Case: 08-31204 Document: 0051999576 Page: 1 Date Filed: 01/08/2010
IN THE UNITED STATES COURT OF APPEALS
FOR THE FIFTH CIRCUIT United States Court of Appeals
Fifth Circuit
FILED
January 8, 2010
No. 08-31204 Charles R. Fulbruge III
Clerk
JULIE DEMAHY
Plaintiff-Appellee
v.
ACTAVIS, INC., Individually and as Successor in Interest of Purepac
Pharmaceutical Company
Defendant-Appellant
Appeal from the United States District Court
for the Eastern District of Louisiana
Before KING, HIGGINBOTHAM, and CLEMENT, Circuit Judges.
PATRICK E. HIGGINBOTHAM, Circuit Judge:
This case presents one issue on appeal: whether the federal regulatory
regime governing pharmaceuticals preempts state-law failure-to-warn claims
against manufacturers of generic drugs. The Supreme Court held, in Wyeth v.
Levine, that such claims are not preempted against name brand drug
manufacturers.1 While not directing our result, it shadows our conclusion that
the federal regulatory regime governing generics is also without preemptive
effect.
1
129 S. Ct. 1187 (2009).
Case: 08-31204 Document: 0051999576 Page: 2 Date Filed: 01/08/2010
No. 08-31204
I
Julie Demahy’s physician prescribed the drug Reglan to treat her
gastroesophageal reflux. For the next four years, Demahy’s pharmacy filled her
prescription with the generic form of Reglan, metoclopramide, manufactured by
Actavis.2 Demahy alleges that its long-term ingestion caused her to develop
tardive dyskinesia, a neurological movement disorder.
The Food and Drug Administration (FDA) approved Reglan in 1980, and
Actavis began manufacturing generic metoclopramide thereafter. In 1985, the
FDA required that Reglan’s label be updated to include a warning regarding the
risk of developing tardive dyskinesia. Actavis revised its labeling to comport
with these changes to the Reglan label. There is no dispute that the generic
drug’s label was at all relevant times the same as Reglan’s. In February 2009,
the FDA issued a labeling revision for metoclopramide meant to warn of the risk
of prolonged use, defined as use for more than 12 weeks.3
Demahy asserts claims of personal injury under the Louisiana Products
Liability Act for, inter alia, failure to warn of the risks of neurological disorder
after long-term use of metoclopramide. 4 Specifically, Demahy argues that
2
Formerly Purepac Pharmaceutical Company.
3
Letter from Joyce Korvick, Deputy Dir. for Safety, Div. of Gastroenterology Products,
Center for Drug Evaluation and Research, FDA, to NDA holders for Reglan, at 3 (Feb. 26,
2009) (“Prolonged treatment (greater than 12 weeks) with metoclopramide should be avoided
in all but rare cases where therapeutic benefit is thought to outweigh the risks to the patient
of developing tardive dyskinesia.”).
4
LA . REV . STAT . 9:2800.51 et seq. The district court dismissed without prejudice
Demahy’s initial claims against defendants Wyeth, Inc., the manufacturer of brand-name
metoclopramide, and Schwarz, Inc., another generic manufacturer, after Demahy’s pharmacy
records indicated she had taken only Actavis-manufactured metoclopramide.
2
Case: 08-31204 Document: 0051999576 Page: 3 Date Filed: 01/08/2010
No. 08-31204
Actavis ignored scientific and medical literature establishing a higher risk of
developing tardive dyskinesia, failed to request a labeling revision from the
FDA, failed to change the label itself even though no prior FDA approval was
required, and failed to report safety information directly to the medical
community.
Actavis moved to dismiss Demahy’s claims, arguing that they rested on
duties imposed by state law that could not be met under federal law—that they
were conflict preempted. The district court denied the motion as to the failure-
to-warn claims.5 Since then, one sister circuit—the Eighth 6 —has considered the
issue, which has split a rapidly growing number of district courts;7 it held that
5
Actavis’s 12(b)(6) motion also argued—successfully—that Demahy’s fraud-on-the-
FDA claims were conflict preempted. The only claims at issue on appeal are those alleging a
failure to warn.
6
Mensing v. Wyeth, 588 F.3d 603 (8th Cir. 2009).
7
District court cases finding state law preempted include Gaeta v. Perrigo Pharm. Co.,
___ F. Supp. 2d ____, 2009 WL 4250690 (N.D. Cal. Nov. 24, 2009) (order denying plaintiffs’
motion for reconsideration); Smith v. Wyeth, 2009 WL 425032 (W.D. Ky. Feb. 20, 2009)
(unpublished), Morris v. Wyeth, 642 F. Supp. 2d 677 (W.D. Ky. 2009), Wilson v. Pliva, 640 F.
Supp. 2d 879 (W.D. Ky. 2009), Masterson v. Apotex Corp., 2008 WL 3262690 (S.D. Fla. Aug.
7, 2008) (unpublished), and Mensing v. Wyeth, 562 F. Supp. 2d 1056 (D. Minn. 2008),
preemption holding overruled by Mensing v. Wyeth, 588 F.3d 603 (8th Cir. 2009). Those
finding no preemption include Munroe v. Barr Labs., Inc., ___ F. Supp. ____, 2009 WL
4047949 (N.D. Fla. Oct. 15, 2009); Bartlett v. Mutual Pharm. Co., ___ F. Supp. ____, 2009 WL
3126305 (D.N.H. September 30, 2009); Stacel v. Teva Pharm., 620 F. Supp. 2d 899 (N.D. Ill.
2009), Kellogg v. Wyeth, 612 F. Supp. 2d 421 (D. Vt. 2008), Tucker v. SmithKline Beecham
Corp., 596 F. Supp. 2d 1225 (S.D. Ind. 2008) (reversing its earlier ruling dismissing the case
on preemption grounds), Barnhill v. Teva Pharm., 2007 U.S. Dist. LEXIS 44718 (S.D. Ala.
2007) (unpublished), and Laisure-Radke v. Par Pharm., Inc., 2006 WL 901657 (W.D. Wash.
March 29, 2006) (unpublished). See also McKenney v. Purepac Pharm. Co., 167 Cal. App. 4th
72 (Cal. App. 2008) (finding no preemption) Barhoum v. Barr Pharm., Inc. (N.J. Super., L.
Div., Aug. 1 2008) (unpublished) (same); Kelly v. Wyeth, Inc., 22 Mass. L. Rep. 384 (Sup. Ct.
2007) (same).
Actavis also urges this court to consider Colacicco v. Apotex, Inc., 521 F.3d 253 (3d Cir.
3
Case: 08-31204 Document: 0051999576 Page: 4 Date Filed: 01/08/2010
No. 08-31204
state tort law is not preempted.8 Appeals involving materially identical
preemption claims are now pending before the Sixth Circuit.9 Our review here
is de novo.10
II
All prescription drugs marketed in this country must first receive FDA
approval. Manufacturers of new drugs must submit a new drug application
(NDA) to the FDA that demonstrates the drug’s effectiveness and safety for its
intended use.11 The 1962 Food, Drug and Cosmetics Act (FDCA) established this
avenue for pioneer drugs, with the core objective of ensuring that drugs are both
safe and effective;12 the FDA has codified the NDA regulations at 21 C.F.R. Part
314. New drug approval requires, among other deliverables, the results of
2008) as persuasive. Colacicco, a pre-Levine case, found FDA failure-to-warn preemption as
to both generic and name brand manufacturers, relying entirely on the rationale rejected by
Levine. The Supreme Court later vacated and remanded for further consideration in light of
Levine. Colacicco v. Apotex, Inc., 129 S. Ct. 1578 (2009). Unlike the instant case, the
heightened labeling requirement allegedly required by state law in Colacicco had been
expressly considered and rejected by the FDA. 521 F.3d at 269. Given the vacatur of the Third
Circuit’s opinion, as well as the case’s factual dissimilarity, we find it neither applicable nor
persuasive here.
8
Mensing, 588 F.3d at 612.
9
Smith v. Wyeth, No. 09-5460 (6th Cir.); Wilson v. Pliva, 09-5466 (6th Cir.); Morris v.
Wyeth, 09-5509 (6th Cir.).
10
See Carden v. General Motors Corp., 509 F.3d 227, 230 (5th Cir. 2007) (citing Frank
v. Delta Airlines, Inc., 314 F.3d 195, 197 (5th Cir. 2002)).
11
See Abbreviated New Drug Application Regulations, 57 Fed. Reg. 17950 (1992)
(outlining the NDA process).
12
FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 142 (2000).
