Amicus Curiae Brief — Merck Sharp & Dohme Corp. v. Albrecht (No. 17-290)

No. 17-290

In the Supreme Court of the United States

MERCK SHARP & DOHME CORP., PETITIONER

vz.

DortIs ALBRECHT, ET AL.

ON WRIT OF CERTIORARI

TO THE UNITED STATES COURT OF APPEALS

FOR THE THIRD CIRCUIT

BRIEF FOR THE UNITED STATES

AS AMICUS CURIAE SUPPORTING PETITIONER

Acting Solicitor General

Counsel of Record

JosePH H. HUNT

BRIAN R. STIMSON Assistant Attorney General

+ —~ ~ ieee EDWIN S. KNEEDLER

ELIZABETH H. DICKINSON ———

Associate ANG

Assistant to the Solicitor

General

ANNAMARIE KEMPIC

Chief Counsel Scort R. McINTOSH

"Vaidation JOSHUA M. SALZMAN

KAREN E. SCHIFTER Attorneys

Senior Counsel Justice

——— of Health and Ww C. 20530-0001

uman Services S gov

Washington, D.C. 20201 (202) 514-2217

QUESTION PRESENTED

Whether a state-law failure-to-warn claim alleging

the insufficiency of brand-name drug labeling is pre-

empted by the Federal Food, Drug, and Cosmetic Act,

21 U.S.C. 301 et seqg., when the Food and Drug Admin-

istration, after the drug manufacturer provided it with

the relevant scientific data, rejected the manufacturer's

application to modify its labeling to warn about the risk

underlying the tort claim.

(I)

TABLE OF CONTENTS

Page

Interest of the United States ................c.ccccceseensecneenneeesnnnnnnnnenes 1

IIIS ccnssenneccmnssnnsesresssrersesesnazemnsencsemssessnemanesnnqnssacesnennenens 1

TN GE GRIEG ccccsncscccsczsssscssesecssscscssanasscsrsmenssnesneneenssness 14

Argument:

The court of appeals erred in rejecting petitioner’s

UA GU RID cxcsccssscscssmscscemssenmesessmnseeemmesnenvesseneesnese 16

A. Petitioner’s preemption defense turns on the

meaning and effect of FDA’s 2909 Complete

Response Letter, which are legal questions that

a court must resolve 17

1. Courts, not juries, determine the meaning

of federal agency decisions ..................<.c000--0-0--- 18

2. Factual findings necessary to interpret a

federal agency decision are also matters for

courts to resolve =

B. age ty meets ee on

to a clear-and-convincing-evidence standard ............ 21

1. Wyeth did not address whether courts or

juries should construe an actual FDA

Qe cecrerensesmmenmsmmsemnenene 22

2. Any factual findings necessary to resolve

a preemption defense need only rest on a

preponderance of the evi ence.................0-0.00-0-0+- 24

3. Resolving this case does not require factual

findings =

C. FDA's May 2009 decision rejected a change to

Fosamax’s warnings and precautions because the

data at that time was insufficient to justify a

change 30

ee 35

(IIT)

IV

TABLE OF AUTHORITIES

Cases: Page

Addington v. Texas, 441 U.S. 418 (1979) ......ccccccccccecevee 2A, 25

Arizona v. Inter Tribal Council of Ariz., Inc.,

DIO UB. 3 GIBB) enccccccccscsccscsuscsccscscscsvsscsnsesveesesecscccccececeseeee 25

Armour & Co. v. Wantock, 323 U.S. 126 (1944) ................. 24

Block v. Community Nutrition Inst., 467 U.S. 340

(1984) 26

Boyle v. United Techs. Corp., 487 U.S. 500 (1988)....... 28, 29

Buckman Co. v. Plaintiffs’ Legal Comm.,

GD TE, a Go ecennensssnsnssnsssnvmensernsnssssnsssssccscssssescscsseseces 28

Cohens v. Virginia, 19 U.S. (6 Wheat.) 264 (1821)............. 24

Dolin v. GlaxoSmithKline LLC, No. 17-3030,

2018 WL 4001208 (7th Cir. Aug. 22, 2018) ..............000000 33

Emich Motors Corp. v. General Motors Corp.,

GD GE, Ga Ca rascccsnsnsscensnsscsnssssszasssscemsssnenemcssescccsssecss 19

English v. General Elec. Co., 496 U.S. 72 (1990) ..........00000 26

Geier v. American Honda Motor Co., 529 U.S. 361

(2000) -» 26

Grogan v. Garner, 498 U.S. 279 (1991) 24

Landgraf v. USI Film Prods. , 511 U.S. 244 (1994) ........... 24

Markman v. Westview Instruments, Inc.,

is 15, 19

Microsoft Corp. v. isi Ltd. P’ship, 564 U.S. 91

(2011) 2A, 25

Mutual Pharm. Co. v. Bartlett, 510 U.S. 472 (2013) coco. 4

PLIVA, Ine. v. Mensing, 564 U.S. 604

(2011) 4, 17, 18, 21,

Perez v. Mortgage Bankers Ass'n, 135 8. Ct. 1199

(2015) 20

Riegel v. Medtronic, Inc., 552 U.S. B12 (2008) ..cccccccoccoo 29

Rivera v. Minnich, 483 U.S. 574 (1987) ......0c00ccccecererenenenens 25

Cases—Continued: Page

Santosky v. Kramer, 455 U.S. 745 (1982) ......cccccceceeeeeeeenees 24

Sprietama v. Mercury Marine, 537 U.S. 51 (2002)............ 20

Teva Pharm. USA, Inc. v. Sandoz, Inc.,

135 S. Ct. 881 (201B) ......-.cccneseresnenenesesnsnennsnsnsnenensnnnseenennns 21

Wyeth v. Levine, 555 U.S. 555 (2009) .000....cccccccceeenves passim

Statutes, regulations, and rules:

Administrative Procedure Act, 5 U.S.C. 551 et seq.,

701 et seq. —

5 U.S.C. 551(6) - 18

5 U.S.C. 561(7) 18

5 U.S.C. 706 —_ 15, 18

Federal Food, Drug, and Cosmetic Act,

21 U.S.C. 301 et seq.:

a eR rncnscnccsnssecnsnsccemsesemsnsccmnsmnesememncsennies 2

21 U.S.C. 352(f) =

21 U.S.C. 3565(b)(1) 4

21 U.S.C. 355d) 2,4

21 U.S.C. 355(02) Ai) 10

21 U.S.C. 365(0)(4) 10, 11, 33, 34

21 U.S.C. 355(0)(4)(A) 10, 33

21 U.S.C. 355(0)(4)(B) 10

21 U.S.C. 355(04)(B)(D) we OO

21 U.S.C. 365(04)(C) il

21 U.S.C. 355(0)(4)(D) il

21 U.S.C. BECO AE) .....000ecceeceesensnnseneennsnnennnsnnennnennennees il

BD ae ID diccsnesecnsesnsscsscssecscssstnstsccssessssonssssetees 4

21 C.F.R.:

Pt. 201:

a cerecerercsccrscsesmennsemesscteesensensniens 2

Boction BO1 BRE) 222c.ccceccocecscecsccccscsescsesssessnseossscsoosees 4

Vi

Regulations and rules—Continued: Page

Section BO1 BBW) 1) ..0.-ccccccsecessereccsvseosersersrenseceessnsensonse 2

Section 201 .56(e) 2

Section B01 IT ...0c.0ccecsersevsevsessersnssersessersnssersersnssosssososoess 2

Section BILGE) .ccccccscosscecsececesrsecsessescssesscsssscresensensones 3

Section 201.57(¢)(6) ie 2,3

Section 201.57(¢)(6)(i) sasenenhsinmmenaesvnnt 4, 31, 33

Section 201.57(e)(7) ... 2, 3, 4, 30

Section 201.57(e)T)GIMB) «0... cecersercereennenenneennnnnnnne 4

Sloction BO1 BD....cccccecserseceecsecsevsevscssesenssessssenssssensseeees 2,3

Section B01. 10D(]) ....00ccscesessrscsessessossosnvscssesecscsesenssneeses 2

Pt. 314:

