Reply Brief — Teva Pharmaceuticals USA, Inc., Petitioner v. GlaxoSmithKline LLC, et al.

No. 22-37

IN THE

Supreme Court of the United States

TEVA PHARMACEUTICALS USA, INC.,

v.

Petitioner,

GLAXOSMITHKLINE LLC, ET AL.,

Respondents.

On Petition for a Writ of Certiorari to the United

States Court of Appeals for the Federal Circuit

REPLY BRIEF FOR PETITIONER

DARYL L. WIESEN

CHRISTOPHER T. HOLDING

ELAINE HERRMANN BLAIS

ROBERT FREDERICKSON, III

GERARD J. CEDRONE

WILLIAM E. EVANS

GOODWIN PROCTER LLP

100 Northern Avenue

Boston, MA 02210

WILLIAM M. JAY

Counsel of Record

GOODWIN PROCTER LLP

JAIME A. SANTOS

1900 N Street, NW

Washington, DC 20036

(202) 346-4000

wjay@goodwinlaw.com

September 7, 2022

Counsel for Petitioner

RULE 29.6 STATEMENT

The corporate disclosure statement included in the

petition remains accurate.

i

TABLE OF CONTENTS

Introduction ................................................................ 1

Argument .................................................................... 2

I.

The decision below upends settled legal

principles and an important legislative

compromise. .................................................... 2

A. The Federal Circuit eviscerated an

essential element of inducement

law. ........................................................... 2

B. The Federal Circuit’s rule is

inconsistent with the HatchWaxman regime. ...................................... 4

II. This issue warrants the Court’s prompt

attention. ......................................................... 6

A. The issue has not been “mooted” by

subsequent regulations. ........................... 7

B. GSK ignores the unpredictability

that the Federal Circuit’s opinion

engenders. ................................................ 9

III. This case presents an ideal opportunity

to address these important issues................ 10

Conclusion ................................................................. 13

ii

TABLE OF AUTHORITIES

Cases

Bayer Schering Pharma AG v. Lupin, Ltd.,

676 F.3d 1316 (Fed. Cir. 2012)............................... 3

Caraco Pharm. Labs., Ltd. v. Novo Nordisk

A/S,

566 U.S. 399 (2012) ................................................ 6

HZNP Medicines LLC v. Actavis Labs. UT, Inc.,

940 F.3d 680 (Fed. Cir. 2019)................................. 3

Life Technologies Corp. v. Promega Corp.,

137 S. Ct. 734 (2017) .............................................. 3

Takeda Pharms. U.S.A., Inc. v. W.-Ward

Pharm. Corp.,

785 F.3d 625 (Fed. Cir. 2015)................................. 3

Statute

35 U.S.C. § 271(b) ........................................................ 2

Regulations

21 C.F.R. § 314.53(b)(1) (2003) .................................... 7

68 Fed. Reg. 36,676 (Jun. 18, 2003) .................... 5, 6, 8

81 Fed. Reg. 69,580 (Oct. 6, 2016)........................... 5, 8

iii

1

INTRODUCTION

The decision below erodes foundational principles

of patent law while upsetting a careful legislative

compromise that ensures ready access to low-cost generic medicines. As Teva, amici, and countless commentators have explained, the resulting upheaval will

harm patients, taxpayers, and the public at large.

Not to worry, GSK insists—generic manufacturers

can still claim “the carve-out statute’s protection” if

they “fully and truly carve out” patented indications.

Opp.33. But there’s the rub: under the Federal Circuit’s decision, the carve-out statute provides no protection, not even to a company that follows FDA’s directions and carves out everything the brand manufacturer identifies.

The facts here are no aberration. Teva did what

generic applicants routinely do—and Congress authorized—when a drug’s active ingredient is no longer

patented: it adopted an FDA-drafted “skinny label”

that carved out the one indication GSK identified as

patented in its sworn representations to FDA. Seven

other generics used the same skinny label. Nevertheless, the Federal Circuit held that a jury could find

Teva liable for “actively” inducing infringement

through information scattered across different portions of its label—portions GSK never identified to

FDA. That is more than just a misapplication of settled precedent: it is an about-face, because it adopts

what was previously a dissenting view and allows a

jury to find active inducement in virtually any carveout case.