4
Case: 08-31204 Document: 0051999576 Page: 5 Date Filed: 01/08/2010
No. 08-31204
successful clinical trials 13 and labeling that accurately portrays the benefits and
risks of the drug, as indicated by those trials and other data.14 “Before approving
an NDA . . . [the] FDA undertakes a detailed review of the proposed labeling,
allowing only information for which there is a scientific basis to be included in
the FDA-approved labeling.”15 The FDA will reject the proposed labeling if
“based on a fair evaluation of all material facts, such labeling is false or
misleading in any particular.”16
Contrast this with the simpler, less demanding approval process required
of generic drugs. In 1984, Congress passed the Hatch-Waxman Amendments to
the FDCA, which altered the federal regulatory regime governing generics.
Thanks to these Amendments, once a pioneer drug loses patent protection, a
drug company may seek permission to market a generic version through a
significantly simplified process, known as the abbreviated new drug application
procedure, or ANDA.17 ANDA drugs must be the “same as” a name brand drug
that has already been approved by the FDA as to active ingredients, route of
administration, dosage form, strength, and conditions of use recommended in the
13
21 U.S.C. § 355(b)(1) (requiring certain specified data including “full reports of
investigations which have been made to show whether or not such drug is safe for use and
whether such drug is effective in use” and “a full list of the articles used as components of such
drug”). Approval is assured absent specified grounds for denial, such as a failure to “include
adequate tests . . . to show whether or not such drug is safe for use under the conditions
prescribed, recommended, or suggested in the proposed labeling thereof” or “results of such
tests [that] show [the] drug is unsafe for use under such conditions.” 21 U.S.C. § 355(d).
14
See 21 U.S.C. § 355; 21 C.F.R. § 314.105(b).
15
73 Fed. Reg. 49603, 49604 (2008).
16
See 21 U.S.C. § 355; 21 C.F.R. § 314.105(b).
17
See 57 Fed. Reg. 17950, 17951 (describing the Hatch-Waxman Amendments).
5
Case: 08-31204 Document: 0051999576 Page: 6 Date Filed: 01/08/2010
No. 08-31204
labeling.18 Under Hatch-Waxman, generic drug manufacturers need not repeat
the clinical work of their name brand counterparts, but instead must only
establish the generic drug’s bioequivalence with the name brand drug.19 By
avoiding “unnecessary,” “wasteful,” and “unethical” duplication of previously-
performed human clinical trials,20 Congress meant “to provide a careful balance
between promoting competition among pioneer . . . and generic drugs, and
encouraging research and innovation.” 21 In turn, this increased competition,
coupled with the elimination of “retesting” of a drug that has already been
determined to be safe and effective,22 would result in significant cost savings to
the American public.23 Indeed, the Congressional Budget Office estimated that
generic drugs save American consumers between $8 billion and $10 billion each
year.24 Generic drugs now account for seven out of ten prescriptions filled in the
United States.25
18
21 U.S.C. § 355(j)(2)(A)(iii).
19
21 U.S.C. § 355(j)(2)(A)(iv).
20
H.R. Rep. No. 98-857, pt. 1, at 16, 1984 U.S.C.C.A.N. at 2649.
21
Abbreviated New Drug Application Regulations, 54 Fed. Reg. 28872, 28874 (proposed
July 10, 1989). See also Mead Johnson Pharm. Group v. Bowen, 838 F.2d 1332, 1333 (D.C.
Cir. 1988) (Hatch-Waxman’s purpose “was to increase competition in the drug industry by
facilitating the approval of generic copies of drugs.”).
22
H.R. Rep. No. 98-857, pt. 1, at 16, 17
23
Mead Johnson, 838 F.2d at 1333.
24
See How Increased Competition from Generic Drugs Has Affected Prices and Returns
in the Pharmaceutical Industry, A Congressional Budget Office Study, July 1998.
25
Susan Okie, Multinational Medicines—Ensuring Drug Quality in an Era of Global
Manufacturing, 361 NEW ENG . J. MED . 737, 738 (2009).
6
Case: 08-31204 Document: 0051999576 Page: 7 Date Filed: 01/08/2010
No. 08-31204
In their application, generic manufacturers must also show “that the
labeling proposed for the new drug is the same as the labeling approved for the
listed drug.”26 Applying to market a generic drug, then, requires “[a] statement
that the applicant’s proposed labeling is the same as the labeling of the reference
listed drug except for” enumerated differences irrelevant here; without such a
statement, the FDA will deny the application.27
III
The Supreme Court ruled in Levine that the federal regulatory regime
governing pharmaceuticals does not preempt a state-law failure-to-warn claim
against the manufacturer of a name brand drug. Actavis urges that generic
drugs are different because the manufacturer of a name brand drug may change
its label unilaterally—through the “changes being effected” (CBE)
process—while seeking the FDA’s approval of the change. According to Actavis,
a generic manufacturer, in contrast, must produce the same drug and use the
same label as the name brand drug manufacturer.
The Levine Court did rely in part on the availability of the CBE process to
reject the claim—advanced by a name brand manufacturer—that it was
“impossible . . . to comply with both the state-law duties underlying those claims
and its federal labeling duties.” 28 The Court explained that once the risk to
consumers has become “apparent,” triggering a state-law duty to warn of it, “the
26
21 U.S.C. § 355(j)(2)(A)(v) (emphasis added); 21 U.S.C. § 355(j)(4)(G) .
27
Abbreviated New Drug Application Regulations, 57 Fed. Reg. 17950, 17985–86 (Apr.
28, 1992) (later codified at 21 C.F.R. §§ 314.94(a)(8)(iii)–(iv) (1993)).
28
Levine, 129 S. Ct. at 1196.
7
Case: 08-31204 Document: 0051999576 Page: 8 Date Filed: 01/08/2010
No. 08-31204
CBE regulation permit[s] [the manufacturer] to provide such a warning before
receiving the FDA’s approval.” 29 Though “the FDA retains authority to reject
labeling changes made pursuant to the CBE regulation,” the Court declined to
“conclude that it was impossible for [the manufacturer] to comply with both
federal and state requirements” without “clear evidence that the FDA would not
have approved a change” to implement the warning.30
Justice Breyer wrote separately “to emphasize the Court’s statement that
‘we have no occasion in this case to consider the pre-emptive effect of a specific
agency regulation bearing the force of law’” 31 and to accent the FDA’s ability to
“determine whether and when state tort law acts as a help or a hindrance to
achieving the safe drug-related medical care that Congress sought” through
“lawful specific regulations describing, for example, when labeling requirements
serve as a ceiling as well as a floor.” 32 Because no such regulation was at issue
in Levine, Breyer agreed with the majority that state law was not preempted.33
Actavis rightly points out that Levine is not the case before us. It does,
however, carry important implications for Actavis’s situation as well.
29
Id. at 1198.
30
Id.
31
Levine, 129 S. Ct. at 1204 (Breyer, J., concurring) (quoting id. at 1203).
32
Id.
33
Id.
8
Case: 08-31204 Document: 0051999576 Page: 9 Date Filed: 01/08/2010
No. 08-31204
IV
Here, as in every preemption case, “[t]he purpose of Congress is the
ultimate touchstone.” 34 Congressional intent to preempt state law can either be
expressed in statutory language or implied in the aim and structure of federal
law.35 Implied preemption comes in two forms: field and conflict preemption.
Field preemption is inferred where federal law is so pervasive that it leaves no
room for state supplementation.36 When Congress has not completely displaced
the possibility of state regulation, preemption may nonetheless occur when state
law “actually conflicts” with federal law.37 This conflict might be with a federal
statute or an “agency regulation with the force of law.” 38 Actavis asserts that
Demahy’s claims are conflict preempted: that it is impossible to comply with
both federal and state law,39 or, alternatively that state law poses an
unacceptable “obstacle to the accomplishment and execution of the full purposes
and objectives of Congress.”40 More specifically, Actavis contends that it is
34
Medtronic, Inc. v. Lohr, 518 U.S. 470, 485 (1996) (plurality) (internal quotation
marks omitted).
35
Hillsborough County, Fla. v. Automated Med. Labs., Inc., 471 U.S. 707, 712–13
(1985).
36
Id.
37
English v. Gen. Elec. Co., 496 U.S. 72, 79 (1990).
38
Levine, 129 S. Ct. at 1200.
39
Crosby v. Nat’l Foreign Trade Council, 530 U.S. 363, 372–73 (2000) (emphasis
added). Instances where it is impossible to comply with both federal and state law are very
rare. Fla. Lime & Avocado Growers v. Paul, 373 U.S. 132, 143 (1963).
40
Crosby, 530 U.S. at 372–73; see also Levine, 129 S. Ct. at 1193–94. But see id. at
1204–06 (Thomas, J., concurring in the judgment) (questioning “far-reaching implied pre-
9
Case: 08-31204 Document: 0051999576 Page: 10 Date Filed: 01/08/2010
No. 08-31204
impossible to comply with both the federal regulatory regime governing generic
drugs and the putative state-imposed duty to heighten warning labels, or that
Louisiana law obstructs the goals of the FDCA, as amended by the Hatch-
Waxman Amendments and implemented by FDA regulation.