Bastion BIG BW) ...ccceccecoevsecseesecsessescessessnssesseseses 5, 27, 28

Section B14. 70D) .....0.0ec-ccerssssssesseevserseneessensensessosenseess 5

Blection BIA TODA) ..cccccccccsecescecsessescoscoscssccssscsssessonsenes 5

Section 314. 7OQD)MZ) ....0.-0--c-scsscersessereesersesensensssnsensenssees 5

Section 314. 7OMD)M2Z)V) ........---c-eceeeesnerereereennenneennnennnnes 5

Section 314. 7OCD)Z)(v MA) .........-eeveeneeneenreennennnennennns 5

Section 314.70(b)(3) 5

Section 314. 7O(CMG)..........c-ceeceesersnenvserenneeenennneanennennns 4

Section 314.70(e)(6)(iii) - 4, 5

Section 314. 7O(e (GIA) ......------ecceeeeererennennnnee 5, 27

Section 314. 7O(C)(7) ........-ec-eceecevsnveenenecnerneneennnnnnnnes 5, 17

Section B14.71() .....csccececsscsscssvsvsessosersrsesenseresssssnsseseses 5

Section 314.71(c) covcasestsenenessmasnnenesdsacennsees 5

Section B14. 10B(a) ........0.0-c-scesssssssesseesersnssesensssesenssessees 6

Section 314.102(b) 6

Section 314. 105(D) .........-ceee 6

Section 314.1 10(a) ..........-« 6

Section 314.110(a)(1) .......-..-..00 - 6

Section 314.1 10(a)(2) ........ccccecereeenseneneenenennne 6, 31, 32

Vil

Regulations and rules—Continued: Page

ee Be secencnsecvecsccccczsscesasecseressseeees 32

RE 6

ID sereesecessnosscemsccscnescnnssccsnemeseneensnes 6

ed, TR, BI Sie ecenessneversnevenseescassscssvsnssscerecscssssensssssssseemesnetn 25

ar, GR, BR, BID ccccunscscsnsnszssvsrsnsccensesessreseesvemmnesssesssonnemmnecsmnate 27

Miscellaneous:

Center for Drug Evaluation & Research, FDA,

CDER 21st Century Review Process: Desk

Reference Guide (2014), https://www fda.gov/

downloads/AboutF DA/CentersOffices/C DE R/

ManualofPoliciesProcedures/UCM218757.pdf.................. 6

Dorland’s Iliustrated Medical Dictionary

Ee Fie terrectnstnentecntemevesenenssnmtnsnmennmne

FDA:

Approved Drug Products with Therapeutic

Equivalence Evaluations (38th ed. 2018),

https://www.fda.gov/downloads/drugs/

ee anand

uem071436.pdf ... A

Guidance for intuey ‘Safety perrrnas

Changes—I mplementation of Section

505(0)(4) of the FD&C Act (July 2013),

https://www.fda.gov/downloads/

Drugs/GuidanceCompliance Regulatory

Information/Guidancea/UCM250783. pdf .................. 5

(Guidance for Industry: Warnings and

Precautions, Contraindications, and Boxed

Warning Sections of Labeling for Human

Prescription Drug and Biological Products—

Content and Format (Oct. 2011),

https://www .fda.gov/downloads/Drugs/

GuidanceCompliance Regulatory

Information/Guidances/UC M075096. pdf .................. 3

Vill

Miscellaneous—Continued:

Guidance for Review Staff and Industry:

Good Review Management Principles and

Practices for PDUFA Products (Apr. 2005),

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

71 Fed. Reg. 3922 (Jan. 24, 2006)

73 Fed. Reg. 49,608 (Aug. 22, 2008) 2...

H.R. Rep. No. 1980, 79th Cong., 2d Sess. (1946)..........

U.S. Dep’t of Justice, Attorney General’s Manual on

the Administrative Procedure Act (1947)...

Page

In the Supreme Court of the United States

No. 17-290

MERCK SHARP & DOHME CorP., PETITIONER

vz.

Doris ALBRECHT, ET AL.

ON WRIT OF CERTIORARI

TO THE UNITED STATES COURT OF APPEALS

FOR THE THIRD CIRCUIT

BRIEF FOR THE UNITED STATES

AS AMICUS CURIAE SUPPORTING PETITIONER

INTEREST OF THE UNITED STATES

This case concerns the circumstances under which a

decision of the Food and Drug Administration (FDA)

rejecting proposed changes to the labeling of a brand-

name drug preempts state-law tort claims that allege

that the drug manufacturer failed to provide adequate

warnings on its labe!ing. At the Court’s invitation, the

United States filed a brief as amicus curiae at the peti-

tion stage of this case.

STATEMENT

1. This case involves a series of tort claims alleging

that petitioner’s labeling for its Fosamax drug products

insufficiently warned of the drugs’ risks. Petitioner has

argued, among other things, that many of the failure-to-

warn claims are preempted because, in 2009, FDA re-

jected its attempt to strengthen relevant warnings on

(1)

2

that labeling. The regulatory scheme for drug labeling

establishes the framework for that defense.

a. Congress has charged FDA with ensuring that

each “drug is safe for use under the conditions pre-

scribed, recommended, or suggested” in its “labeling.”

21 U.S.C. 355(d); ef. 21 U.S.C. 321(m) (defining “label-

ing”); 21 U.S.C. 352(f) (misbranding). FDA regulations

govern the content and format of prescription-drug

labeling. See, eg., 21 C.F.R. 201.56, 201.57; see 21

C.F.R. 201.100(c). Those regulations are intended to

organize labeling information to more effectively com-

municate to healthcare professionals the “information

necessary for the safe and effective use of prescription

drugs.” 71 Fed. Reg. 3922, 3928 (Jan. 24, 2006). Two

separate labeling sections now generally required on

prescription-drug labeling are relevant here: the Warn-

ings and Precautions section and the Adverse Reactions

section. See 21 C.F.R. 201.57(c)(6) and (7).'

' The specific requirements for labeling content and format dis-

cussed in the text generally apply to prescription drugs subject to a

new drug application (NDA) or efficacy supplement approved on or

after June 30, 2001. 21 C.F.R. 201.56(b)(1). The specific labeling

requirements for older drug products differ in certain respects. See

21 C.F.R. 201.56(e), 201.80.

This case involves the labeling of three FDA-approved Fosamax

products: Fosamax tablets (NDA 20560, approved 1995); Fosamax

oral solution (NDA 21575, approved 2003); and Fosamax Plus D tab-

lets (NDA 21762; approved 2005). See J.A. 510-511; see also FDA,

Drug Products with Therapeutic Equivalence Evalua-

tions 3-12, 6-14 (38th ed. 2018) (listing Fosamax products). Al-

though the newer labeling requirements discussed in this brief did

not apply to all of those products at the time relevant here, no party

has suggested that the differences between the two sets of labeling

requirements are relevant to this case. The government agrees

This brief therefore follows the path taken by the court of appeals,

3

The Warnings and Precautions section must identify

“clinically significant adverse reactions” and certain

other safety hazards where “reasonable evidence of a

causal association” between the drug and such hazards

exists. 21 C.F.R. 201.57(c)(6); see 73 Fed. Reg. 49,603,

49,604 (Aug. 22, 2008) (stating that a “‘preponderance’

of evidence” is not required). FDA adopted that causal

standard in part to “prevent overwarning” of potential

risks, which, if included in the Warnings and Precau-

tions section, could dilute other “more important warn-

ings” or “deter appropriate use” of the drug. 73 Fed.

Reg. at 49,605-49,606. FDA thus reserves this section

for only a “discrete set” of hazards serious enough

to affect prescribing decisions. FDA, Guidance for

Industry: Warnings and Precautions, Contraindica-

tions, and Boxed Warning Sections of Labeling for Hu-

man Prescription Drug and Biological Products—

Content and Format 3 (Oct. 2011), https://www.fda.gov/

downloads/Drugs/GuidanceCompliance Regulatory

Information/Guidances/UCM075096.pdf.”

The Adverse Reactions section, by contrast, de-

scribes “the overall adverse reaction profile of the

drug.” 21 C.F.R. 201.57(c)(7). The causal threshold for

including an adverse reaction in this section is lower

than that for the Warnings and Precautions section: An

adverse reaction must be listed if “some basis” exists

which based its decision on the newer labeling requirements in Sec-

tion 201.57(c) without discussing Section 201.80. See Pet. App. 7a-9a

& nn.6, 9-10, 16, 51a n.130, 68a-64a nn.154, 156.

* The 2011 guidance describes FDA’s interpretation of its 2006 la-

beling regulations. FDA has informed this Office that the 2011 guid-

ance accurately reflects how FDA treated the Warnings and Pre-

cautions section during the period relevant here.

4

“to believe there is a causal relationship between the

drug and the occurrence of the adverse event.” /bid.

b. A brand-name drug “manufacturer bears respon-

sibility for the content of its labelfing] at all times.”

Wyeth v. Levine, 555 U.S. 555, 570-571 (2009); see

21 U.S.C. 355(0)(4)(1). When new information becomes

available about a new risk or a new aspect of a known

risk that causes existing labeling to become inadequate,

the manufacturer is responsible for pursuing a revision

to its labeling. See 21 C.F.R. 201.57(c)(6)(i) (stating that

updated warning must be added “as soon as” sufficient

causal evidence of a clinically significant hazard exists);

21 C.F.R. 201.57(c)(7)(ii)(B) (requiring list of adverse

reactions identified in postmarketing experience); cf.

