lacked the required
data. AR 63. That typographical error delayed the FDA’s acceptance of the ANDA. But, the
error was Amneal’s, and it was not the product of any “change in or review of the requirements
for approval of” the ANDA. lndeed, the only possible basis for arguing otherwise returns to
Amneal’s contention that the exception to the 30-month rule applies in every case in which the
ANDA sponsor faces an unanticipated hurdle after delivering its application to the FDA. As
explained above, that is not the law.
C. Remaining Issues

Having concluded that Amneal has failed to identify a relevant change in the
requirements for approval of its ANDA, the Court need not reach Amneal’s further argument
that the FDA misconstrued the phrase “caused by” and impermissibly imposed a rigid, single-
factor causation test. Dkt. 25 at 23_38. The Court recognizes that the FDA’s causation test is
difficult to square with the statutory text, but Amneal can prevail here only by demonstrating that
the FDA misconstrued or misapplied both the “requirement” prong and the “caused by” prong of
the proviso. Because Amneal has not cleared the first hurdle, the second hurdle is beside the
point

Finally, having concluded that Amneal’s case fails on the merits and that the FDA is
entitled to summary judgment, the Court need not decide whether Amneal is entitled to a
preliminary injunction or temporary restraining order. “The purpose of a preliminary injunction
is merely to preserve the relative positions of the parties until a trial on the merits can be held.”

Univ. ofTex. v. Camenisch, 451 U.S. 390, 395 (1981). lt follows that, “[i]n the absence of a

36

pending claim for relief, there is no basis for the Court to issue an order designed to maintain the

status quo while the merits of the dispute are resolved.” Justice v. Koskinen, 109 F. Supp. 3d

142, 151 (D.D.C. 2015). Amneal’s motion for a preliminary injunction or, in the altemative, for

a temporary restraining order is, accordingly, now moot.

CONCLUSION
For the reasons set forth above, the Court will deny Amneal’s motion for summary

judgment, Dkt. 25, grant the FDA’s and Lupin’s cross-motions for summary judgment, Dkt. 30;
Dkt. 34, and deny Amneal’s motion for a preliminary injunction or, in the altemative, a
temporary restraining order, Dkt. 62, as moot

A separate order will issue

tsi Randolp h D. Moss
RANDOLPH D. MOSS
United States District Judge

Date: January 23, 2018

37