Supplemental 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.,

Petitioner,

v.

GLAXOSMITHKLINE LLC,

SMITHKLINE BEECHAM (CORK) LIMITED,

Respondents.

On Petition For a Writ of Certiorari to

The United States Court of Appeals

for the Federal Circuit

SUPPLEMENTAL BRIEF FOR

RESPONDENTS

MICHAEL J. KANE

ELIZABETH M. FLANAGAN

Fish & Richardson P.C.

60 South Sixth Street

Suite 3200

Minneapolis, MN 55402

(612) 335-5070

kane@fr.com

eflanagan@fr.com

JUANITA R. BROOKS

Counsel of Record

JONATHAN E. SINGER

Fish & Richardson P.C.

12860 El Camino Real

Suite 400

San Diego, CA 92130

(858) 678-5070

brooks@fr.com

singer@fr.com

Counsel for Respondents

i

QUESTION PRESENTED

When a generic drug is doubly indicated for a patented use, and there is strong record evidence of intent and inducing conduct, can the generic manufacturer evade liability for induced patent infringement

merely because it did not include on its label (i.e.

“carved out”) one of the two indications corresponding

to the patented use?

ii

RULES 24(B) AND 29.6 STATEMENT

All parties are identified in the caption of this

brief. Respondent GlaxoSmithKline LLC, is a forprofit Delaware company. Respondent SmithKlineBeecham (Cork) Limited is a for-profit organization.

GSK plc owns 10% or more of the stock in respondents.

iii

TABLE OF CONTENTS

Page

QUESTION PRESENTED.......................................... i

RULES 24(B) AND 29.6 STATEMENT .................... ii

I.

The Factual Underpinnings of the

Government’s Argument for Certiorari Are

Unsupported by the Record and Will Be the

Subject of Further Proceedings ..................................4

II.

The Government’s View of the CarveOut Process Contradicts Decades of FDA

Published Practice, Policy and Guidance ...................8

iv

TABLE OF AUTHORITIES

Page(s)

Cases

AstraZeneca LP v. Apotex, Inc.,

633 F.3d 1042 (Fed. Cir. 2010) ............................ 10

Caraco Pharm. Labs., Ltd. v. Novo

Nordisk A/S,

566 U.S. 399 (2012) ................................................ 9

Global-Tech Appliances, Inc. v. SEB S.A.,

563 U.S. 754 (2011) .............................................. 11

Metro-Goldwyn-Mayer Studios Inc. v.

Grokster, Ltd.,

545 U.S. 913 (2005) .............................................. 11

Stilz v. United States,

269 U.S. 144 (1925) .............................................. 11

Takeda Pharmaceuticals U.S.A., Inc. v.

West-Ward Pharmaceutical Corporation,

785 F.3d 625 (Fed. Cir. 2015) .............................. 10

Statutes

35 U.S.C. § 271(e)(1) .................................................. 10

Regulations

21 CFR 314.94(a)(8)(iv) ............................................... 3

68 Fed. Reg. 36,683 (June 18, 2003) ....................... 8, 9

v

80 Fed. Reg. 6828 (February 6, 2015) ...................... 10

81 Fed. Reg. 69,599 (October 6, 2016) ........................ 9

Other Authorities

Coreg Package Insert (02/2007)

available at https://www.accessdata.fda.gov/drugsatfda_docs/label/2007/020297s022lbl.pdf. .................................. 8

U.S. Food and Drug Administration,

Summary of FDA’s FY2024

Legislative Proposals, available at

https://www.fda.gov/media/166049/d

ownload................................................................. 12

1

Rather than address the actual facts of this case

and the documented history of the section viii “carveout” process, the government rewrites both to reach

the result it wants. Certiorari should not be granted

based on such false premises, which will lead to messy

merits proceedings dominated by factual disputes,

making this case a poor vehicle for addressing the issues raised by the government.

A properly-instructed jury found Teva willfully intended to cause infringement with its “skinny label”

generic carvedilol that carved out indication 1.1 of

GSK’s Coreg® label, but left on indication 1.2. There

is no doubt substantial evidence supports this finding.

The government focuses on one piece of that evidence

– the label itself – to the exclusion of everything else,

as if the issue of intent boils down only to that.