Contrary to GSK’s attempt to grandfather this

case, the regulatory obligations that FDA applies

2

today are materially identical to those it applied when

Teva adopted its skinny label. And, regardless, the

Federal Circuit’s new rule makes any regulatory

change irrelevant, because its decision gave no weight

to the Hatch-Waxman statute or regulations. That is

precisely why certiorari is needed.

The proceedings below spanned two oral arguments, sua sponte panel rehearing, a divided en banc

vote, and eight total opinions. Those are not the hallmarks of a factbound dispute. The Federal Circuit has

replaced a regime of predictability and certainty with

one of doubt and risk. The result will be less competition and higher prices. This Court should intervene

without delay.

ARGUMENT

I. The decision below upends settled legal principles and an important legislative compromise.

The Federal Circuit now permits inducement liability based on label language that does not recommend infringement and that the brand never identified as patented. That decision guts a key element of

inducement law and distorts Hatch-Waxman’s careful

balance.

A. The Federal Circuit eviscerated an essential element of inducement law.

1. GSK contends (at 24-28) that the Federal Circuit

correctly recited (and the jury was properly instructed)

that § 271(b) requires proof that the defendant took

“active steps” to induce infringement. Anything beyond that top-line point, GSK says, is too factbound to

review. But the Federal Circuit’s legal analysis does

not stop at the highest level of generality.

3

The court fractured on a pivotal legal question:

what qualifies as “active” inducement? Teva’s FDAdrafted skinny label excised the patented use GSK

had identified. So GSK pointed to scattered phrases

elsewhere in the label that, according to its expert,

just “mentioned” elements of the patented method

(Pet.13-14, 18). Until now, those pieces of Teva’s label

would not have qualified as active inducement under

this Court’s precedents because they do not recommend the patented use (Pet.22-27), and they would

have been protected by Hatch-Waxman (Pet.28-32).

But the Federal Circuit held GSK’s theory legally

proper. That is a repudiation of precedent, not just a

one-time misapplication (Opp.27-28). See, e.g., Professors’ Br. 5; Mylan Br. 11; Pet.18-19.

This Court has repeatedly reviewed such legal-sufficiency questions in patent cases, e.g., Life Technologies Corp. v. Promega Corp., 137 S. Ct. 734 (2017), and

this case likewise warrants review. Virtually any

skinny-label case can involve similar facts; after all,

when a drug is approved for multiple uses, the unpatented use is often related enough to the patented one

that even a skinny label could “mention” patented elements. E.g., Takeda Pharms. U.S.A., Inc. v. W.-Ward

Pharm. Corp., 785 F.3d 625, 630 (Fed. Cir. 2015) (label discussing unpatented use, preventing gout flareups, mentioned but did not instruct patented use,

treating gout flareups). As a result, every significant

case will now go to trial.

2. GSK insists (at 29-30) that it can harmonize the

decision below with the long line of cases holding that

passive description is not active inducement. For example, GSK says that in cases like Takeda, HZNP,

and Bayer, there was no evidence the defendant’s

4

label actually encouraged the alleged infringement.

But the supposed “encouragement” in those cases (description of infringing elements in scattered sections

of a label) is materially indistinguishable from Teva’s

supposed “encouragement.” Pet.25-27, 30-31. The

dissenter’s view in those cases has become the majority holding here. Pet.31.

GSK insists that those cases must be factually different—but just what facts make them different, GSK

never says. The closest it comes is claiming (at 8, 17)

that Teva “manipulat[ed]” its label to “capture” the

carvedilol market. But Teva simply followed the

standard carve-out path, which Congress created precisely so manufacturers would not have to “wait until

[a method patent] expire[s]” to market low-cost generics, as GSK would prefer (Opp.6). Indeed, Teva’s

carve-out was not even unique compared to other generic carvedilol manufacturers: seven others launched

with the same skinny label that FDA supplied to Teva.

Pet.11. This is a run-of-the-mill skinny-label case,

making the Federal Circuit’s conclusion that a jury

can find inducement on these facts a dramatic legal

shift.

B. The Federal Circuit’s rule is inconsistent

with the Hatch-Waxman regime.

GSK argues (at 8, 11, 28, 31-32) that the decision

below does not disturb the Hatch-Waxman framework

because generic applicants must independently determine which parts of a brand’s label implicate patented

uses. That argument distorts the statutory and regulatory regime.