Such a conclusion is not to be found lightly. As Levine reminded, “[i]n all
pre-emption cases, and particularly in those in which Congress has legislated .
. . in a field which the States have traditionally occupied,” a court must begin
this inquiry into congressional intent “with the assumption that the historic
police powers of the States were not to be superseded by the Federal Act unless
that was the clear and manifest purpose of Congress.” 41 “We rely on the
presumption because respect for the States as ‘independent sovereigns in our
federal system’ leads us to assume that ‘Congress does not cavalierly pre-empt
[state law].’” 42 Though there is ongoing disagreement among Supreme Court
jurists as to if, when, and how this presumption applies—particularly in implied
conflict preemption cases—five members of the Levine Court held that it applies
to conflict preemption cases at least where, as here, the question is whether
federal regulation of prescription drugs preempts state-law failure-to-warn
claims.43
emption doctrines” and explaining Thomas’s increasing skepticism of the Court’s “purposes
and objectives pre-emption jurisprudence”).
41
Levine, 129 S. Ct. at 1194–95 (quoting Lohr, 518 U.S. at 485) (internal citations and
quotation marks omitted); Rice v. Santa Fe Elevator Corp., 331 U.S. 218, 230 (1947).
42
Levine, 129 S. Ct. at 1195 n.3 (quoting Lohr, 518 U.S. at 485).
43
Levine, 129 S. Ct. at 1195–96 & n.3 (“[T]he dissent argues that the presumption
against pre-emption should not apply to claims of implied conflict pre-emption at all . . . but
this Court has long held to the contrary.”). But see id., 129 S. Ct. at 1229 n.14 (Alito, J.,
10
Case: 08-31204 Document: 0051999576 Page: 11 Date Filed: 01/08/2010
No. 08-31204
In this case, the bar to a finding of preemption is set even higher because
federal law provides no remedy for an injured consumer. Preemption of state
failure-to-warn claims would foreclose a remedy that was traditionally available
and for which federal law provides no substitute. Courts have been particularly
reluctant to find preemption in such cases without an unambiguous signal of
congressional intent.44 This is especially true in cases that involve health and
safety concerns, because “[s]tates traditionally have had greater latitude under
their police powers to legislate as to the protection of the lives, limbs, health,
comfort, and quiet of all persons.”45 Moreover, Congress has already expressly
dissenting) (“[I]t is not true that this Court has long applied a presumption against pre-
emption in conflict pre-emption cases.”) (internal quotation marks omitted); see also Wells
Fargo Bank of Texas, N.A. v. James, 321 F.3d 488, 491 (5th Cir. 2003). Though Justice
Thomas joined in the Levine majority’s judgment, he declined to say whether a presumption
against preemption should apply. Levine, 129 S. Ct. at 1208, n.2 (Thomas, J., concurring)
(“Because it is evident from the text of the relevant federal statutes and regulations
themselves that the state-law judgment below is not pre-empted, it is not necessary to decide
whether, or to what extent, the presumption should apply in a case such as this one, where
Congress has not enacted an express-pre-emption clause.”). In its initial brief (submitted
prior to the Supreme Court’s decision in Levine) Actavis argues that this “presumption against
preemption” does not apply to conflict preemption cases. Similarly, it appears that each
district court to have found failure-to-warn suits conflict-preempted has not applied the
presumption against preemption. See, e.g., Mensing, 562 F. Supp. 2d at 1061.
44
Lohr, 518 U.S. at 487 (“It is, to say the least, ‘difficult to believe that Congress would,
without comment, remove all means of judicial recourse for those injured by illegal conduct.’”)
(quoting Silkwood v. Kerr-McGee Corp., 464 U.S. 238, 251 (1984)); Bates v. Dow Agrosciences
LLC, 544 U.S. 431, 449 (2005) (“If Congress had intended to deprive injured parties of a long
available form of compensation, it surely would have expressed that intent more clearly.”).
See also Silkwood, 464 U.S. at 263–64 (“Because the Federal Government does not regulate
the compensation of victims, and because it is inconceivable that Congress intended to leave
victims with no remedy at all, the pre-emption analysis . . . comfortably
accommodates—indeed it compels—the conclusion that compensatory damages are not pre-
empted . . . .”) (Blackmun, J., dissenting).
45
Lohr, 518 U.S. at 485. Cf. Altria Group, Inc. v. Good, 555 U.S. ____, 129 S. Ct. 538,
543 (2008).
11
Case: 08-31204 Document: 0051999576 Page: 12 Date Filed: 01/08/2010
No. 08-31204
preempted state failure-to-warn claims for some products governed by the
FDCA—medical devices 46 —and its choice not do so for other FDA-regulated
products militates further against a finding of preemption here.47 In Levine, the
Court found Congress’s enactment of an express preemption for medical devices
telling, particularly given the historic coexistence of state tort remedies and
federal regulation of prescription drugs.48 As the Supreme Court has repeatedly
instructed, “[t]he case for federal pre-emption is particularly weak where
Congress has indicated its awareness of the operation of state law in a field of
federal interest, and has nonetheless decided to ‘stand by both concepts and to
tolerate whatever tension there [is] between them.’”49
V
Against this backdrop, Actavis first argues “that it would have been
impossible for [it] to comply with the state-law duty to modify [its product’s]
labeling without violating federal law,”50 i.e., that “‘compliance with both federal
46
Riegel v. Medtronic, Inc., 128 S. Ct. 999, 1009 (2008). In 1997, Congress preempted
certain state requirements concerning over-the-counter medications but expressly preserved
product liability actions. See 21 U.S.C. §§ 379r(e), 379s(d).
47
Levine, 129 S. Ct. at 1200.
48
Id.
49
Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 166–67 (1989) (quoting
Silkwood, 464 U.S. at 256 (1984)).
50
Levine, 129 S.Ct. at 1199.
12
Case: 08-31204 Document: 0051999576 Page: 13 Date Filed: 01/08/2010
No. 08-31204
and state [law] is a physical impossibility.’” 51 Requiring that the conflict be one
of “physical impossibility” readily suggests that this is a “demanding defense.”52
Here, Actavis urges that federal law requires that it maintain at all times a label
that is the “same as” the name brand’s, thus preventing simultaneous
compliance with a state law requiring additional warnings.
There is no dispute here that Hatch-Waxman proscribes the approval of
an application to produce a generic drug with labeling that is not the “same as”
that of the listed drug. Demahy acknowledges that a generic’s label must
initially conform to a listed drug’s and she does not allege that Actavis is liable
under state law for failure to warn adequately of the risks of tardive dyskinesia
at the time of approval. Rather, she seeks to hold Actavis liable for failing to
take steps to change the label after approval in order to provide adequate
warning once additional risks emerged. And, while Congress plainly intended
for a generic drug manufacturer to submit labeling identical to—or, the “same
as”—the brand name drug when seeking ANDA approval, the statutory scheme
“is silent as to the manufacturer’s obligations after the ANDA is granted.”53
Of course, this statutory silence does not end our preemption inquiry,
because “state laws can be pre-empted by federal regulations as well as by
51
Fid. Fed. Sav. & Loan Ass’n v. de la Cuesta, 458 U.S. 141, 153 (1982) (quoting Fla.
Lime & Avocado Growers, Inc. v. Paul, 373 U.S. 132, 142–43 (1963)). See also Barnett Bank
of Marion County, N.A. v. Nelson, 517 U.S. 25, 31 (1996) (explaining that impossibility conflict
would exist “if the federal law said, ‘you must sell insurance,’ while the state law said, ‘you
may not’”).
52
Levine, 129 S. Ct. at 1193.
53
See Bartlett, ___ F. Supp. 2d ___, 2009 WL 3126305, at *12 (quoting Stacel, 620 F.
Supp. 2d at 907); see also 21 U.S.C. § 355(j)(2)(A)(v); 21 U.S.C. § 355(j)(4)(G).
13
Case: 08-31204 Document: 0051999576 Page: 14 Date Filed: 01/08/2010
No. 08-31204
federal statutes.” 54 The regulations on which Actavis relies, however, do not
purport to bar generic labeling modifications following initial approval. Instead,
they require only that a generic’s label initially conform to the listed drug’s; if
the label does not, these regulations provide that an ANDA application will be
denied. They do not address post-approval modifications at all.55
On the contrary, “through many amendments to the FDCA and to FDA
regulations, it has remained a central premise of federal drug regulation that the
manufacturer bears responsibility for the content of its label at all times. It is
charged both with crafting an adequate label and with ensuring that its
warnings remain adequate as long as the drug is on the market.” 56 As for
maintaining an adequate label, the regulatory framework makes plain that
manufacturers—name brand and generic alike—must act to warn customers
when they learn that they may be marketing an unsafe drug. For their part,
generic manufacturers are subject to the requirement that their labeling “be
54
Hillsborough County, Fla. v. Automated Med. Labs, Inc., 471 U.S. 707, 713 (1985).