21 C.F.R. 201.56(a)(2) (requiring updated labeling “when

new information becomes available that causes the la-

beling to become inaccurate, false, or misleading”).’

i. After FDA has approved a new drug application

(N DA) for a drug, 21 U.S.C. 355(b)(1) and (d), two mech-

anisms exist for changing a brand-name prescription

drug’s labeling, both of which require that the manufac-

turer file a supplemental NDA for FDA approval. First,

the sponsor may submit a Changes Being Effected

(CBE) supplement for certain labeling changes, which

allows the manufacturer immediately to implement its

proposed labeling changes upon FDA’s receipt of the

supplement. 21 C.F.R. 314.70(c)(6) and (iii); see Wyeth,

555 U.S. at 571, 573. A CBE supplement may be sub-

mitted, for instance, to add or strengthen a warning,

* Becaus. a generic drug’s labeling generally must track that of

its reference listed drug, generic-drug manufacturers cannot inde-

pendently change such labeling. See, eg., Mutual Pharm. Co. v.

Bartlett, 570 U.S. 472, 486, 488 (2013); PLIVA, Inc. v. Mensing,

564 U.S. 604, 613-615, 618 (2011).

5

precaution, or adverse reaction to reflect “newly ac-

quired information” if “the evidence of a causal associa-

tion satisfies the [relevant] standard for inclusion in the

labeling.” 21 C.F.R. 314.70(ce)(6)(iii A); see 21 C.F.R.

314.3(b) (defining “[njewly acquired information”). If

FDA later disapproves the supplement, however, it may

order the manufacturer to cease distributing the drug

with the labeling changes. 21 C.F.R. 314.70(c)(7).

Second, the sponsor may submit a Prior Approval

Supplement (PAS) to propose labeling changes. 21 C.F.R.

314.70(b) and (2)(v). Under that procedure, FDA ap-

proval is required before the changes are made. 21 C.F .R.

314.70(b)(3). A PAS “must be submitted” for certain

types of changes that “include, but are not limited to,”

certain labeling changes other than those described in

Section 314.70(c)(6)(iii) for CBE supplements. 21 C.F .R.

314.70(b)(1), (2), and (v)(A). Historically, FDA has also

accepted PAS applications instead of CBE supple-

ments, as occurred in this case, particularly where sig-

nificant questions exist on whether to revise or how to

modify existing drug labeling.‘

ii. “All procedures and actions that apply to an ap-

plication” submitted to FDA generally apply “to supple-

ments.” 21 C.F.R. 314.71(b) and (c). FDA has accord-

ingly confirmed to this Office that it follows many of the

general principles applicable to its review of an NDA

when undertaking the more limited task of reviewing sup-

plements that propose safety-related labeling changes.

More specifically, FDA communicates with the appli-

cant about “scientific, medical, and procedural issues that

* Cf. FDA, Guidance for Industry: Safety Labeling Changes—

Implementation of Section 505(0)(4) of the FD&C Act 7 (July 2013),

https://www.fda.gov/downloads/Drugs/GuidanceCompliance

Regulatory Information/Guidances/UCM250783. pdf.

6

arise” in the course of its review. 21 C.F.R. 314.102(a).

The “{djevelopment of final labeling” generally is then

“an iterative process between the applicant and FDA”

involving a series of communications.’ If FDA reviewers

identify “easily correctable deficiencies” in a supplement,

they will “make every reasonable effort to communicate

[them] promptly to applicants.” 21 C.F.R. 314.102(b).

And if only “editorial or similar minor deficiencies in the

| proposed] labeling” exist, FDA may approve the sup-

plement on the condition that the applicant makes ap-

propriate corrections and submits a copy of the final la-

beling before marketing the drug with that labeling.

21 C.F.R. 314.105(b).

FDA will reject a supplement, however, if the pro-

posed labeling change is false or misleading or if it does

“not comply with the requirements for labels and label-

ing in {21 C.F.R. PJart 201.” 21 C.F.R. 314.125(b)(6) and

(8). In such circumstances, FDA will send the applicant

a “complete response letter.” 21 C.F.R. 314.110(a). A

complete response letter reflects FDA’s “complete re.

view of the data submitted” and “will describe all of the

specific deficiencies that the agency has identified.”

21 C.F.R. 314.110(a)(1) and (2).

2. a. Petitioner is the manufacturer of Fosamax, a

brand-name drug that FDA approved in tablet form in

* See Center for Drug Evaluation & Research, FDA, CDER 21st

Century Review Process: Desk Reference Guide 37 (2014), https-//

www.fda gov/downloads/AboutF D A/Centers Offices/CD E R/Manua!

of Policies Procedures/UCM218757.pdf, see also FDA, Guidance for

Review Staff and Industry: Good Review Management Principles

and Practices for PDUF A Products 21 (Apr. 2005), https-//www.fda.

gov/downloads/drugs/guidancecomplianceregulatory information/

guidances/ucm0797 48. pdf (addressing “communication between the

FDA and applicants” during “labeling discussions”).

7

1995 for the treatment of osteoporosis in postmenopau-

sal women. Pet. App. 5a, 12a; see p. 2 n.1, supra. Evi-

dence later began to emerge suggesting a connection

between Fosamax and an increased risk of an unusual

type of thigh-bone fracture known as an “atypical fem-

oral fracture[],” which occurs with no or minimal exter-

nal trauma and results in a complete fracture of the fe-

mur. Pet. App. 6a, 13a-14a. Petitioner kept FDA in-

formed of those studies. /d. at 13a.

In June 2008, FDA informed petitioner that it was

aware of reports regarding the occurrence of atypical!

femoral fractures in patients using bisphosphonates

like Fosamax. C.A. App. A1935-A1936. FDA stated

that it was “concerned about this developing safety sig-

nal” and asked petitioner to submit any information it

had on the issue. /bid. Petitioner promptly complied.

Pet. App. 14a.

b. In September 2008, petitioner submitted three

Prior Approval Supplements for its three Fosamax

products that proposed changing the relevant labeling

to address atypical femoral fractures in two respects.

Pet. App. 14a-15a; see p. 2 n.1, supra.®

First, in the Adverse Reactions section (J.A. 715-

729), petitioner proposed adding a reference to “low-

energy femoral shaft fracture.” J.A. 728; see Pet. App.

16a. Petitioner also proposed including a parenthetical

cross-reference to direct readers to a discussion of

“Low-Energy Femoral Shaft Fracture” proposed for

the Warnings and Precautions section. J.A. 728 (“(see

* Relevant portions of one of the PAS applications are available at

J.A. 669-761; see C.A. App. A2697-A2928 (reproducing additional

portions). FDA has confirmed to this Office that petitioner pro-

posed the same relevant labeling language for each of its three

Fosamax products.

8

PRECAUTIONS, Low-Energy Femoral! Shaft Frac-

ture)”) (emphasis omitted).

Second, in the Warnings and Precautions section

(J.A. 703-715), petitioner proposed adding a new

subsection with an identical title: “Low-Energy Femo-

ral Shaft Fracture.” J.A. 707 (emphasis omitted); see

Pet. App. 15a. That subsection stated that “[ljow-

energy fractures of the subtrochanteric and proximal

femoral shaft have been reported in a small number of

bisphosphonate-treated patients.” J.A. 707. The pro-

posed warning added that “[sjome” of those fractures

were “stress fractures (also known as insufficiency frac-

tures),” and the remainder of petitioner’s proposed text

repeatedly referenced stress fractures. /bid.'

Petitioner supported its applications with evidence

regarding femoral fractures in Fosamax users. Pet.

App. 16a; see, eg., J.A. 745-761; C.A. App. A2748-

A2928. The applications stated, as relevant, that peti-

tioner’s use of the term “stress fracture” in connection

with reports of “low-energy subtrochanteric/mid femo-

ral shaft fractures” referred to an “insufficiency frac-

ture” that occurs with no “identifiable external trau-

matic event.” J.A. 746, 748; see J.A. 748-749, 751. The

treatment data in the applications indicated that 91% of

the fractures resulted in surgical intervention and the

other 9% involved patients who sustained only “incom-

plete stress fractures.” J.A. 753.

" An “insufficiency f[racture!’—-which can be associated with

“osteoporosis”—is “a stress fracture that occurs during norma!

stress on a bone of abnormally decreased density”; it is thus differ-

ent from the type of “stress f[racture!” experienced by athletes (a

fatigue fracture) “caused by unusual or repeated stress on a bone.”