Hatch-Waxman relies on a simple information exchange. Brand manufacturers must provide sworn

5

representations about what methods, and what corresponding sections of labeling, their patents claim;

FDA uses those representations to assess and approve

skinny labels for generics, which are exceptions to the

same-labeling requirement generics otherwise must

follow. Pet.7-10; AAM Br. 6, 11-12. That is the rule

now—and it was the rule when FDA prepared Teva’s

skinny label. FDA made clear in 2003 that it would

require brand manufacturers “to identify specifically

the approved uses claimed by [any] method-of-use patent, with reference to the approved labeling,” so that

a generic applicant can “assess whether [it] is seeking

approval for a use the sponsor states is claimed in the

listed patent.” 68 Fed. Reg. 36,676, 36,682 (Jun. 18,

2003) (emphasis added). 1 Indeed, FDA contrasted its

chosen approach against one requiring the generic “to

make its own independent decision on whether a

listed method-of-use patent claims the use for which

the [generic] applicant seeks approval.” Id.

GSK repeatedly cites (at 8, 11, 31-32) FDA’s admonition that the 240-character “use codes” that

brands submit to FDA “are not meant to substitute for

the [generic] applicant’s review of the patent and the

approved labeling.” 68 Fed. Reg. at 36,683. But as

FDA explained in the very next sentence, while Tweetlength use codes “may not fully describe the use as

claimed in the patent,” that sort of detail is supplied

by the brand manufacturer’s sworn “declaration,

which includes the complete description of the

method-of-use-claim and the corresponding language

in the labeling of the approved drug.” Id. (emphasis

added). FDA and generic applicants do use that

1 FDA’s 2016 clarifications echoed the same point.

69,580, 69,597 (Oct. 6, 2016).

81 Fed. Reg.

6

“publicly available” declaration—and the portions of

the label it identifies—in drafting skinny labels. Id.

The limitations on use codes that GSK emphasizes are

precisely why FDA insists on the sworn declarations.

GSK also emphasizes (at 31-32) that FDA plays a

“ministerial” role that does not involve a “substantive

review” of any listed patents. But that is because FDA

defers entirely to the brand manufacturer’s representations about what its patents cover. Caraco Pharm.

Labs., Ltd. v. Novo Nordisk A/S, 566 U.S. 399, 406407 (2012). That just highlights why a carve-out consistent with those representations cannot be active inducement—and why the Federal Circuit’s decision, requiring the generic applicant to beware even a carveout drafted by FDA, leaves the skinny-label system

broken.

The government endorses carve-outs as “critical,”

Pet.35-36, and never contemplated that its own carveout drafting would lead generics into massive liability.

The Federal Circuit declined requests to invite the

government’s views before ruling. This Court should

not credit GSK’s arguments without first hearing

from the government.

II. This issue warrants the Court’s prompt attention.

The decision below will seriously compromise manufacturers’ ability to market low-cost generic versions

of unpatented drugs—to the detriment of patients,

taxpayers, and the U.S. healthcare system. See

Pet.32-36; AAM Br. 14-18; Mylan Br. 13-19; Professors’ Br. 8-11 (providing data showing the reach of the

decision below). And it will have ripple effects beyond

just generics. See Alvotech Br. 16-22 (explaining the

7

chilling effect on alternatives to high-priced “biologic”

medicines). GSK fails to refute the urgent need for

this Court’s review.

A. The issue has not been “mooted” by subsequent regulations.

GSK argues (at 30-32) that 2016 regulatory

changes “mooted” the need for this Court’s review by

making “the regulatory scheme today” “dramatically

different.” Both the premise and the conclusion are

wrong.

First, the Federal Circuit’s (and GSK’s) view of the

law makes the regulations irrelevant to inducement

liability. GSK persuaded the Federal Circuit that it

does not matter what representations a brand manufacturer makes in its sworn declarations (or how specific they are), because the generic manufacturer can

be liable even if it carves out everything the brand

manufacturer identifies. See Pet.App.22a-25a; Opp.28.

GSK cannot argue against certiorari by pointing to a

supposed regulatory change that has no bearing under the legal rule it convinced the Federal Circuit to

adopt.

Second, GSK’s claim that the regulations are now

materially different is demonstrably false. At all relevant times, brand manufacturers have had to submit

detailed representations about their patents and labels. As early as 2003, for example, brand manufacturers had to “identify with specificity the section of

the approved labeling that corresponds to the method

of use claimed by the patent submitted.” 21 C.F.R.