55
21 C.F.R. § 314.94(a)(8)(i) (explaining that generic drug manufacturer must provide
a copy of the currently-approved labeling for the listed name brand drug as part of its ANDA
application); 21 C.F.R. § 314.94(a)(8)(iv) (requiring that, as part of its ANDA application, a
generic manufacturer give a side-by-side comparison of its proposed labeling with the
approved labeling for the listed drug, with all differences annotated and explained); 21 C.F.R.
§ 314.127(a)(7) (noting that the FDA will not approve an ANDA application if information
submitted is “insufficient to show that the labeling proposed for the drug is the same as
labeling approved for the listed drug”); 57 Fed. Reg. 17950, 17957 cmt. 20 (“An ANDA
applicant who believes the labeling for a proposed drug product should differ from that
approved for the reference listed drug should contact FDA to discuss whether labeling for both
generic and listed drugs should be revised.”) (emphasis added).
56
Levine, 129 S. Ct. at 1197–98 (citing 21 C.F.R. § 201.80(e) and 21 C.F.R. § 314.80)(b)).
See also 21 C.F.R. § 314.80(b) (placing responsibility on the manufacturer for post-marketing
surveillance).
14
Case: 08-31204 Document: 0051999576 Page: 15 Date Filed: 01/08/2010
No. 08-31204
revised . . . as soon as there is reasonable evidence of an association of a serious
hazard with a drug.”57 Demahy claims that Actavis failed to comply with this
requirement despite reasonable evidence that long-term use of metoclopramide
poses a serious hazard. Actavis responds that this requirement is on the name
brand manufacturer alone, or that it is overridden as to generics, in light of the
putative requirement that it conform to the name brand’s label at all times.
FDA commentary supports the requirement—advanced by Demahy—that
“at a minimum a generic manufacturer should alert the agency to any new
safety hazard associated with its product.”58 First, in commentary accompanying
57
21 C.F.R. § 201.80(e). This requirement applies to drugs, such as metoclopramide,
that are not tethered to a new drug application or supplement submitted prior to July 2001.
See id. § 201.56(b).
Congress once again revised the prescription drug regime in 2007, through the Food
and Drug Administration Amendments Act. These amendments provide that, when a drug is
no longer marketed by its pioneering manufacturer (the NDA holder), ANDA holders must
submit a supplemental application proposing labeling changes to reflect new safety
information identified by the agency, or to explain why no change is warranted. 21 U.S.C. §
355(o)(4)(B). Congress was careful to remind, however, that this obligation does not “affect
the responsibility” of a generic manufacturer holding an approved ANDA “to maintain its label
in accordance with existing requirements, including subpart B of part 201” and the CBE
provision. Id. at § 355(o)(4)(I). Here, also, the FDA intended—in apparent contradiction to
the plain text of 21 C.F.R. § 201.80—to “affirm that a CBE supplement is appropriate to
amend the labeling for an approved product only to reflect newly acquired information and to
make it clear that a CBE supplement may be used to add or strengthen a contraindication,
warning, precaution, or adverse reaction only if there is sufficient evidence of a causal
association with the drug . . . .” 73 Fed. Reg. at 49604.
Though in this case the pioneering manufacturer (Wyeth) still markets the listed drug,
this provision is instructive nevertheless. If, as Actavis contends, neither this provision nor
the CBE process impose any “responsibility” on a generic manufacturer “to maintain its label,”
we are then left wondering why Congress would have deemed it necessary to clarify that the
2007 Amendments did not alter that responsibility, when we assume, as we must, that
“Congress is aware of existing law when it passes legislation.” South Dakota v. Yankton Sioux
Tribe, 522 U.S. 329, 351 (1998) (quoting Miles v. Apex Marine Corp., 498 U.S. 19, 32 (1990)
(internal quotation marks omitted)).
58
Mensing, 588 F.3d at 609.
15
Case: 08-31204 Document: 0051999576 Page: 16 Date Filed: 01/08/2010
No. 08-31204
the FDA’s implementation of the Hatch-Waxman Amendments, the FDA stated:
“After approval of an ANDA, if an ANDA holder [a generic manufacturer]
believes that new safety information should be added, it should provide adequate
supporting information to FDA, and FDA will determine whether the labeling
for the generic and listed drugs should be revised.” 59 Generic manufacturers
must also follow the same record keeping and reporting of adverse drug
experiences postmarketing as name brand manufacturers. As the FDA
explained, “ANDA applicants [must] submit a periodic report of adverse drug
experiences even if the ANDA applicant has not received any adverse drug
experience reports or initiated any labeling changes.” 60 At the very least, then,
the FDA contemplates that generic manufacturers will initiate label changes in
addition to echoing changes to the name brand label.
Nevertheless, Actavis points out that 21 U.S.C. § 355(e) “authorizes the
withdrawal of approval of an application if ‘there is a lack of substantial
evidence that the drug will have the effect it purports or is represented to have
under the conditions of use prescribed, recommended, or suggested in the
labeling thereof.’ This provision applies to both [generic] and [listed] drug
products.”61 Further, 21 C.F.R. § 314.150, explains that the FDA will initiate
proceedings to withdraw approval for a generic drug if “the labeling for the drug
. . . is no longer consistent with that for the listed drug . . . except for differences
59
57 Fed. Reg. 17950, 17961 cmt. 40 (Apr. 28, 1992) (emphasis added).
60
57 Fed. Reg. 17950, 17965 cmt. 53 (Apr. 28, 1992) (emphasis added).
61
57 Fed. Reg. at 17962 (quoting 21 U.S.C. § 355(e)).
16
Case: 08-31204 Document: 0051999576 Page: 17 Date Filed: 01/08/2010
No. 08-31204
approved in the ANDA . . . .”62 Actavis reads this language to prevent it from
revising the label of its generic metoclopramide, because doing so would result
in the FDA’s withdrawal of approval.
The FDA promulgated § 314.150 in response to comments that it “should
create a new provision authorizing the agency to withdraw an [ANDA] if the
[ANDA] holder failed to modify its labeling to match labeling changes in the
reference listed drug.”63 After being “revised . . . accordingly,” the final version
now “states that the ANDA applicant’s failure to maintain drug labeling that is
consistent with that of the listed drug may be grounds for withdrawing approval
of the [ANDA].” 64
Similarly, in response to a comment that the “FDA should create a
mechanism to compel ANDA holders to revise their labeling to conform to the
listed drug product once the ANDA is approved,” the FDA observed that 21
U.S.C. § 355(e)(2) already “authorizes the withdrawal of approval of an
application if ‘there is a lack of substantial evidence that the drug will have the
effect it purports or is represented to have under the conditions of use
prescribed, recommended, or suggested in [its] labeling.’” 65 Because this
provision applies to both listed and generic drug products and “an ANDA must
have labeling that is the same as the . . . listed drug,” the FDA explained that “a
generic drug . . . whose labeling is inconsistent with the listed drug’s labeling
62
21 C.F.R. § 314.150(b)(10) (emphasis added).
63
57 Fed. Reg. at 17970 cmt. 78.
64
Id.
65
57 Fed. Reg. 17950, 17961 (quoting 21 U.S.C. §355(e)(2)).
17
Case: 08-31204 Document: 0051999576 Page: 18 Date Filed: 01/08/2010
No. 08-31204
might not be considered safe and effective.” 66 Given that, the FDA crafted §
314.150 “to permit the agency to withdraw approval of the ANDA if an applicant
fails to maintain labeling in compliance with the requirements of the” Hatch-
Waxman Amendments.67
The overarching import of these remarks suggests, as one district court
put it, that “the purpose of [the] regulation was not to prevent a generic
manufacturer from improving or strengthening its warnings. It was, instead, to
ensure that the FDA could require a generic manufacturer to modify its labeling
to match labeling changes in the reference listed drug.”68 In Levine, the
Supreme Court found it “difficult to accept” that “the FDA would bring an
enforcement action against a manufacturer for strengthening a warning.” 69 Nor
is “a drug . . . misbranded simply because the manufacturer has altered an FDA-
approved label”; rather, the misbranding provisions concern the accuracy of the
label’s substance and the adequacy of its warnings and the FDA “contemplates
that federal juries will resolve most misbranding claims.” 70
This is not to say that generic drug manufacturers—or any drug
manufacturers, for that matter—are free to make whatever changes they see fit.
66
Id.
67
Id.
68
Bartlett, ___ F. Supp. 2d. ____, 2009 WL 3126305, at *18 (quoting Barnhill, 2007 U.S.