Dorland’s Illustrated Medical Dictionary 710-711 (29th ed. 2000).

9

In May 2009, FDA issued a Complete Response Let-

ter (J.A. 510-513) informing petitioner that FDA could

not “approve the[] applications in their present form.”

J.A. 511; see Pet. App. 18a. FDA stated that it “agree/d]

that atypical and subtrochanteric fractures should be

added” to the Adverse Reactions labeling section. J.A.

511. FDA therefore recommended that petitioner mod-

ify its proposed text (J.A. 728) for that section to read

“low energy femoral shaft and subtrochanteric frac-

tures.” J.A. 512. With respect to petitioner’s Warnings

and Precautions proposal, however, FDA determined

that the “justification for the proposed [Warnings and

Precautions] section language is inadequate.” J.A. 511.

The letter also stated that “[iJdentification of ‘stress

fractures’ may not be clearly related to the atypical sub-

trochanteric fractures that have been reported in the

literature” and that “[d]iscussion of the risk factors

for stress fractures is not warranted and is not ade-

quately supported by the available literature and post-

marketing adverse event reporting.” J.A. 511-512.

In June 2009, petitioner updated the Adverse Reac-

tions section of its Fosamax labeling (J.A. 253, 267-274)

using FDA’s recommended text. J.A. 274; see J.A. 279.

Petitioner then withdrew its three pending PASs, J.A.

654-655, and submitted new CBE supplements for that

labeling change as the “quickest route to upd. e” its la-

beling, J.A. 657-658. FDA later approved these CBE

supplements.*

* See FDA, Supplement Approval 1-2 (Mar. 1, 2010), https://www.

accessdata fda gov/drugsatfda docs/appletter/2010/020560805 1805

58057 ,021575801280168018ltr.pdf (NDA 20560/S-057 and 21575/

S-018); FDA, Supplement Approval 1-2 (Mar. 1, 2010), https://www.

accessdata. fda. gov/drugsatfda docs/appletter/2010/02 17628005800

98S010itr.pdf (NDA 21762/S-010).

10

c. Nearly a year after its Complete Response Let-

ter, and after reviewing additional data submitted by

petitioner and other manufacturers, FDA issued a

Safety Announcement in March 2010 (J.A. 519-522)

stating that “FDA’s review of these data did not show

an increase in [a] risk [of atypical subtrochanteric fe-

mur fractures] in women using [bisphosphonates],” but

that FDA was working with an outside expert task force

to “gather additional information.” J.A. 519-520 (stat-

ing that the data “hajd] not shown a clear connection

between bisphosphonate use and [that] risk”); see Pet.

App. 19a.

On September 14, 2010, the task force completed its

report, which identified an apparent association be-

tween long-term bisphosphonate use and certain atypi-

cal femoral fractures. Pet. App. 20a; see J.A. 523.

d. On October 13, 2010, after studying the report,

FDA exercised its authority under 21 U.S.C. 355(0)(4)

to initiate safety-based changes to Fosamax’s labeling.

See J.A. 526-527, 546 (FDA letter dated Oct. 13, 2010).

Under Section 355(0)(4), the Secretary of Health and

Human Services, acting through FDA, “shall promptly

notify” a brand-name drug manufacturer as the “re-

sponsible person” for a drug “(i]f the Secretary becomes

aware of new safety information that the Secretary be-

lieves should be included in the labeling of the drug.”

21 U.S.C. 355(0)(4)(A); see 21 U.S.C. 355(0)(2)(A)(ii)

(defining “responsible person”). Within 30 days, the

manufacturer then must either submit a supplemental

application to change the labeling or notify the Secre-

tary that (and explain why) it believes that no change is

warranted. 21 U.S.C. 355(0)(4)(B). The Secretary must

“promptly review and act upon such supplement” and,

11

if the Secretary disagrees with the manufacturer’s la-

beling proposal or disagrees with the manufacturer's

view that no changes are warranted, the Secretary “shal!

initiate discussions” to reach agreement within 30 days

regarding a labeling change. 21 U.S.C. 355(0)(4)(C) and

(D). The Secretary may thereafter “issue an order di-

recting” the manufacturer “to make such a labeling

change as the Secretary deems appropriate to address

the new safety information.” 21 U.S.C. 355(0)(4)(E).

FDA announced its decision to the public by explain-

ing that it was requiring bisphosphonate manufacturers

to modify their labeling to include information regard-

ing the risk of such fractures in, among other places, the

Warnings and Precautions section. J.A. 246, 249. Ina

briefing to reporters, FDA’s Deputy Director for the

Office of New Drugs explained that the data that FDA

had previously reviewed was insufficient to allow the

agency to “tease out the association between [bisphos-

phonates] and these rare atypical fractures,” but that

the task force’s September 2010 report had helped FDA

“understand the[] fractures” better, “provide[d] more

information that more closely associate[d] the[] atypical

fractures with long-term bisphosphonate use,” and per-

suaded FDA that such fractures were “more closely re-

lated to these drugs * * * than[F'DA] previously had ev-

idence for.” J.A. 488-489, 493-494; see J.A. 486.

FDA’s October 2010 letter invoking Section 355(0)(4)

likewise notified petitioner that FDA “ha[d] become

aware of a possible increased risk of atypical subtro-

chanteric and diaphyseal femoral fractures in patients

taking bisphosphonates, including Fosamax,” and that

FDA now believed that new safety information about

that risk should be included in Fosamax’s labeling, in-

12

cluding changes to the Warnings and Precautions sec-

tion. J.A. 527, 528-529. In response, petitioner pro-

posed a labeling change that utilized the term “stress

fractures.” Pet. App. 22a. FDA thereafter struck those

references from petitioner’s proposal in the course

of approving the labeling change because, an FDA

employee explained, the term would suggest to most

practitioners “a minor fracture” that “would contradict

the seriousness of the atypical femoral fractures” at is-

sue. J.A. 566, 606-607; see Pet. App. 22a-23a; cf. J.A.

547, 549-628 (redline showing FDA edits to petitioner’s

proposed labeling).

3. Over 1000 plaintiffs subsequently filed separate

state-law tort actions against petitioner, alleging that

they had sustained atypical femoral fractures caused by

taking Fosamax. Pet. App. 4a, 23a. Although plaintiffs

asserted an array of tort theories, they generally al-

leged that petitioner had failed to provide adequate

warnings on its Fosamax labeling. /d. at 4a, 24a. The

cases were consolidated for pretrial proceedings as a

multidistrict litigation (MDL). Jd. at 23a.

One bellwether case within the MDL was selected

for trial on its failure-to-warn claim. Pet. App. 24a-25a

& n.64. Petitioner moved for summary judgment on

preemption grounds, but the district court did not im-

mediately rule on the motion. /d. at 163a. After a jury

rendered a verdict for petitioner on case-specific grounds,

the court rendered a post-trial decision holding that the

bellwether plaintiffs’ failure-to-warn claim was pre-

empted. /d. at 25a-26a, 153a-174a.

The district court subsequently extended that pre-

emption holding to the other MDL cases in which the

plaintiff was injured before September 14, 2010—the

13

date of the expert task force report—and granted judg-

ment to petitioner in those cases. Pet. App. 152a; see

id. at 113a-152a.

4. The court of appeals vacated and remanded. Pet.

App. la-74a.

The court of appeals concluded that its impossibility -

preemption analysis was controlled by this Court’s de-

cision in Wyeth v. Levine, supra, which the court viewed

as teaching that a drug-focused failure-to-warn claim is

preempted “where there is ‘clear evidence that the FDA

would not have approved a change’ to the label,” Pet.

App. 32a-33a (quoting Wyeth, 555 U.S. at 571). See id.

at 28a-55a. The court concluded that Wyeth’s “clear ev-

idence” discussion “announce|d] a standard of proof”

that is “synonymous with ‘clear and convincing evi-

dence’” and required proof showing it “is highly proba-

ble that the FDA would not have approved a change to

the drug’s label.” /d. at 35a, 37a; see id. at 33a-37a. The

court further concluded that the relevant preemption

determination—which involves “predicti{ing] how the

FDA would have reacted in a hypothetical scenario” in-

volving a new proposal to strengthen labeling warnings,

id. at 5la-52a—is a factual determination for a jury, not

a legal one for a court. /d. at 38a-55a.