At trial, however, the jury heard voluminous additional evidence of Teva’s intent, including testimony

from Teva’s commercial witness (admitting Teva intended to capture the entire heart failure market),

statements in Teva’s product brochures calling Teva’s

product a “Generic of COREG® Tablets,” and 2004

and 2007 press releases publicizing that Teva’s generic carvedilol should be used to treat heart failure.

C.A.Appx10488, 4245, 6347, 6342. Those press releases remained on Teva’s website for years, with the

former explicitly lumping indications 1.1 and 1.2 together as “heart failure” and the latter deliberately

describing Teva’s product as a “generic version of

GSK’s cardiovascular agent COREG® (Carvedilol)

Tablets” without any hint Teva’s product had a skinny

label. C.A.Appx11857, 10542; Pet.App.35a. GSK’s

physician expert testified unrebutted he read the releases and understood them to mean Teva’s generic

carvedilol should be used in accordance with the

2

method of use claimed in GSK’s ’000 patent.

C.A.Appx11656, 11659. The jury also heard evidence

Teva later switched to a full label with no carve out

and without the required notice to GSK – a brazen

move from which the jury could draw inferences regarding what Teva’s intent was all along.

C.A.Appx10569-10572, 11044, 11049-50; Pet.App.8a9a.

Although no additional support was needed, the

label itself was also evidence of Teva’s intent. The

government’s contrary assertions rest on a clear misunderstanding about how the carved-out label arose.

As demonstrated, neither FDA nor Teva could possibly have relied on GSK’s 2008 Form 3542 in crafting

Teva’s “skinny label.” Opp. 12-13. GSK submitted the

form in February 2008, six months after Teva

launched its “skinny label” generic carvedilol in September 2007. C.A.Appx6880-6882. The form had

nothing to do with FDA’s or Teva’s conduct, so it cannot serve as the basis for overturning the jury’s willful

infringement finding.

But even if the form were probative of any issue

on appeal, the government’s statements about it are

unmoored from the record. First, the government fails

to even mention the use code identified by GSK, U233, let alone argue choosing that use code was improper. U-233, “Decreasing Mortality Caused by Congestive Heart Failure,” reads directly on Claim 1 of

the asserted patent (“A method of decreasing mortality caused by congestive heart failure…”) and is strikingly similar to indication 1.2 that Teva retained on

its label (“Carvedilol is indicated to reduce cardiovascular mortality…”). C.A.Appx45, 7665. Second, contrary to the government’s assertion, GSK did not “represent” on that form that the ’000 patent was limited

3

to indication 1.1. U.S. Br. 15. The description GSK

provided on the form applies equally to indication 1.1

and indication 1.2. C.A.Appx6881. Simply put, GSK

made no misrepresentations to FDA or Teva about

what approved uses were patented.

Just as puzzling is the government’s unsupported

assertion FDA purportedly “gave” the carve-out label

to Teva “based on” GSK’s submissions. U.S. Br. 6. For

nearly two decades, the process has been the opposite,

with generic applicants required by regulation to propose carve-outs based on their review of the use code,

patent, and brand label, and FDA conducting a review

of those carve-outs for safety and efficacy. 21 CFR

314.94(a)(8)(iv). That Teva apparently decided not to

follow this process does not demonstrate an “intent

not to infringe,” but rather recklessness with respect

to its statutory and regulatory responsibilities.

But this Court need not rely on GSK’s, Teva’s or

the government’s word on any of these factual matters. A supplemental proceeding on remand is already

slated to address these questions in the context of

Teva’s equitable estoppel defense. The government

attempts to short-circuit this process by resolving contested issues against GSK and dispensing with evidence and law that does not fit its policy objectives.

This is no basis for recommending a grant of certiorari. The petition should be denied.

4

I.

The Factual Underpinnings of the Government’s Argument for Certiorari Are Unsupported by the Record and Will Be the Subject

of Further Proceedings

At the heart of the government’s brief are two erroneous factual assertions: 1) GSK “represented” on

its 2008 Form 3542 that its patented use was limited

to indication 1.1 on Coreg’s label through the use of

language “essentially identical” to indication 1.1 (U.S.

Br. 6); and 2) “based on” information provided by GSK,

FDA “gave” to Teva the redlined document that ultimately became its skinny label (id.). Neither assertion is supported by the record.