Dist. LEXIS 44718, at *13). But see Mensing, 562 F. Supp. 2d at 1062 (finding that this
statement “underscore[s] the notion that the ANDA drug’s label must remain the same as that
of the listed drug”).
69
Levine, 129 S. Ct. at 1197.
70
Levine, 129 S. Ct. at 1197; see also 21 U.S.C. § 355(e).
18
Case: 08-31204 Document: 0051999576 Page: 19 Date Filed: 01/08/2010
No. 08-31204
Most labeling modifications must be pursued through a “major changes”
procedure, which requires prior FDA approval before any modification takes
place.71 As the Levine Court observed, “[g]enerally speaking, a manufacturer
may only change a drug label after the FDA approves a supplemental
application.” 72 A manufacturer may, however, “make certain changes to its label
before receiving the agency’s approval,” through the so-called “changes being
effected,” or CBE, process delineated in 21 C.F.R. § 314.70(c)(iii).73
Accordingly, Demahy posits that Actavis could have complied with FDA
regulations and state law by using either the CBE process, the “major changes”
procedure (otherwise known as the prior approval process), or a third
method—warnings sent directly to healthcare providers. A finding of
preemption would require that all be foreclosed to generic manufacturers. We
consider each in turn.
A. The “Changes Being Effected” Process
Where, as here, the requisite labeling change would “add or strengthen a
contraindication, warning, precaution, or adverse reaction,” 74 to reflect
“information not previously submitted to the [FDA]” and is based on “sufficient
71
21 C.F.R. § 314.70(b)(2)(v); 21 U.S.C. § 356a(c)(1).
72
129 S. Ct. at 1196.
73
Id.
74
Id. (quoting 21 C.F.R. § 314.70(c)(6)(iii)(A)) (internal quotation marks omitted).
19
Case: 08-31204 Document: 0051999576 Page: 20 Date Filed: 01/08/2010
No. 08-31204
evidence of a causal association,” it qualifies for the CBE process,75 meaning the
manufacturer “may make the labeling change upon filing its supplemental
application with the FDA; it need not wait for FDA approval.” 76 The district
court focused much of its analysis on the availability of the CBE process to
generic manufacturers.
“Just as nothing in the text of the Hatch-Waxman Amendments forbids a
generic manufacturer from changing its drug’s label from the listed version’s
post-approval,”77 the CBE regulation also does not, on its face, distinguish
between generic and name brand drug manufacturers; that is, it does not forbid
a generic manufacturer from using the CBE process to unilaterally change a
label. Located at 21 C.F.R. § 314.70, the regulation provides that “the holder of
an approved application”—not just an approved new drug application—“may
commence distribution of the drug product involved upon receipt by the agency
of a supplement for the change.”78
Unsupported by the regulation’s text, then, Actavis must seek support
from FDA commentary for its argument that generic manufacturers cannot use
75
73 Fed. Reg. 49603, 49604 (2008) (“Expressly requiring that a CBE supplement
reflect newly acquired information and be based on sufficient evidence of a causal association
will help to ensure that scientifically accurate information appears in the approved labeling
for such products.”); 21 C.F.R. § 601.12(f)(2) (explaining that the CBE is available only when
the proposed change “reflect[s] newly acquired information”); 21 C.F.R. § 314.70(c)(6)(iii)(A)
(requiring that a CBE be supported by “evidence of a causal association[that] satisfies the
standard for inclusion in the labeling under 201.57”); 21 C.F.R. § 814.3(o) (defining “[n]ewly
acquired information” as “data, analyses, or other information not previously submitted to the
agency”).
76
Levine, 129 S. Ct. at 1196 (quoting 21 C.F.R. § 314.70(c)(6)(iii)(A)).
77
Bartlett, 2009 WL 3126305 at *14.
78
21 C.F.R. § 314.70(c)(6).
20
Case: 08-31204 Document: 0051999576 Page: 21 Date Filed: 01/08/2010
No. 08-31204
the CBE process. Specifically, when the Agency revised the CBE regulation
following the Hatch-Waxman Amendments, it reminded “ANDA applicants that,
as noted in paragraph 4 above, the labeling for an ANDA product must, with few
exceptions, correspond to that for the reference listed drug.”79
At first glance, this provision seems to provide arguable support for
Actavis’s view: the comment, unlike others it cites, refers to “the labeling for an
ANDA product,” rather than “the labeling proposed for a product in an ANDA
application” and thus could be read to mandate that a generic’s labeling remain
the “same as” the listed drug’s after approval. A closer examination—one that
accounts for the comment’s reference to “paragraph 4”—indicates, however, that
the FDA’s explanation was still fixed on the pre-approval label.
“Paragraph 4” turns down a suggestion that the FDA “accept ANDA’s with
warnings or precautions in addition to those on the reference listed drug’s label,”
by noting that “the applicant’s proposed labeling [must] be the same as that of
the listed reference drug” with exceptions not relevant here.80 So, once again we
encounter an admonition that the content of a generic drug’s labeling during the
ANDA approval process conform to that of the corresponding listed drug; there
is no direction, however, as to what may happen afterwards. It would be a
stretch then, in light of paragraph 4’s cameo in the FDA’s comment to the CBE
regulation, to read it as Actavis would have us do, and expand the directive that
“the labeling proposed for the drug [be] the same as the labeling approved for the
79
57 Fed. Reg. at 17955.
80
Id. at 17953.
21
Case: 08-31204 Document: 0051999576 Page: 22 Date Filed: 01/08/2010
No. 08-31204
listed drug,” into the post-approval period, an undemanded play with
displacement of traditional state regulation.81
In addition, these comments accompanied the only change that the FDA
made to the CBE regulations when implementing the Hatch-Waxman
Amendments: the addition of a paragraph requiring applicants to “comply with
the patent information requirements under [21 U.S.C. §355(c)(2)].” 82 Actavis’s
preferred reading of these comments is not plausible unless we accept the
proposition that the FDA would, in a minor and unrelated revision, express the
novel view that generic manufacturers are altogether forbidden from using the
CBE process.
On the other hand, what the FDA clearly did do to implement Hatch-
Waxman cuts against Actavis’s position, or at the very least muddies
congressional and agency intent: it promulgated 21 C.F.R. § 314.97, entitled
“Supplements and other changes to an approved abbreviated application.”83 This
rule provides that ANDA applicants “shall comply with the requirements of §§
314.70 and 314.71 regarding the submission of supplemental applications and
other changes to an approved abbreviated application.” 84 The requirements of
§ 314.70 include, of course, the CBE process.
81
21 U.S.C. § 355(j)(4)(G).
82
57 Fed. Reg. at 17983 (later codified at 21 C.F.R. § 314.70(e) (1993)).
83
57 Fed. Reg. at 17987. The FDA received no comments on this change.
84
21 C.F.R. § 314.97. We acknowledge that the CBE regulations appear at 21 C.F.R.
§ 314.70, which is located in Subpart B of Part 314. The FDA entitled Subpart B
“Applications,” while giving Subpart C the moniker “Abbreviated Applications.”
22
Case: 08-31204 Document: 0051999576 Page: 23 Date Filed: 01/08/2010
No. 08-31204
Had the FDA intended to deny generic manufacturers access to the CBE
procedures, notwithstanding § 314.97’s plain language, we might expect the FDA
to say so, either in § 314.97 or in the CBE provision itself. If this was the FDA’s
intent—and the industry’s understanding at the time—it passed over the
opportunity to make that clear in these provisions promulgated in Hatch-
Waxman’s wake. As the regulations stand, however, we cannot tack the words
“only when a listed drug manufacturer has first revised its label” onto the end
of § 314.97—as Actavis’s argument begs.
Nonetheless, Actavis insists that the CBE process is only available to
generic manufacturers to implement changes already made to the name brand’s
label: in other words, that § 314.97 only requires that generics follow name
brand labeling if and when a change is made by the name brand manufacturer.
For that reason, Actavis says, the FDA rejected a comment suggesting that
ANDA holders be required to submit drug labeling at periodic intervals to ensure
that the generic label matched its listed counterpart. The FDA found such a
procedure unnecessary because “existing reporting requirements at 21 C.F.R. §
314.70 [including the CBE provision] ensure that labeling changes are brought
to FDA’s attention in an appropriate and timely fashion.” 85 The FDA will then
“advise ANDA holders of changes to be made after approval, but postapproval
changes resulting from the expiration of exclusivity or patent protection are the
responsibility of the ANDA holder.”86 While this comment surely indicates that
a generic drug manufacturer should use the CBE process to enact changes
endorsed by the FDA, it does not say, however, that the process was not also
85
57 Fed. Reg. at 17961 cmt. 39.
86
Id.
23
Case: 08-31204 Document: 0051999576 Page: 24 Date Filed: 01/08/2010
No. 08-31204
intended to allow ANDA holders to act independently from the name brand
manufacturers.