Applying those principles, the court of appeals held

that summary judgment should not have been granted

to petitioner. Pet. App. 55a-74a. As relevant here, the

court determined that a “reasonable jury” could con-

clude that petitioner could have revised the Warnings

and Precautions section of its labeling before Septem-

ber 2010. /d. at 56a-57a, 67a. A jury, the court rea-

soned, could find that FDA’s 2009 decision to reject pe-

titioner’s proposed revision to that section was not

based on FDA’s determination that the evidence at the

14

time was insufficient to indicate that Fosamax caused

atypical femoral fractures, but rather was based on

FDA's dissatisfaction with the proposal’s use of the

term “stress fractures,” which medical practitioners

might misunderstand to refer to fractures less serious

than the femoral fractures in question, id. at 64a-66a.

See id. at 59a-68a.°

SUMMARY OF ARGUMENT

Petitioner provided FDA with the relevant scientific

data about Fosamax’s risks and, in May 2009, FDA re-

jected petitioner’s proposal to add a warning about

atypical femoral fractures in the Warnings and Precau-

tions section of Fosamax’s labeling. The court of ap-

peals erroneously rejected petitioner’s impossibility-

preemption defense to respondents’ state-law failure-

to-warn claims, based on its determination that preemp-

tion required petitioner to establish by “clear and con-

vincing” evidence that FDA would have denied a CBE

supplement, which the court deemed to be a factual

question for a jury to decide. The proper focus here is

on whether F DA’s May 2009 decision embodied a deter-

mination by FDA that insufficient causal evidence existed

to warrant strengthening the Warnings and Precau-

tions section of the Fosamax labeling to address atypi-

cal femoral fractures. That is a question of law for a

court to resolve, not a question of fact for a jury. More-

over, FDA’s 2009 decision, construed in light of the

governing regulatory regime, and FDA’s subsequent

* The court of appeals also concluded that the district court erred

in granting summary judgment on claims based on the Adverse

Reactions section of Fosamax’s labeling before its 2009 revision,

Pet. App. 69a-73a, and on non-failure-to-warn claims, id. at 74a.

15

actions show that petitioner could not have altered

F'osamax’s labeling until late 2010.

A. FDA’s May 2009 Complete Response Letter is a

legal document reflecting the agency’s exercise of legal

authority to adjudicate petitioner’s regulatory labeling

application. The meaning and effect of such agency

action is a legal question within the exclusive province

of a court. Indeed, courts have long been vested with

authority to interpret agency action as a question of law.

See, e.g., 5 U.S.C. 706. Judges, rather than lay juries,

are best suited to evaluate the scope of an agency’s de-

termination because judges are trained and experi-

enced in construing legal documents and are far better

equipped to understand agency decisions in light of the

governing statutory and regulatory context. Vesting

that legal construction in judges familiar with adminis-

trative law also fosters the type of interpretive uni-

formity appropriate when determining the scope and ef-

fect of federal agency action. That holds true even when

subsidiary factual questions are relevant to the court’s

legal determination of the meaning and effect of agency

action. Cf. Markman v. Westview Instruments, Inc.,

517 U.S. 370, 388-389 (1996).

B. The court of appeals concluded that Wyeth v.

Levine, 555 U.S. 555 (2009), required a clear-and-

convincing-evidence standard, Pet. App. 35a-37a, that a

jury must apply in this case, id. at 38a-55a. But Wyeth

addressed distinct circumstances and did not address

how to determine the meaning and effect of an actual

FDA labeling decision. The longstanding rule is that

judges, not courts, interpret such agency action as a

question of law. Nor does Wyeth’s passing reference to

“clear evidence” alter the preponderance-of-the-evidence

16

standard that applies in civil actions involving monetary

disputes between private litigants.

C. FDA’s May 2009 Complete Response Letter, the

relevant regulatory context, and the agency’r subse-

quent actions concerning Fosamax demonstrate that

FDA determined that existing information about atypical

femoral fractures was insufficient to warrant a change

to Fosamax’s Warnings and Precautions section. The

record amply demonstrates that petitioner could not

have updated its labeling until late 2010. Accordingly,

respondents’ corresponding state-law failure-to-warn

claims are preempted.

ARGUMENT

THE COURT OF APPEALS ERRED IN REJECTING

PETITIONER’S PREEMPTION DEFENSE

The court of appeals erred in holding that a jury

must determine whether FDA’s May 2009 decision—

which declined to approve petitioner’s proposal to re-

vise Fosamax’s Warnings and Precautions section to

warn against “[l]ow-energy fractures of the subtrochan-

teric and proximal femoral shaft,” J.A. 707—-preempted

respondents’ state-law failure-to-warn claims arising

from that same type of injury. Where, as here, FDA

renders a decision declining to approve a drug-labeling

change, the interpretation of that administrative deci-

sion and its significance for a failure-to-warn claim are le-

gal questions for a court to resolve, not factual ques-

tions for a jury. Moreover, because FDA’s action pre-

vented petitioner from modifying the relevant labeling

before late 2010, the court of appeals erred in rejecting

petitioner’s impossibility-preemption defense.

17

A. Petitioner’s Preemption Defense Turns On The Meaning

And Effect Of FDA’s 2009 Complete Response Letter,

Which Are Legal Questions That A Court Must Resolve

“The question for ‘impossibility’ [preemption] is

whether the private party could independently do under

federal law what state law requires of it.” PLIVA, Inc.

v. Mensing, 564 U.S. 604, (2011). A state-law failure-

to-warn claim that rests on the contention that a drug

manufacturer failed to provide adequate warnings on

drug labeling is therefore preempted if, under federal!

law, the manufacturer could not have “independently”

altered the labeling at the relevant time without FDA’s

“special permission and assistance.” /d. at 618-619, 623-

624 & n.8. As a general matter, a name-brand drug

“manufacturer may only change a drug label after the

FDA approves a supplemental application.” Wyeth v.

Levine, 555 U.S. 555, 568 (2009). When that is so, po-

tential failure-to-warn claims are preempted.

Some failure-to-warn claims, however, are not pre-

empted under the Court’s formulation in Mensing. A

manufacturer may in certain circumstances submit a

CBE supplement, which, upon its receipt by FDA, “per-

mits [the] manufacturer to make certain [safety-based |

changes to its labelling] before receiving the agency’s

approval.” Wyeth, 555 U.S. at 568 (emphasis added);

see pp. 4-5, supra (discussing CBE process). FDA may

thereafter disapprove the supplemental application and

order the manufacturer to cease distributing the drug

with such labeling changes. 21 C.F.R. 314.70(c)(7).

This Court has accordingly concluded that a brand-

name drug manufacturer will establish an impossibility-

preemption defense by sufficiently establishing that

FDA “would not have approved [the relevant] change to

18

{the drug’s] label{ing]” under its CBE regulation. Wyeth,

555 U.S. at 571; see Mensing, 564 U.S. at 624 n.8.

In this case, the court of appeals held that the rele-

vant judicial task in resolving petitioner’s preemption

defense “is to predict how the FDA would have reacted”

under the requirements of its “CBE regulation” “in a

hypothetical scenario” involving “a different label

amendment than the one it actually rejected in [its] May

2009 [Complete Response] letter.” Pet. App. 5la-52a.

That is incorrect. The threshold inquiry here is whether

F DA’s May 2009 decision embodied a determination by

FDA that insufficient causal evidence existed to warrant

strengthening Fosamax’s Warnings and Precautions

section to address atypical femoral fractures. If FDA’s

May 2009 decision did reflect such a determination, then

petitioner has established an impossibility-preemption

defense. And the meaning of FDA’s Complete Response

Letter is a question of law for a court to resolve, not a

question of fact for a jury.

I. Courts, not juries, determine the meaning of federal

agency decisions

A federal agency’s written decision on a regulatory

application is a legal document: an agency action that

embodies the agency’s exercise of legal authority to ad-

judicate the application. Cf. 5 U.S.C. 551(6) and (7) (de-

fining agency adjudication). The meaning and effect of

such agency action present legal questions within the

exclusive province of a court.

When Congress enacted the Administrative Proce-

dure Act (APA), 5 U.S.C. 551 et seq., 701 et seq., it spec-

ified that a “reviewing court shall * * * determine the

meaning or applicability of the terms of an agency ac-

tion.” 5 U.S.C. 706. That provision reflects the long-

standing view that “questions respecting the * * *

19

terms of any agency action” and the “application”

thereof are “questions of law” and therefore matters for

“courts * * * to decide.” H.R. Rep. No. 1980, 79th Cong..,

2d Sess. 44 (1946); see U.S. Dep’t of Justice, Attorney

General’s Manual on the Administrative Procedure

Act 108 (1947) (APA’s review provision “restates the

present law”).