§ 314.53(b)(1) (2003); supra pp. 4-6. That was the rule

in place in September 2006, when GSK submitted a

8

form 2 requiring it to identify the scope of its patent

protections “with specificity” and “with reference to

the proposed labeling.” Pet.App.57a (Prost, J., dissenting); C.A.App.6895. FDA drafted Teva’s skinny

label, and Teva launched its generic product with that

label, consistent with those representations. Pet.11.

GSK is simply wrong to suggest (at 15) that it had

not yet identified the protected portions of its label

when Teva launched in 2007. The only thing that

happened in 2008 was the reissue of GSK’s patent

with narrower claims. GSK resubmitted its patent

declaration after reissue, but identified no change

in the portions of its labeling claimed. Compare

C.A.App.6880-6887, with C.A.App.6890-6907.

Nothing about subsequent regulatory amendments materially changed brand manufacturers’ obligations. FDA repeatedly described the 2016 amendments on which GSK relies as mere “clarifications.”

See generally 81 Fed. Reg. 69,580 (using forms of the

words “clarify” more than 100 times). Most notably,

the amendments clarified that if a method patent

claims just a subsection of an indication, the brand

manufacturer must identify that specific subsection.

See id. at 69,581. That minor change is not even

2 As Teva explained, Pet.8 n.3, manufacturers submit two forms:

Form 3542a (with applications) and Form 3542 (after approval).

GSK suggests (at 10-12) that only the latter form matters. That

is wrong: FDA uses post-approval forms “to determine whether

a patent is eligible for listing” in the Orange Book, 68 Fed. Reg.

at 36,697, but that is an entirely different determination. Both

forms require sworn representations identifying the specific labeling sections claimed by any method patent, to make the carveout process function. Pet.8 n.3.

9

relevant here—GSK has never contended that its patent claims just a portion of an indication.

B. GSK ignores the unpredictability that the

Federal Circuit’s opinion engenders.

Like the Federal Circuit, GSK claims (at 33) that

the decision below will not affect generic manufacturers that “truly” carve out patented uses. But as already discussed, this is a “true” carve-out case: nothing sets Teva apart from other skinny-label generics.

See supra pp. 3-4. Going forward, no generic manufacturer can be confident it will avoid infringement liability by carving out everything the brand identifies

to FDA. Even if some manufacturers can ultimately

win a jury trial, they will all now face years of uncertainty and expense before any definitive resolution.

See Pet.32-36. That risk will cause many generic

manufacturers to forgo launching in the first place, as

amici explain.

Hatch-Waxman was designed to avoid exactly that

result. Generic manufacturers are supposed to be

able to obtain certainty pre-launch—either by resolving patent disputes before launch, without a jury trial

or damages (if seeking approval for patented indications), or by avoiding patent litigation altogether (if

seeking approval only for unpatented indications

through a section viii carve-out). But now, a carve-out

is the worst of all worlds: it provides no pre-launch

resolution and no protection from massive exposure.

The result will be precisely what Congress did not

want: one narrow method patent holding generics off

the market entirely.

10

III. This case presents an ideal opportunity to

address these important issues.

GSK invents several unpersuasive reasons why

this case is supposedly a poor vehicle.

First, GSK suggests (at 8-13) that, as a factual

matter, it did identify the post-MI LVD indication in

its declaration—FDA and Teva just failed to recognize

it. That brand-new argument is a brazen distortion of

the record. GSK’s label identified three uses for carvedilol, including the CHF indication:

Congestive Heart Failure: COREG is indicated for the treatment of mild

to severe heart failure of ischemic or cardiomyopathic origin, usually in addition to diuretics, ACE inhibitor, and digitalis, to increase survival and,

also, to reduce the risk of hospitalization (see CLINICAL TRIALS).

Left Ventricular Dysfunction Following Myocardial Infarction: COREG

is indicated to reduce cardiovascular mortality in clinically stable patients

who have survived the acute phase of a myocardial infarction and have a

left ventricular ejection fraction of ≤ 40% (with or without symptomatic heart

failure) (see CLINICAL TRIALS).

Hypertension: COREG is also indicated for the management of essential

hypertension. It can be used alone or in combination with other antihypertensive agents, especially thiazide-type diuretics (see PRECAUTIONS,

Drug Interactions).