Actavis also directs us to the FDA’s rejection of yet another comment as
part of administrative implementation of Hatch-Waxman. This time, it is one
that suggested empowering “ANDA applicants to deviate from the labeling for
the reference listed drug to add contraindications, warnings, precautions,
adverse reactions, and other safety-related information.87 In response, the FDA
reiterated that “the ANDA product’s labeling must be the same as the listed
drug’s product labeling because the listed drug product is the basis for ANDA
approval.” 88 Because “[c]onsistent labeling will assure physicians, health
professionals, and consumers that a generic drug is as safe and effective as its
brand-name counterpart,” the FDA instructed “an ANDA applicant [who]
believes new safety information should be added to a product’s labeling . . . [to]
contact FDA, and FDA will determine whether the labeling for the generic and
listed drugs should be revised.” 89
Nothing in this response deviates from the now-familiar distinction
between the near-unqualified ban on labeling differences pre-approval and their
availability “after approval of an ANDA.” During that latter period, “if an ANDA
holder believes that new safety information should be added, it should provide
87
57 Fed. Reg. at 17961 cmt. 40.
88
Id.
89
Id.
24
Case: 08-31204 Document: 0051999576 Page: 25 Date Filed: 01/08/2010
No. 08-31204
adequate supporting information to FDA, and FDA will determine whether the
labeling for the generic and listed drugs should be revised.”90
Actavis maintains that this passage demonstrates that, for generic drugs,
the FDA “decides whether a revision is necessary, not the drug manufacturer.”
We agree, but the FDA is the ultimate arbiter for all changes—whether
prompted by a pioneer manufacturer or a generic one. Every submitted change
requires FDA approval, even one that takes effect immediately through the CBE
process. The FDA makes these approval decisions “based on a comprehensive
scientific evaluation of the product’s risks and benefits under the conditions of
use prescribed, recommended, or suggested in the labeling.”91 Through this
review, the FDA is in effect and in fact “determin[ing] whether the labeling for
the generic and listed drugs should be revised.” This is no less true when the
FDA reviews the change only after it has been made, as is the case with CBE
changes. At best, then, this language is reason for pause, not for a conclusive
reading, or a finding of impossibility preemption.
Lastly, Actavis emphasizes the FDA’s recent statements suggesting that
generics cannot use the CBE process and that federal law preempts state-law
failure-to-warn claims brought against generics. Actavis first points to two amici
briefs filed by the FDA in Colacicco v. Apotex Corp., a case heard in the Eastern
District of Pennsylvania 92 and then in the Third Circuit.93 The briefs argued
90
Id.
91
73 Fed. Reg. at 49604.
92
Colacicco v. Apotex, Inc., 432 F. Supp. 2d 514 (E.D. Pa. 2006).
93
Colacicco v. Apotex, Inc., 521 F.3d 253 (3d Cir. 2008).
25
Case: 08-31204 Document: 0051999576 Page: 26 Date Filed: 01/08/2010
No. 08-31204
that Hatch-Waxman and implementing FDA regulations preempt state failure-
to-warn claims as to both listed and generic drug manufacturers.94 In so doing,
the FDA took the position that generic manufacturers could not utilize the CBE
process at all.95 The Third Circuit found the plaintiffs’ failure-to-warn claims
preempted.96 The Supreme Court later decided Levine, and in light of that
holding, granted certiorari in Colacicco, vacated the judgment of the Third
Cicruit, and remanded for further consideration.97 In response to Levine, the
United States withdrew as amicus in Colacicco and notified the Third Circuit
that the United States “does not take a position on whether [the state-law
failure-to-warn claims] are preempted” and “has not yet conducted an
examination of various preemption issues following the Supreme Court’s
decision in Wyeth [Levine] that would be necessary to inform a position of the
United States in this case.”98 Now withdrawn, the FDA’s amicus views are
muted and we do not consider them.
Similarly, Actavis has recently lost another pillar of support. In its briefs,
Actavis relies on a footnote in the FDA’s proposed 2008 revision to the CBE
regulations, which states that “CBE changes are not available for generic drugs
94
Brief of United States as Amicus Curiae, Colacicco v. Apotex, Inc., C.A. No. 05-5500-
MMB (E.D. Pa. filed May 10, 2006); Brief of United States as Amicus Curiae, Colacicco v.
Apotex, Inc., No. 06-3107 (3d Cir. filed Dec. 4, 2006).
95
Id. at 7–8, 18
96
Colacicco, 521 F.3d at 276.
97
Colacicco v. Apotex, No. 08-437 (Mar. 9, 2009).
98
Letter from Sharon Swingle, U.S. Department of Justice - Civil Division, Appellate,
to Marcia M. Waldron, Clerk, United States Court of Appeals for the Third Circuit (Apr. 28,
2009).
26
Case: 08-31204 Document: 0051999576 Page: 27 Date Filed: 01/08/2010
No. 08-31204
approved under an ANDA under 21 U.S.C. 355(j). To the contrary, a generic
drug manufacturer is required to conform to the approved labeling for the listed
drug.”99 But the final version of the rule omits this footnote’s language. In fact,
the final rule—which seeks to “affirm that a CBE supplement is appropriate . .
. only to reflect newly acquired information . . . and [when] there is sufficient
evidence of a causal association with the drug”—is virtually silent as to generics
altogether.
The final rule does, however, specifically note that the FDA is “amending
its regulations regarding changes to an approved [name brand] application.”100
Noting this statement, Actavis points out that the rule does not mention that the
changes affect generic applications as well. It then posits that this omission is
a sign that the CBE process simply does not apply to generics.
We might agree absent three salient insights. First, in its commentary to
the final version of the regulation, the FDA removed the footnote that explicitly
stated what Actavis asks this court to now hold: that the CBE process does not
apply to generics. Second, these regulations were promulgated prior to the
Supreme Court’s decision in Levine and since that time, the FDA has withdrawn
its amicus in Colacicco, as part of an apparent reconsideration of its preemption
position. Lastly, the 2008 final rule mentions generics only once—in a citation
after a sentence that reads: “The agency has clarified by regulation and guidance
the types of supplements that should be filed to satisfy a sponsor’s obligations
to change a drug’s labeling . . . .” In support of this statement, the FDA cites 21
99
73 Fed. Reg. at 2849 n.1.
100
73 Fed. Reg. at 49604.
27
Case: 08-31204 Document: 0051999576 Page: 28 Date Filed: 01/08/2010
No. 08-31204
C.F.R. § 314.70 (which contains the CBE regulation) and a document entitled
“Guidance for Industry: Changes to an Approved NDA or ANDA,” produced in
November 1999. The guidance document notes that “[a]ll labeling changes for
ANDA products must be consistent with [21 U.S.C. § 355(j)],” but otherwise does
not distinguish between label changes made by pioneer manufacturers and their
generic counterparts.101
Without explicit reference to the use of the CBE process by generic
manufacturers, we decline to read in a bar to its use. The FDA’s “earlier
position,” either as amicus or commentator, is instead “deprived of all claim to
deference, by the fact that it is no longer the agency's position.”102
B. Prior Approval Process
Nor does anything in the FDCA or Hatch-Waxman Amendments explicitly
forbid generic manufacturers from proposing a label change through the so-
called prior approval process.103 While FDA regulations provide for permissive
use of the CBE process for warning enhancements, and the prior approval
process is required for “major changes,” there is no indication of an agency
policy, let alone congressional intent, to prevent generic manufacturers from
proposing any and all labeling changes—no matter the significance of the
change—through the prior approval process. Rather, the regulations governing
101
Guidance for Industry: Changes to an Approved NDA or ANDA, at *24 (November
1999). The guidance document refers to section 505(j) of the FDCA, which is codified at 21
U.S.C.§ 355(j).
102
Riegel v. Medtronic, Inc., 128 S. Ct. 999, 1009 (2008).
103
21 C.F.R. § 314.70(b)(2)(v)(A).
28
Case: 08-31204 Document: 0051999576 Page: 29 Date Filed: 01/08/2010
No. 08-31204
labeling changes repeatedly use the nonrestrictive phrase “[t]hese changes
include, but are not limited to” in describing the changes manufacturers may
propose through each type of supplement.104 Indeed, manufacturers are required
to use the prior approval process for most “labeling changes.”105
C. “Dear Doctor” Letters
In addition to the CBE and prior approval processes, Demahy posits that
Actavis could have satisfied its state-law duty to warn by communicating
directly with doctors, through a “Dear Doctor” letter. These letters—addressed
to medical professionals and intended to explain the risks associated with
prolonged use of metoclopramide—would also be subject to FDA regulation
because they fall within the agency’s broad definition of “labeling.” 106 But, when
promulgating its labeling regulations well before the Hatch-Waxman
Amendments, the FDA made clear that the requirements “do not prohibit a
manufacturer . . . from warning health care professionals whenever possibly
harmful adverse effects associated with the use of the drug are discovered.”107
Thus, though generic manufacturers cannot send “Dear Doctor” letters
without prior FDA approval, they can suggest that the FDA send such letters on
104
Id. at §§ 314.70(b)(2), (c)(2), (d)(2).