The legal nature of that inquiry is consistent with

this Court’s precedent addressing the meaning and

effect of a prior judicial adjudication. When issues

adjudicated in prior litigation are relevant to factfinding

in a subsequent civil action, this Court has held that the

question of “[w|hat issues were decided [in that prior]

litigation is * * * a question of law” that the trial court

must itself decide by examining relevant materials from

the prior case. E’mich Motors Corp. v. General Motors

Corp., 340 U.S. 558, 571-572 (1951). That holds true

even where a jury must consider “the scope and effect

of the former judgment on the case at trial.” Jd. at 572;

see id. at 568. In such circumstances, the trial court

must first determine the prior adjudication’s scope and

effect before “instruct[ing] the jury” on its legal deter-

mination. /d. at 571. Just as the scope of a prior judicial!

adjudication is a question of law for a court to decide, so

too is the scope of a federal agency adjudication like the

one at issue here.

No sound reason exists for treating the meaning and

effect of an FDA administrative determination any dif-

ferently. Judges, rather than lay juries, are best suited

to evaluate the scope of an agency’s determination, both

because judges are trained and experienced in “(the

construction of written instruments,” Markman v.

Westview Instruments, Inc., 517 U.S. 370, 388 (1996), and

because judges are far better equipped to understand and

20

interpret agency decisions in light of the governing stat-

utory and regulatory context. Moreover, framing the

decision as a question of law to be decided by judges fa-

miliar with principles of administrative law will foster

the type of uniformity appropriate when determining

the scope and effect of federal agency action. See id. at

390-391; id. at 391 (concluding that uniformity would “be

ill served by submitting issues of document construction

to juries”). Whether the agency action is a notice-and-

comment regulation or something less formal like the

adjudicatory decision at issue here, the meaning of

agency action is a legal question that a court should de-

cide. See, e.g., Sprietsma v. Mercury Marine, 537 U.S.

51, 64-68 (2002); cf. Perez v. Mortgage Bankers Ass'n,

135 S. Ct. 1199, 1208 n.4 (2015) (“{I]t is the court that

ultimately decides wh{at] a given regulation means.”).

2. Factual findings necessary to interpret a federal

agency decision are also matters for courts to resolve

In some contexts, disputed factual questions can be

relevant to a court’s legal determination of the mean-

ing and effect of an agency decision. For example, if

FDA rejected a labeling supplement on the ground that

the information supporting it was insufficient to war-

rant a labeling change, the meaning and scope of that

decision can depend on what information FDA had be-

fore it. Yet in tort litigation between private parties

(which typically will lack compilation of an official ad-

ministrative record for an FDA decision), litigants may

dispute whether a drug manufacturer submitted all ma-

terial data to FDA. In addition, to the extent that an

agency decision read in the proper regulatory context is

insufficiently clear, a private litigant might seek to sub-

mit extrinsic material to provide further context for in-

terpreting the decision. Where consideration of such

21

material requires resolution of factual matters for a

court to determine the meaning and scope of the agency

decision, the court itself should resolve them.

This Court in Markman confronted analogous circum-

stances when it held that “the construction of a patent”

is a “purely legal” issue “exclusively within the province

of the court.” 517 U.S. at 372, 391. The Court reasoned

that “judges, not juries, are the better suited” for dis-

cerning the meaning of patent terms, even though fac-

tual questions involving “credibility determinations”

are sometimes “subsumed” within the relevant analysis.

Id. at 388-389. In other words, the “ultimate issue of

the proper construction of a [patent] claim [is] treated

as a question of law” for a court to decide, even though

“subsidiary factfinding is sometimes necessary.” Teva

Pharm. USA, Inc. v. Sandoz, Inc., 135 8S. Ct. 831, 838

(2015). The same rationale applies here. To the extent

extrinsic evidence may sometimes be relevant in litiga-

tion between private parties to determine the meaning

and effect of FDA’s agency action, the court’s evalua-

tion of such subsidiary facts does not alter the ultimate

legal character of the inquiry or the court’s exclusive

authority to resolve it.

B. The Court Of Appeals Erred In Holding That, Under

Wyeth v. Levine, A Jury Must Resolve Petitioner's

Preemption Defense As A Factual Matter Subject To A

Clear-And-Convincing-Evidence Standard

The court of appeals should have treated the inter-

pretation of FDA’s 2009 labeling decision as a legal

question for the courts to decide. As explained further

below, the basis for FDA’s 2009 Fosamax labeling deci-

sion is properly determined as a matter of law from

FDA’s Complete Response Letter, read in the context

of petitioner’s underlying labeling supplement and the

22

governing regulatory framework and related FDA ac-

tions. See pp. 30-34, infra.

The court of appeals, however, concluded that peti-

tioner’s preemption defense presented a question of

fact for a jury and this Court’s decision in Wyeth v. Lev-

ine, supra, required petitioner to provide “clear evi-

dence” from which a jury could find that FDA “would

{have} reject{ed] [the] plaintiff’s proposed warning” for

Fosamax if petitioner hac proposed that warning to

FDA. Pet. App. 33a, 54a. In addressing that “hypothet-

ical scenario,” the court of appeals determined that a

jury could reasonably conclude that “FDA rejected [peti-

tioner’s] proposed warning about femoral fractures in

2009 not because” FDA deemed the “causal link be-

tween Fosamax and fractures” to be insufficient, but be-

cause FDA was dissatisfied with petitioner’s proposed

text. Jd. at 51a, 64a-65a. There was nothing “hypothet-

ical” about FDA’s actual 2009 decision in this case,

and nothing in Wyeth addresses how courts should de-

termine the meaning and effect of such actual agency

action.

I. Wyeth did not address whether courts or juries

should construe an actual FDA decision

In Wyeth, this Court determined that a state-law

failure-to-warn claim involving a brand-name drug was

not foreclosed by the doctrine of impossibility preemp-

tion, because the drug manufacturer had a duty to en-

sure the adequacy of its own labeling and could have in-

voked FDA’s CBE regulation to update its labeling

promptly to provide additional warnings. 555 U.S. at

570-573. The Court recognized that “FDA retains au-

thority to reject [a manufacturer’s unilateral] labeling

changes made pursuant to the CBE regulation,” id. at

571, but it explained that Wyeth did “not argue” that

23

any actual FDA decision had “prohibited” it from

strengthening its labeling, id. at 572.

The Court instead viewed Wyeth as arguing that

FDA had “intended to prohibit it” from changing the la-

beling when FDA approved prior applications for the

relevant drug, a contention that the state trial and su-

preme courts had rejected. 555 U.S. at 572 & n.5. Thus,

without any relevant agency decision at hand, this

Court stated that it would not conclude that it was

impossible for Wyeth to comply with both federal and

state requirements “absent clear evidence that the

FDA would not have approved a change to [the] la-

bel{ing]” in question. /d. at 571 (emphasis added). In

other words, the Court reasoned that Wyeth needed to

make a clear showing that “FDA would have rescinded

any change in the label” that Wyeth made through the

CBE process in order to establish the impossibility of

such achange. Mensing, 564 U.S. at 624 n.8. The Court

concluded that Wyeth, which did “not argue” that it had

provided FDA with an analysis of the “specific dangers”

in question, had failed to show that “FDA would have

prevented it from adding a stronger warning.” Wyeth,

555 U.S. at 572-573.

Because Wyeth discussed the question whether FDA

would have rejected a CBE labeling change if the man-

ufacturer in that case had unilaterally made such a

change to strengthen its labeling, Wyeth did not resolve

how to determine the meaning and effect of an actual

FDA decision rejecting a proposed labeling change.

For that reason, the court of appeals erred in trans-

planting Wyeth’s discussion about the need for “clear

evidence” of what FDA “would have [done]” to this con-

text, which concerns the meaning and effect of what

FDA actually did. This Court has long cautioned that

24

it is “often misleading” to transplant “[|gjeneral expres-

sions” from one opinion “to other facts” because every

opinion must be “read in the light of the facts of the case

under discussion.” Armour & Co. v. Wantock, 323 U.S.

126, 133 (1944); see Landgraf v. USI Film Prods.,

511 U.S. 244, 266 (1994) (citing Cohens v. Virginia,

19 U.S. (6 Wheat.) 264, 399 (1821) (Marshall, C.J.)). Wy-

eth simply does not speak to the circumstances pre-

sented here.

2. Any factual findings necessary to resolve a preemption

defense need only rest on a preponderance of the

evidence

The court of appeals erroneously determined (Pet.

App. 35a-37a) that Wyeth requires “a factual showing”

of impossibility by “clear and convincing evidence” in

order to establish a preemption defense. To the extent

any factual disputes must be resolved to establish

preemption, the proper standard of proof is a prepon-

derance of the evidence.

a. The preponderance-of-the-evidence standard ap-

plies in “civil actions between private litigants unless

‘particularly important individual interests or rights are

at stake’” or a statute prescribes a different standard.