With respect to the former, the form at issue allowed NDA holders to identify the approved use covered by their patents by “indication or method of use

information.” C.A.Appx6881 (emphasis added). GSK

chose the latter. Id. Indeed, GSK’s 2008 Form 3542

neither states “indication 1.1” in the one-inch high by

five-inch wide box provided on the form, nor cites indication 1.1.

Instead, GSK used language describing its approved (and patented) “method of use” in language

from the approved label, the other option on the form.

The language used contains a portion of the wording

of indication 1.1, specifically, a portion that also covers the patented aspect of indication 1.2 – namely,

treatment of those patients with MI/LVD who have

symptomatic heart failure. As both parties’ experts

agreed, these patients by definition have congestive

heart failure, one of the conditions listed by GSK in

the box on the form.

C.A.Appx10602-10606

(McCullough); C.A.Appx11132; 11226 (Zusman). Significantly, the language used on the form also omits

language from indication 1.1 – reducing the risk of

5

hospitalization – that had not been shown in the clinical trial supporting indication 1.2 and thus would not

cover that indication. C.A.Appx5545-5549.

At trial, the jury heard testimony explaining these

distinctions from a scientific point of view, primarily

from Mary Anne Lucas, a listed inventor on the ’000

patent. Dr. Lucas explained to the jury the different

stages of heart failure, and that patients who have recently suffered a heart attack and are symptomatic

have “mild” heart failure. C.A.Appx10359-10360,

10381-10382. They heard her describe the clinical

trial supporting indication 1.2 (CAPRICORN), which

showed a decrease in cardiovascular mortality but not

a reduction in the risk of hospitalization, while the

clinical trials supporting indication 1.1 (COPERNICUS AND COMET) showed both. C.A.Appx1035910360, 10378-10383, 5545-5549. And they heard both

experts agree that symptomatic patients under indication

1.2

have

congestive

heart

failure.

C.A.Appx10602-10606, 11132, 11226.

Accordingly, if any generic applicant were to utilize GSK’s Form 3542 in the “carve-out” process, the

form’s description of the method of use would have signaled the need to carve out both indication 1.1 and indication 1.2 to avoid patent infringement.

But regardless of how one reads Form 3542, its

discussion of the label was irrelevant to Teva, as GSK

did not submit the form until February 2008, six

months after Teva launched its skinny-labeled product.1 What was available to Teva to inspect, at the

relevant time, was the ’069 patent, e.g., Claim 1 “A

method of decreasing mortality caused by congestive

1 The record is silent as to when the form ultimately

became public.

6

heart failure. . .,” and the corresponding “use code,”

“Decreasing Mortality Caused by Congestive Heart

Failure,” which GSK later repeated on its 2008 Form

3542. In fact, the use code contains language that is

closer to indication 1.2 than to indication 1.1.

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 less than or equal

to 40% (with or without symptomatic

heart failure) [see Clinical Studies (14.2)].

C.A.App.7665 (emphases added).

The government gives no weight to any of this, instead preferring to emphasize that the small box on

the 2008 Form 3542 does not contain specific excerpts

of language from indication 1.2, like the words “Left

Ventricular Dysfunction.” But that incorrectly presumes physicians would not understand “mild” congestive heart failure to be what indication 1.2 is describing in symptomatic patients. The physician experts at trial agreed to the opposite of that presumption, both testifying indication 1.2 does, in fact,

describe patients suffering from congestive heart failure. C.A.Appx10602-10603, 11132.

All of this is to say what the government asserts

about GSK’s 2008 Form 3542 is inaccurate or, at minimum, the subject of dispute. If certiorari is denied, a

bench trial on equitable estoppel will address the very

issues raised by the government in its brief, as the majority at the Federal Circuit understood. Pet.App.25a.

This Court should not grant review at this stage based

on faulty factual assumptions by the government that

7

contradict the jury’s implied fact finding and will be

further refuted at the equitable estoppel trial.

At that bench trial, GSK will present evidence

that, at all relevant times, FDA was well aware GSK

and heart failure practitioners understood symptomatic patients under indication 1.2 were suffering

from “mild” heart failure, and use of carvedilol could

increase survival, one of the “methods of use” described in the box on GSK’s 2008 Form 3542. For example, at the FDA advisory committee meeting regarding approval of indication 1.2 in 2002, GSK provided extensive testimony and documentation explicitly telling FDA that indication 1.2 covered patients

with “mild” heart failure. C.A.Appx11968, 1193-65.