105
Id. at § 314.70(b)(2)(v)(A).
106
21 C.F.R. §§ 202.1(1), (2); see also 21 C.F.R. § 200.4 (regulating content and
appearance of mailings about drugs to physicians).
107
44 Fed. Reg. 37434, 37447 (June 26, 1979).
29
Case: 08-31204 Document: 0051999576 Page: 30 Date Filed: 01/08/2010
No. 08-31204
their behalf;108 the FDA will then send letters out if it determines that they are
a necessary part of a risk evaluation and mitigation strategy.109 Under
Louisiana law, a drug manufacturer may discharge its duty to warn through
notice to the prescribing physician.110
_________
Of the three avenues for complying with both state and federal law that
Demahy identifies—the CBE process, the prior approval process, and letters sent
directly to healthcare providers—each shares the same fundamental attributes:
the manufacturer bears primary responsibility for maintaining its label
consistent with safe and effective use of its product; when reports indicate that
a label requires revision, the manufacturer must alert the FDA and provide
supporting scientific data; and the FDA then makes the decision whether such
a labeling change is supported by science. Even though with the CBE process,
the decision is made after the label has been changed, the key feature remains:
108
21 U.S.C. § 355-1(i)(2) (establishing this process as part of the 2007 amendments
to the FDCA). Although this process was not in effect when Demahy took metoclopramide, the
FDA’s later endorsement of it provides support for Actavis’s contention that Congress did not
intend that generic manufacturers send out “Dear Doctor” letters independently. Although
“[o]ccasionally, drug manufacturers and distributors mail important information about their
drugs to health care professionals” with or without FDA involvement, longstanding FDA
instructions for these mailings are directed at pioneer drug companies. NDAs: “Dear
Healthcare Professional” Letters (July 2, 2003), available at
http://www.fda.gov/cder/mapp/6020.10.pdf.
109
Id.
110
See Stahl v. Novartis Pharm. Corp., 283 F.3d 254, 266 (5th Cir. 2002) (applying
Louisiana law); see also Marks v. OHMEDA, Inc., 2003-1446, p. 11 (La. App. 3 Cir. 3/31/04);
871 So. 2d 1148, 1157; Mikell v. Hoffman-LaRoche, Inc., 94-0242, pp. 7–8 (La. App. 1 Cir.
12/22/94); 649 So. 2d 75, 79–80.
30
Case: 08-31204 Document: 0051999576 Page: 31 Date Filed: 01/08/2010
No. 08-31204
the FDA is still the final arbiter of labeling changes, while the manufacturer
retains primary responsibility for the content of its label. The federal interest
is in maintaining safe and effective labeling that is consistent across name brand
and generic bioequivalent versions of the same drug. Who prompts the FDA to
consider necessary changes to that shared label is immaterial.
At best, Actavis has demonstrated that even an interpretation of the CBE
process most favorable to Actavis (and one that fails to persuade this court) is
decidedly equivocal. Yet, equivocation falls short of the “clear and manifest
purpose of Congress” required for a finding of preemption.111 Even assuming
that the CBE regulation cannot be used by an ANDA holder to amend its label
without FDA pre-approval, Levine’s principles still apply with full force, and we
agree with Demahy that generic drug manufacturers may use two other means
of complying with both federal and state law—the prior approval process and
correspondence sent directly to healthcare providers.
Though most courts to have considered the question—including the one
below—focus on the CBE process, the CBE regulation was not the exclusive, or
even the primary, basis for rejecting preemption in Levine. Rather, the Court
explained that the brand name drug manufacturer’s quest for preemption was
grounded in a “more fundamental misunderstanding” of the regulatory regime:
that the “FDA, rather than the manufacturer, bears primary responsibility for
drug labeling.” 112 As we have seen, the opposite is true: “a central premise of
111
Levine, 129 S. Ct. at 1195 (quoting Lohr, 518 U.S. at 485).
112
Id. at 1197.
31
Case: 08-31204 Document: 0051999576 Page: 32 Date Filed: 01/08/2010
No. 08-31204
federal drug regulation [is] that the manufacturer bears responsibility for the
content of its label at all times.” 113
Actavis urges that Congress intended to preempt state law because FDA
regulations do not impose a duty on generic manufacturers to change their drug
labels. This argument does not address the question we must answer.
Preemption is not found because state law imposes duties and federal law does
not. We look instead to the state-law imposition of duties, and to whether those
duties make simultaneous compliance with federal law impossible. What the
FDA might have done once Actavis suggested these changes is immaterial to the
imposition of liability; Levine makes plain that uncertainty about the FDA’s
response makes federal preemption less likely: “absent clear evidence that the
FDA would not have approved a change to [the drug’s] label, we will not conclude
that it was impossible for the [the manufacturer] to comply with both federal
and state requirements.” 114 The record here contains nothing, let alone “clear
evidence,” that suggests the FDA would have rejected a labeling proposal from
Actavis. In fact, as discussed, the FDA mandated earlier this year that
manufacturers of metoclopramide revise their labels to disclose further risks of
tardive dyskinesia associated with long-term use.115
Finally, Actavis’s use of Justice Breyer’s concurrence in Levine is here
unavailing: he said if an agency sets a floor and a ceiling, its actions may very
113
Id. at 1197–98.
114
Id. at 1198.
115
Letter from Joyce Korvick, Deputy Dir. for Safety, Div. of Gastroenterology Products,
Center for Drug Evaluation and Research, FDA, to NDA holders for Reglan, at 3 (Feb. 26,
2009)
32
Case: 08-31204 Document: 0051999576 Page: 33 Date Filed: 01/08/2010
No. 08-31204
well enjoy preemptive effect. No such regulation bearing the force of law is
before us, while the FDA’s retreat from its earlier position on preemption and
the use of the CBE provision casts further doubt on Actavis’s argument. That
being said, had the FDA gone further than Justice Breyer would require and
specifically examined the risk of long-term use of metoclopramide at the time
that Demahy’s cause of action arose, the argument for preemption would be on
surer footing.116
VI
Even if compliance with state and federal law is not “impossible,” state law
is nonetheless preempted if it stands as an “obstacle to the accomplishment and
execution of the full purposes and objectives of Congress” as embodied in the
Hatch-Waxman Amendments and the FDCA.117 Here, if preemption is to be
found, two additional conclusions necessarily follow: first, that Congress
intended the name brand drug manufacturer to bear the sole burden of coping
116
See Dusek v. Pfizer, Inc., No. Civ. A. H-02-3559, 2004 WL 2191804, at *10 (S.D. Tex.
Feb. 20, 2004) (unpublished) (“The Court does not hold that FDA drug approvals in general
preempt failure to warn claims. The Court merely rules that permitting Plaintiffs’ claim
would be authorizing judicially what the FDA already has expressly disallowed. . . . In the face
of numerous contentions from several different sources that [the drug] should contain a
warning that [it] can cause suicide, the FDA has consistently determined no such explicit
causation admonition is justified scientifically. . . . Therefore, on the specific facts of this case
. . ., Plaintiffs’ failure to warn claim is preempted because it is in direct, actual conflict with
federal law.”). Cf. Perry v. Novartis Pharm. Corp., 456 F. Supp. 2d 678, 686–87 (E.D. Pa. 2006)
(relying on the fact that the FDA had made no conclusive determination on the specific risk
and label at issue, prior to the time the cause of action arose, in holding that the plaintiff’s
failure-to-warn claim was not preempted). See generally Catherine M. Sharkey, Products
Liability Preemption: An Institutional Approach, 76 GEO . WASH . L. REV . 449, 513–20 (2008)
(advocating for the approaches taken in Dusek and Perry).
117
Crosby, 530 U.S. at 372–73.
33
Case: 08-31204 Document: 0051999576 Page: 34 Date Filed: 01/08/2010
No. 08-31204
with incipient risks, even when it has ceased manufacturing the drug and left
the market to generics; and two, that Congress intended either that the name
brand manufacturer be liable for all failure-to-warn claims—even those arising
out of the use of generic substitutes—or, that the injured plaintiff be left with no
remedy. In assessing congressional objectives, these corollaries cannot be put
aside.