Grogan v. Garner, 498 U.S. 279, 286 (1991); see Mi-

crosoft Corp. v. i4i Ltd. P’ship, 564 U.S. 91, 100-103

(2011) (interpreting patent statute to require clear-and-

convincing evidence in certain contexts)."° Accordingly,

‘© The clear-and-convincing evidence standard applies when, for

instance, the government seeks to terminate an individual’s paren-

tal rights, Santosky v. Kramer, 455 U.S. 745, 769-770 (1982), or

seeks to deprive an individual of liberty in involuntary commitment

proceedings, Addington v. Texas, 441 U.S. 418, 433 (1979).

25

in a “typical civil case involving a monetary dispute be-

tween private parties,” the preponderance standard ap-

plies because private litigants should “share the risk of

error in roughly equal fashion.” Addington v. Texas,

441 U.S. 418, 423 (1979); see, e.g., Rivera v. Minnich,

483 U.S. 574, 576, 579 (1987) (action by mother to obtain

child-support payments by establishing paternity).

This is a typical private civil action for money dam-

ages, and nothing warrants a higher standard of proof

to resolve the preemption question here. The Court’s

preemption decisions do not suggest that a higher

evidentiary showing would generally be required where

resolution of factual issues is necessary to establish that

a state-law duty conflicts with federal law. Although the

Court has applied a presumption against implied con-

flict preemption, that presumption is a “rule of con-

struction [that] rests on an assumption about congres-

sional intent,” Arizona v. Inter Tribal Council of Ariz.,

Inc., 570 U.S. 1, 13 (2013) (emphasis added), not an evi-

dentiary principle governing the standard of proof. Cf.

Microsoft Corp., 564 U.S. at 103 (noting that the exist-

ence of a presumption alone does not “establish{| the

governing standard of proof”); Fed. R. Evid. 301 (evi-

dentiary presumptions shift the burden of production

unless a statute or rule provides otherwise).

b. The court of appeals based its “clear and convinc-

ing evidence” requirement on a single sentence in Wy-

eth, Pet. App. 35a & n.94, 37a, in which the Court stated

that it “w{ould] not conclude that it was impossible” for

Wyeth to strengthen a drug’s labeling under the CBE

process “absent clear evidence that the FDA would not

have approved a change to [the] label,” 555 U.S. at 571

(emphasis added). But Wyeth did not squarely address

or definitively resolve the relevant standard of proof

26

with the phrase “clear evidence.” Cf. Pet. App. 28a, 33a-

35a (characterizing Wyeth’s discussion as “cryptic”).

Wyeth provides no analysis or any citation to suggest

the Court believed it was imposing a higher burden of

proof with respect to factual issues, and Wyeth’s pass-

ing reference to “clear” evidence is not properly read to

require one.

This Court has used the term “clear evidence” in con-

texts in which the Court analyzes preemption issues as

a question of law, reflecting an interpretive presump-

tion against preemption. See, e.g., Geter v. American

Honda Motor Co., 529 U.S. 861, 885 (2000) (noting that

“a court should not find pre-emption too readily in the

absence of clear evidence of a conflict”); English v. Gen-

eral Elec. Co., 496 U.S. 72, 86 (1990) (“[Wle find no evi-

dence of a ‘clear and manifest’ intent on the part of Con-

gress to pre-empt tort claims like petitioner’s.”). The

Court has also held that its use of the even more specific

phrase “clear and convincing evidence” in another con-

text involving a question of law should not be under-

stood in its “strict evidentiary sense,” but merely as a

“useful reminder” that a general presumption concern-

ing a legal interpretation should control if substantial

doubt exists about “congressional intent.” Block v.

Community Nutrition Inst., 467 U.S. 340, 350-351

(1984). The Court in Wyeth appears to have similarly

used the phrase not as a shorthand for a formal clear-

and-convincing evidentiary standard, but merely to in-

dicate that a manufacturer asserting preemption must

show that a labeling change was not warranted under

the relevant statutory and regulatory framework. If

the Court had intended to depart from the preponder-

ance standard that applies to factual issues in almost all

27

civil actions between private litigants, it presumably

would have explained such an unusual departure.

3. Resolving this case does not require factual findings

a. As the case comes to this Court, the preemption

question is limited to a situation in which FDA consid-

ered the “relevant scientific data” in a prior labeling de-

cision, Pet. i. See Gov’t Cert. Amicus Br. 19 n.10; see

also Pet. Br. 35 (discussing Sup. Ct. R. 15.2). This case

is therefore properly resolved solely on the basis of a

legal interpretation of FDA’s 2009 labeling decision in

light of the governing regulatory regime and subse-

quent FDA action. Respondents argue (Br. in Opp. 26)

that, because petitioner relied below on an internal com-

pany document describing a telephone conversation with

an FDA employee about petitioner’s proposed labeling

change, see Pet. App. 17a; J.A. 764-767, a factfinder

should assess the reliability and weight of that evidence.

But assuming that it would be appropriate to rely on

such material in interpreting an agency decision, as ex-

plained above (pp. 20-21, supra), such a subsidiary fac-

tual question would be for courts to resolve in constru-

ing the meaning of FDA’s 2009 decision; and as ex-

plained below (pp. 30-34, infra), there is in any event no

need here for resort to such evidence because the mean-

ing of FDA’s 2009 decision is clear.

To be sure, an actual FDA labeling decision might

not in itself resolve preemption if, for instance, FDA did

not consider certain safety information in approving

name-brand drug labeling or in denying a labeling

change because the information was not provided to

FDA or because it arose after FDA’s decision. In such

a situation, a plaintiff could argue that information that

FDA did not consider constitutes “newly acquired infor-

mation,” 21 C.F.R. 314.3(b), 314.70(c)(6)(iii)( A), showing

28

that the drug caused a sufficiently serious hazard to

have allowed the manufacturer to update its labeling

under the CBE process. Cf. pp. 3-5, supra."' The

proper adjudication of such a contention would need to

account for the meaning and scope of any prior FDA la-

beling decision and the information that FDA previ-

ously evaluated in light of the governing statutory and

regulatory scheme (matters that the court must resolve,

see pp. 18-21, supra) in order to determine whether any

arguably new information is materially different from

the information that FDA previously determined to be

insufficient. This case, however, does not present that

question.

b. Respondents incorrectly analogize (Br. in Opp.

26; Supp. Cert. Br. 11) this case to Boyle v. United Tech-

nologies Corp., 487 U.S. 500 (1988). In Boyle, this Court

discussed the jury’s role in evaluating whether a mili-

tary contractor could raise a federal-common-law de-

fense to a state-law design-defect claim. /d. at 504, 507-

508, 514; see Buckman Co. v. Plaintiffs’ Legal Comm..,

531 U.S. 341, 347 (2001) (describing Boyle). That de-

fense required the contractor to make three factbound

showings: (1) the United States approved reasonably

'! Information—ineluding “new analyses of previously submitted

data”—will qualify as “[njewly acquired information” only if it “re-

veal|s]| risks of a different type or greater severity or frequency than

previously included in submissions to FDA.” 21 C.F.R. 314.3(b).

Accordingly, nominally “new” information concerning risks of a ma-

terially similar type, severity, and frequency as those revealed in

information previously evaluated by FDA is cumulative and not

“newly acquired information” that could justify a CBE supplement.

If for instance, FDA previously determined that that evidence of X

was insufficient to warrant a warning about risk Y, the existence of

additional but similar information about X would be insufficient to

justify a warning.

29

precise specifications; (2) the allegedly defective mili-

tary equipment conformed to those specifications; and

(3) the contractor warned the United States about the

dangers in the use of the equipment that were known to

the contractor but not to the United States. Boyle,

487 U.S. at 512. In that context, the Court concluded

that “whether the facts establish the conditions for the

defense is a question for the jury,” and it “would be er-

ror” for a reviewing court to itself determine that “the

defense had [not] been established” unless “no reason-

able jury” could have found that “that the Government

contractor defense was inapplicable.” /d. at 514.

Boyle, like Wyeth, did not involve the interpretation

of an actual federal agency decision. Boyle thus does

not address the key question here: whether the inter-

pretation of FDA’s 2009 decision is a legal inquiry for a

court. Nor did Boyle address whether a jury should re-

solve the ultimate matter of preemption when preemp-

tion turns on the application of a complex regulatory re-

gime implemented by an expert federal agency like

FDA. Cf. Riegel v. Medtronic, Inc., 552 U.S. 312, 325

(2008) (observing that a jury that “sees only the cost” of

an FDA-approved product in a lawsuit about a plain-

tiff’s injuries is unlikely to “apply cost-benefit analysis

similar to that applied by the experts at the FDA”).