C.A.App.7992 (highlighting added). When asked

which of those uses its patents claimed, GSK recited

the CHF indication nearly verbatim:

4.2a If the answer to 4.2 Use: (Submit indication or method of use

is “Yes,” identify with information as identified specifically in the

approved labeling.)

specificity the use

with reference to the

proposed labeling for Treatment Of Mild-To-Severe Heart Failure Of

Ischemic Or Cardiomyopathic Origin, Usually

the drug product.

In Addition To Diuretics, ACE Inhibitor, And

Digitalis, To Increase SurvivalI

I

r

C.A.App.6895 (highlighting added). It said nothing

about the post-MI LVD indication. Now, GSK seeks

to rewrite its sworn declaration using ellipses,

Opp.12, rather than acknowledge that it tracks the

CHF indication so specifically.

11

To make matters worse, GSK misstates the scope

of both its patent claims and the post-MI LVD indication. GSK claims (at 7) that its patent applies “no

matter the patient’s specific heart failure symptoms.”

But GSK’s patent covers only treatments specifically

intended to reduce mortality caused by symptomatic

heart failure; GSK conceded that the patient must

have been diagnosed with congestive heart failure.

C.A.App.130, available at 2016 WL 3186657, at *19 &

n.22 (D. Del. June 3, 2016). The post MI-LVD indication, by contrast, involves reducing mortality in certain patients after a heart attack, “with or without

symptomatic heart failure”—i.e., the patient need not

be diagnosed with congestive heart failure. GSK also

never mentions that the patent requires co-administration with an ACE inhibitor, diuretic, or digoxin,

Pet.App.5a, but the post-MI LVD indication does not.

The latter indication is “outside the scope of the

claims.” Pet.App.167a n.9. That is why FDA approved eight skinny-labeled generics without requiring them to carve out the post-MI LVD indication—

and why GSK didn’t utter a peep of protest then.

Second, GSK claims (at 32-33) that the Federal

Circuit’s remand for Teva to pursue an equitable-estoppel defense somehow makes the record “not yet

complete.” Nonsense. The jury-trial record is closed

and fully developed, and the question presented is

whether that record required JMOL, making any remand erroneous. This Court routinely grants certiorari in that posture. Pet.36-37.

Finally, GSK wrongly suggests (at 2, 24, 32-33)

that the question presented is “not case dispositive,”

primarily because a GSK witness claimed that “over

70% of the damages” accrued after Teva amended its

12

label at FDA’s direction, Pet.12. 3 But GSK cannot

prove damages during the amended-label period if

Teva’s skinny label did not encourage doctors to infringe. “The reason is simple: nothing about doctors’

prescribing practices changed when Teva amended its

label to the full version,” as the majority did not dispute and GSK conceded. Pet.App.76a-77a (Prost, J.,

dissenting). GSK’s 70/30 argument could not sustain

the verdict in any event: the verdict sheet did not apportion damages between the two periods, and the

jury declined to adopt GSK’s proffered damages

amount. And even if GSK were right that skinny-label damages were “only” $70 million, that is no reason

to leave the Federal Circuit’s erroneous rule unreviewed—and let GSK keep that windfall.

3 GSK also refers (at 27) to marketing materials calling Teva’s

product “AB-rated” or the “generic equivalent” of Coreg. But

the panel majority acknowledged that those materials just

“point physicians to [the supposedly inducing] partial label.”

Pet.App.32a n.7; see id. at 74a-76a (Prost, J., dissenting). If that

label does not induce, these materials add nothing. Id. The majority’s first opinion placed greater reliance on these materials,

Pet.App.100a, but GSK declined to defend that rationale and focused instead on the post-MI LVD indication. GSK Reh’g Resp.

14-17 (Jan. 29, 2021). It cannot backtrack now that the panel,

too, has retreated from relying on those materials.

13

CONCLUSION

The petition should be granted.

Respectfully submitted.

DARYL L. WIESEN

CHRISTOPHER T. HOLDING

ELAINE HERRMANN BLAIS

ROBERT FREDERICKSON, III

GERARD J. CEDRONE

WILLIAM E. EVANS

GOODWIN PROCTER LLP

100 Northern Avenue

Boston, MA 02210

WILLIAM M. JAY

Counsel of Record

GOODWIN PROCTER LLP

JAIME A. SANTOS

1900 N Street, NW

Washington, DC 20036

(202) 346-4000

wjay@goodwinlaw.com

September 7, 2022

Counsel for Petitioner