Compliance with Louisiana’s failure-to-warn law, so says Actavis, will
necessarily burden generic manufacturers with duplicative studies, trials, and
other data-gathering exercises, all putative prerequisites to placing generic
companies on notice that label changes are warranted. These burdens, if
imposed, would no doubt drive up the development costs, and thus, the market
price, of generic drugs—thwarting a key tenet of Hatch-Waxman “to make
available more low cost generic drugs by establishing a generic drug approval
procedure.”118
But what Louisiana law would require of generic drug companies does not
inevitably impose significant investment of time and money. For one, Actavis
fails to identify any statutory or regulatory provision that obligates
manufacturers to justify labeling changes through their own clinical trials or
other similarly onerous efforts.119 Yes, requests to the FDA for label changes
must be buttressed with scientific evidence, but nothing indicates that the
evidence must be—as a matter of regulatory prescription or scientific
reliability—acquired through the manufacturer’s own clinical tests. As FDA
118
H.R. Rep. No. 98-857, pt. 1, at 14, 1984 U.S.S.C.A.N. at 2647.
119
See 21 U.S.C. § 356a (containing no such requirement); 21 C.F.R. § 314.70 (same);
see also Bartlett, 2009 WL 3126305, at *24 (noting the absence of any such requirement).
34
Case: 08-31204 Document: 0051999576 Page: 35 Date Filed: 01/08/2010
No. 08-31204
regulations make plain, the regulatory regime contemplates that drug
companies will effect labeling changes without conducting new clinical trials:
“labeling shall be revised to include a warning as soon as there is reasonable
evidence of a serious hazard with a drug; a causal relationship need not have
been proved.”120 The FDA expects that such “reasonable evidence” might be
derived from, among other sources, “new clinical studies, reports of adverse
events, or new analyses of previously submitted data (e.g., meta-analyses).” 121
While clinical studies, then, may be used to support labeling changes, they are
in no way prerequisites to those changes. When the FDA itself mandated an
enhanced warning for metoclopramide in early 2009, it did not conduct its own
studies, but referenced studies published elsewhere.
Actavis concedes that generic manufacturers already must report “each
adverse drug experience that is both serious and unexpected . . . as soon as
possible but in no case later than 15 calendar days of initial receipt of the
information by the applicant,” 122 and every other “adverse drug experience . . .
at quarterly intervals, for 3 years from the date of approval . . . and then at
annual intervals.” 123 The FDA also requires that generics “develop written
120
21 C.F.R. § 201.80(e); see also 21 C.F.R. § 314.70(c)(6)(iii) (a CBE may be used “[t]o
add or strengthen a contraindication, warning, precaution, or adverse reaction for which the
evidence of a causal association satisfies the standard for inclusion in the labeling under [21
C.F.R.] § 201.57(c)”); Id. at § 201.57(c)(6) (explaining that evidence must be “reasonable” in
order to change a label’s warnings and precautions); Id. at § 201.57(c)(7) (explaining that
there must be “some basis to believe” in order to change a label’s discussion of adverse
reactions).
121
73 Fed. Reg. at 49604.
122
21 C.F.R. § 314.98 (referencing § 314.80); Id. at § 314.80 (c)(1)(i).
123
21 C.F.R. § 314.80(c)(2)(i).
35
Case: 08-31204 Document: 0051999576 Page: 36 Date Filed: 01/08/2010
No. 08-31204
procedures for the surveillance, receipt, evaluation, and reporting of
postmarketing adverse drug experiences to FDA.” 124 In Levine, the Supreme
Court found that multiple reports of adverse drug experiences provided the
substantiation necessary to justify a request for a heightened warning.125
At least in the case of metoclopramide, it seems, the allegedly higher risk
of long-term use was noted in medical literature beginning in the 1980s and
1990s.126 Even if state law prompts Actavis to alert the FDA to this information,
Actavis has provided no evidence that such collection and analysis of existing
data and conclusions would result in significant additional expenditures.
Stepping back, the duty to warn is, in general terms, predicated on the
superior knowledge of the manufacturer.127 In the world of prescription drugs,
a pharmaceutical company manufactures, and the FDA approves, a branded
drug only after extensive research and testing. Pioneer drug manufacturers
thus develop superior knowledge of their product, and the duty to warn is more
fairly imposed, as it was in Levine. Conversely, generic manufacturers
undertake limited research efforts thanks to Hatch-Waxman. They can obtain
approval for their copycat drug and label with the limited showing that their
product is the “same as” a branded drug. Consistency among generic and name
brand manufacturers not only avoids redundant research and monitoring efforts,
124
21 C.F.R. § 314.80(b).
125
Levine, 129 S. Ct. at 1197.
126
See, e.g., Kelly v. Wyeth, 2007 WL 1302589, at *2 n.7 (Mass. Super. 2007).
127
See, e.g., Greenman v. Yuba Prods., Inc., 377 P.2d 897 (Cal. 1963); MATTH EW
BEND ER 1-6 PRODUCTS LIABILITY PRACTICE GUIDE § 6.03(4)(c).
36
Case: 08-31204 Document: 0051999576 Page: 37 Date Filed: 01/08/2010
No. 08-31204
it also assures consumers and physicians that the generic product is safe and
effective.128
That difference in initial regulatory burdens marks a stark tension
between Hatch-Waxman’s quest to quickly and cheaply place generic drugs on
the market and a state law tort regime that represents the lone remedy for
individuals harmed by inadequate labeling of generic drugs. Nevertheless,
Congress did not consider the Hatch-Waxman Amendments in a vacuum, as
Actavis would have us do now. Instead, the Amendments serve as just
that—amendments—to the FDCA and to the modern regulatory regime
governing all prescription drugs that the FDCA first established. Hatch-
Waxman’s goals are thus tethered to those of the overall regulatory
scheme—chief among them the maintenance of safety and efficacy. In this wider
context, nothing about the Hatch-Waxman Amendments, and their goal of
cheaper drugs, obviates the concomitant prescription that all drugs, even
cheaper ones, remain safe. Instead, as Levine explains, “failure-to-warn actions
. . . lend force to the FDCA’s premise that manufacturers, not the FDA, bear
primary responsibility for their drug labeling at all times.” 129 “The FDA has
limited resources to monitor the 11,000 drugs on the market, and manufacturers
have superior access to information about their drugs, especially in the
postmarketing phase as new risks emerge.” 130 In passing the FDCA, Congress
128
Changes in the Labeling of ANDAs Subsequent to Revision of Innovator Labeling,
U.S. Department of Health and Human Services, Food and Drug Administration, Division of
Generic Drugs, Policy and Procedure Guide 38-89, Aug. 21, 1989.
129
Levine, 129 S. Ct. at 1202.
130
Id.
37
Case: 08-31204 Document: 0051999576 Page: 38 Date Filed: 01/08/2010
No. 08-31204
“determined that widely available state rights of action provided appropriate
relief for injured [drug] consumers” and that “state-law remedies further
consumer protection by motivating manufacturers . . . to give adequate
warnings.” 131 We see no reason why the same cannot be said for the Hatch-
Waxman Amendments to the FDCA.
Today’s inquiry does not concern the propriety of state failure-to-warn
claims. It is enough that such a claim exists; we cannot consider whether a
patient—as a policy matter—should have a state-law claim for a drug
manufacturer’s alleged failure to adequately warn of its products’ risks.
Louisiana says that she should. We have examined evidence that Congress
believed she should not. In this case, unless the law would somehow harness
liability onto name brand manufacturers for all failure-to-warn claims,
preemption in this case would leave Demahy without a remedy. Yet, “[i]f
Congress had intended to deprive [Demahy] of a long available form of
compensation, it surely would have expressed that intent more clearly.” 132 To
hold otherwise would leave us with the bizarre conclusion that Congress
intended to implicitly deprive a plaintiff whose doctor prescribes a generic drug
of any remedy, while under Levine, that same plaintiff would have a state-law
claim had she only demanded a name brand drug instead.
Our review of the applicable statutes and regulations has not provided
evidence sufficient to overcome the presumption against preemption; that is,
there is no evidence sufficient for us to say that it was the “clear and manifest
131
Id. at 1199–1200.
132
Bates, 544. U.S. at 449.
38
Case: 08-31204 Document: 0051999576 Page: 39 Date Filed: 01/08/2010
No. 08-31204
purpose” of Congress to preempt state law, or to allow the FDA to do the same.133
The presumption reflects the judiciary’s reluctance to find the intention of a
coordinate federal branch to supplant state law. The preservation of our
federalism requires Congress to do more than it—or the FDA—has chosen to do
here. We cannot make the choice for them. The need for supplanting state
duties here and the attendant calibration of costs and benefits are far beyond
judicial ken—a reality reflected in the legal demands of conflict preemption and
not wholly distant from the demands for implications of private rights of action.
Because state imposition of duties to warn on generic drug manufacturers
neither renders compliance with federal regulation impossible nor obstructs the
goals of that regulation, we AFFIRM the district court’s finding that Demahy’s
state-law failure-to-warn claims are not preempted.
133
Levine, 129 S. Ct. at 1195.
39