Certainly, Boyle did not hold that the ultimate matter

of preemption must always rest with a jury when the

parties dispute what might be characterized as under-

lying factual issues. That question, which is not pre-

sented here, warrants further development in the lower

courts once this Court has clarified that the preemptive

effect of federal agency decisions involves legal ques-

tions for courts rather than juries.

30

C. FDA’s May 2009 Decision Rejected A Change To

Fosamax’s Warnings And Precautions Because The Data

At That Time Was Insufficient To Justify A Change

FDA’s May 2009 decision rejecting petitioner’s pro-

posal to modify Fosamax’s Warnings and Precautions

section to address atypical femoral fractures was based

on the agency’s determination that the information

about that risk was then insufficient to justify such a

warning. That conclusion flows directly from the terms

of the agency’s May 2009 Complete Response Letter,

the relevant regulatory context, and the agency’s sub-

sequent actions. Given FDA’s determination, respond-

ents’ claim that petitioner should have updated its

Warnings and Precautions labeling before late 2010 is

preempted.

FDA’s Complete Response Letter (J.A. 510-513)

shows that FDA determined that the existing data for

atypical femoral fractures was sufficient to update

Fosamax’s Adverse Reactions section, but not its Warn-

ings and Precautions section. FDA determined that

“atypical and subtrochanteric fractures should be

added” as adverse reactions, J.A. 511, reflecting that

FDA found “some basis to believe there |wa]s a causal

relationship between the drug and the occurrence of

th{at] adverse event.” 21 C.F.R. 201.57(c)(7).

Petitioner’s proposed addition to the Warnings and

Precautions section was based on the exact same risk.

Petitioner proposed a title (“Low-Energy Femoral

Shaft Fracture”) for its proposed subsection in Warn-

ings and Precautions, J.A. 707, that was identical (with

only differing capitalization) to its proposed text for the

Adverse Reactions section, J.A. 728. The proposed

warning for “{lJow-energy fractures of the subtrochan-

teric and proximal femoral shaft,” J.A. 707, likewise

31

addressed the same type of adverse reaction that FDA

concluded should be added to Fosamax’s labeling. See

J.A. 512 (FDA’s recommendation to add “low energy

femoral shaft and subtrochanteric fractures” to Ad-

verse Reaction section). And because both proposed la-

beling additions addressed the same risk, petitioner

proposed adding an explicit cross-reference in the label-

ing’s Adverse Reactions section (“[Sjee PRECAUTIONS,

Low-Energy Femoral Shaft Fracture,” J.A. 728) that

would have directed readers to the more fulsome de-

scription of the same hazard proposed for the Warnings

and Precaution section.

Under the governing regulations, however, such an

adverse reaction is to be elevated to the Warnings and

Precautions section only if “reasonable evidence of a

causal association with [the] drug” exists. 21 C.F.R.

201.57(c)(6)(i). Here, FDA rejected petitioner’s addi-

tion because the “justification for the proposed |Warn-

ings and Precautions] section language |wa]s inade-

quate.” J.A. 511 (emphasis added). Petitioner had also

proposed stating that “[slome” of the reported frac-

tures were “insufficiency” “stress fractures,” J.A. 707.

See p. 8 & n.7, supra. But FDA determined that such

fractures “may not be clearly related to the atypical

subtrochanteric fractures * * * reported in the litera-

ture,” and the associated discussion of stress-fracture

risk factors was likewise “not adequately supported by

the available literature and post-marketing adverse

event reporting.” J.A. 511-512 (emphases added).

F'DA’s decision thus was based on the lack of adequate

data to support a warning.

The court of appeals focused instead on the possibil-

ity that “FDA [might have] rejected [petitioner’s| pro-

posed warning” because of the warning’s use of the term

32

“stress fractures.” Pet. App. 64a-65a. But a Complete

Response Letter reflects “FDA’s complete review of

the data submitted”; and FDA’s regulations make clear

that the letter need not address any “proposed product

labeling” if FDA determines that the “data submitted

are inadequate.” 21 C.F.R.314.110(a)(2) and (3) (empha-

sis added). By contrast, if FDA determines that a

safe‘ y-based labeling change is warranted based on the

data, FDA will attempt promptly to identify easily cor-

rectable deficiencies in the proposed text and will develop

final labeling text with the manufacturer in an iterative

process. See pp. 5-6, supra. As a result, the May 2009

letter embodies FDA’s “recommend|ation]” that peti-

tioner modify its proposed Adverse Reactions text with

language (shown here in italics) that FDA itself pro-

posed: “low energy femoral shaft and subtrochanteric

fractures.” J.A. 512 (emphasis added); ef. J.A. 728 (pe-

titioner’s proposal). FDA made no similar suggestions

for revisions to petitioner’s proposed addition in the

Warnings and Precautions section because no warning

was justified based on the information at the time.

That understanding is reinforced by FDA’s subse-

quent actions in late 2010, when it concluded that the

Warnings and Precautions section should be revised.

FDA personnel directly edited petitioner’s proposed

language to remove stress-fracture references deemed

insufficiently clear. Pet. App. 22a-23a; see J.A. 547, 549-

628 (email with redline showing FDA edits to peti-

tioner’s proposed labeling); J.A. 606-607 (FDA edits to

proposed warning about atypical subtrochanteric and

diasphyseal femoral fractions). No sound basis thus ex-

ists for concluding that FDA determined in May 2009

that the data was sufficient to warrant a warning but

33

that it rejected petitioner’s proposal because of peti-

tioner’s proposed text. Cf. Dolin v. GlaroSmithKline

LLC, No. 17-3030, 2018 WL 4001208, at *9 (7th Cir. Aug.

22, 2018) (concluding that it would be “unreasonable” to

conclude that FDA rejected a proposed warning be-

cause the manufacturer had proposed adding it to the

wrong place in the labeling, rather than because FDA

had concluded that the warning was not warranted).

Significantly, FDA’s own regulations require that

the Warnings and Precautions section “must be re-

vised” to add sueh a clinically significant hazard “as

soon as” sufficient causal evidence exists. 21 C.F.R.

201.57(¢)(6)(i) (emphasis added). And if FDA had

“believe[d]” in May 2009 that the “new safety infor-

mation” that petitioner had submitted “should [have]

be[en| included in [Fosamax’s] labeling,” Section

355(0)(4) would have required that FDA “promptly no

tify” petitioner, 21 U.S.C. 355(0)(4)(A), and engage in ex-

pedited discussions to revise the labeling, 21 U.S.C.

355(0)(4)(B)-(D). Given the statutory and regulatory

framework, it would not be reasonable to interpret

FDA’s 2009 decision as reflecting a determination that

a new warning was justified but that FDA rejected pe-

Litioner’s proposed warning because petitioner’s pro-

posed text was inadequate.

Indeed, nearly a year later, in March 2010, FDA

announced—after reviewing further information—that

it had yet to identify an “increase in [a] risk [of atypical!

subtrochanteric femur fractures] in women using [bis -

phosphonates|.” J.A. 520. It was only in October 2010—

after an external task force had completed its report on

the issue—that FDA came to “believe that the infor-

mation” about atypical femoral fractures should be

added to the Warnings and Precautions section and

34

therefore invoked Section 355(0)(4) to revise the label-

ing for Fosamax and other bisphosphonates. See J.A.

527-528; see also pp. 10-12, supra.

In short, FDA’s 2009 Complete Response Letter

rejecting petitioner's PAS proposal to update Fosa-

max’s Warnings and Precautions section, when read in

the proper context of the governing statutory and reg-

ulatory regime—as well as FDA’s subsequent regula-

tory actions regarding Fosamax—demonstrate that, in

F DA’s judgment, an update to Fosamax’s Warnings and

Precautions section to discuss atypical femoral frac-

tures would not have been called for before late 2010.

Accordingly, it would have been impossible under fed-

eral law for petitioner to provide such a warning at an

earlier time. Any state-law failure-to-warn claim pred-

ieated on a state-law duty to provide such a warning is

therefore preempted.

35

CONCLUSION

The judgment of the court of appeals should be

reversed and the case remanded for further proceedings.

Respectfully submitted.

. —— nS JEFFREY B. WALL

Deputy General JosEPH H. HUNT

ELIZABETH H. DICKINSON nee e. omenam

Associate General

5 Deputy Solicitor General

ANTHONY A. YANG

ANNAMARIE KEMPIC

Assistant to the Solicitor

Deputy Chief Counsel, Guana

Scort R. McInTOsH

a JOSHUA M. SALZMAN

of Health and Attorneys

uman Services

SEPTEMBER 2018

* Solicitor General Noel J. Francisco is recused from this matter.

“ General Counsel Robert P. Charrow is recused from this matter.