GSK will also present information currently classified

as confidential that bears on the issue of what, on

GSK’s label, other generic companies believed was

and wasn’t patented.

Also at issue in that bench trial will be the government’s second critical and flawed assertion – that

FDA prepared Teva’s skinny label “based on” submissions from GSK. The present record contains not a

shred of evidence to support this notion. Teva

shielded most discovery into the origins of its skinny

label through assertions of privilege, allowing the jury

to see only emails describing Teva learning another

generic applicant was intending to pursue a skinny label, deciding to go that same route, looking for a

skinny label prepared by a generic company Teva had

acquired, and at the last minute receiving a red-lined

“skinny label” from FDA, with no further identification as to how that label came to be. C.A.Appx7993,

6908-6951.

8

The origins of this skinny label thus remain a

mystery on the present record, despite the government’s claim. And this is not merely a theoretical

problem, but one with real impact on the merits. For

example, the “redline” version of the skinny label provided by FDA to Teva in September 2007 has a different indication 1.1 than GSK’s label for Coreg® approved in February of 2007, with GSK’s label having

the word “chronic” in indication 1.1, while the red-line

“skinny label” version does not. GlaxoSmithKline,

Coreg (carvedilol) package insert, revised 02/2007,

U.S. Food and Drug Administration website, available

at https://www.accessdata.fda.gov/drugsatfda_docs/

label/2007/020297s022lbl.pdf; C.A.Appx6913.

Accordingly, granting certiorari here would result

in only one thing for certain – a dispute over the facts

that would impede consideration of any legal issues.

Certiorari should not be granted when the factual record is in such dispute and a trial designed to address

Teva’s and the government’s unsupported assertions

will be held on remand.

II. The Government’s View of the Carve-Out

Process Contradicts Decades of FDA Published Practice, Policy and Guidance

In addition to relying on faulty factual assertions,

the government’s brief misstates the law and makes

flawed policy arguments more appropriately addressed to Congress.

The government acts as if GSK should have submitted a redline label on its 2008 Form 3542. Not so.

There was no requirement – nor room on the form –

for the NDA holder to identify everything on its label

related to the listed patent. 68 Fed. Reg. 36,686;

36,710-712 (June 18, 2003). And while FDA has never

9

required NDA holders to identify all the specific language on the label relating to the patented method of

use, it did not even impose an obligation to identify

sections and subsections of the labeling describing the

claimed method of use until 2016, almost a decade after the events in question. 81 Fed. Reg. 69,599 (Oct.

6, 2016).

Rather, the focus of the regulatory process for two

decades has been on use codes, which the Court recognized are “pivotal to FDA’s implementation of the

Hatch-Waxman Amendments.” Caraco Pharm. Labs.,

Ltd. v. Novo Nordisk A/S, 566 U.S. 399, 419 (2012).

In this process, FDA has consistently stated that its

role in patent listing is ministerial and does not involve substantive review of patents. Id. at 406-07.

FDA’s guidance available at the time relevant to this

case stated that FDA’s regulations and practices surrounding use codes serve a notice function, but even

use codes were “not meant to substitute for the applicant’s review of the patent and the approved labeling.”

68 Fed. Reg. 36,683 (June 18, 2003).

Here, the use code reads plainly on the ’000 patent

and indications 1.1 and 1.2. U-233, “Decreasing Mortality Caused by Congestive Heart Failure,” is almost

identical to the preamble of Claim 1: “A method of decreasing mortality caused by congestive heart failure

. . . .” The story is the same as to the contested indication 1.2: “Carvedilol is indicated to reduce cardiovascular mortality…in patients … with … symptomatic heart failure.”

After conducting a review of the patent in light of

the use code, the applicant, not FDA, is supposed to

“propose labeling for the generic drug that ‘carves out’

from the brand’s label the still-patented methods of

use.” Caraco, 566 U.S. at 406. Even in proposing the

10

expanded 2015 regulatory requirements, FDA made

clear it would defer to a generic applicant seeking a

carve-out because the applicant has “a strong incentive to interpret the scope of the patent correctly to

avoid being subject to patent infringement litigation

following ANDA approval and potentially enjoined

from marketing its product.” 80 Fed. Reg. 6828 (February 6, 2015).

Teva’s failure to follow this process is the cause of

its current predicament. Contrary to the government’s implications, generics are not bound to slavish

copying of the innovator’s label in the “skinny label”

context. C.A.Appx10548-10550 (GSK’s regulatory expert explaining changes Teva could have requested).

For example, in Takeda Pharmaceuticals U.S.A., Inc.

v. West-Ward Pharmaceutical Corporation, 785 F.3d

625, 630 (Fed. Cir. 2015), the generic label included

statements not on the innovator’s label that disclaimed the patented use. See also AstraZeneca LP v.

Apotex, Inc., 633 F.3d 1042, 1058-59 (Fed. Cir. 2010)

(generic negotiated with FDA regarding wording of its

label). But Teva failed to take basic steps to avoid an

infringing label and, as the jury heard, actively encouraged infringement through multiple channels.

The government’s proposed rewriting of the statutes to relieve Teva of the consequences of its actions

flies in the face of long-standing precedent, practice

and guidance. When Congress wanted to create a

safe-harbor for generics, it did so expressly. See 35

U.S.C. § 271(e)(1).

Likewise, the government’s position finds no support in this Court’s inducement precedent. As GSK

detailed and the government does not dispute, the uncontested jury instructions were in complete agree-

11

ment with Grokster and Global-Tech. Opp. 24-26 (citing Metro-Goldwyn-Mayer Studios Inc. v. Grokster,

Ltd., 545 U.S. 913, 935-37 (2005) and Global-Tech Appliances, Inc. v. SEB S.A., 563 U.S. 754, 760 (2011)).

Moreover, “[i]nfringement is a question of fact.”

Stilz v. United States, 269 U.S. 144, 147 (1925). And

the Federal Circuit majority held substantial evidence

supported the jury verdict of willful infringement and

did not upset the careful balance Congress struck in

Hatch-Waxman. Pet.App.11a-12a, 14a-15a, 27a. The

government’s argument that a skinny label “cannot

provide” evidence of specific intent, U.S. Br. 14, does

not justify taking this factual question away from the

jury.

As noted, the jury had sufficient evidence outside

the label to find intent. See supra p. 1-2. Moreover,

even the government retreats from its per se rule by

acknowledging a generic manufacturer could be liable

for exploiting its skinny label to induce doctors to

practice the patented method. U.S. Br. 16. The dispute here – and in all the relevant cases – is thus factual, not legal.

And while the unique events at issue here occurred long ago and under a very different regulatory

scheme, Opp. 30-31, the government fails to substantiate its claim that even under the current regulations, the section viii carve-out process is under

threat. The government admits liability for inducement by a skinny label is “rarely imposed,” U.S. Br.

22, and does not cite a single example of any change

in practice since the verdict in this case was first affirmed in 2020.

12

To the extent any adjustment to Hatch-Waxman

is needed, Congress is the proper forum for the government’s policy-driven argument. Consistent with

this, FDA announced it will be asking Congress to do

just that. U.S. Food and Drug Administration, Summary of FDA’s FY2024 Legislative Proposals at 3,

available at https://www.fda.gov/media/166049/download. In that request, FDA conceded that the Federal

Circuit majority indicated its decision was narrow,

fact dependent, and does not upset the careful balance

struck by Hatch-Waxman. Id. And while the record

contains no evidence that the decision has discouraged section viii carve-outs, Congress can hear from

all stakeholders – pharmaceutical innovators, generic

manufacturers, insurance providers, doctors and patients – and make a public policy decision about

whether any adjustment to the careful balance of

Hatch-Waxman is needed. This Court should not

preempt that process by accepting the government’s

policy-driven invitation to rewrite the statute and ignore all the evidence that supported the jury’s verdict.

The petition for writ of certiorari should be denied.

Respectfully submitted,

JUANITA R. BROOKS

Counsel of Record

JONATHAN E. SINGER

Fish & Richardson P.C.

12860 El Camino Real

Suite 400

San Diego, CA 92130

(858) 678-5070

brooks@fr.com

singer@fr.com

13

MICHAEL J. KANE

ELIZABETH M. FLANAGAN

Fish & Richardson P.C.

60 South Sixth Street

Suite 3200

Minneapolis, MN 55402

(612) 335-5070

kane@fr.com

eflanagan@fr.com

April 11, 2022

Counsel for Respondents

GlaxoSmithKline LLC,

and SmithKlineBeecham (Cork) Limited