In the Supreme Court of the United States
ON WRIT OF CERTIORARI
TO THE SUPREME COURT OF VERMONT
BRIEF FOR THE UNITED STATES AS AMICUS CURIAE SUPPORTING PETITIONER
PAUL D. CLEMENT
Counsel of Record
GREGORY G. KATSAS
Acting Assistant Attorney
EDWIN S. KNEEDLER
Deputy Solicitor General
Assistant to the Solicitor
DOUGLAS N. LETTER
PETER R. MAIER
Department of Justice
Washington, D.C. 20530-0001
THOMAS R. BARKER
Acting General Counsel
GERALD F. MASOUDI
Associate General Counsel
WENDY S. VICENTE
Department of Health and
Washington, D.C. 20201
Whether the prescription drug labeling judgments imposed on manufacturers by the Food and Drug Ad ministration (FDA) pursuant to FDA's comprehensive safety and efficacy authority under the Federal Food, Drug, and Cosmetic Act, 21 U.S.C. 301 et seq., preempt state law product liability claims premised on the theory that different labeling judgments were necessary to make drugs reasonably safe for use.
In the Supreme Court of the United States
ON WRIT OF CERTIORARI
TO THE SUPREME COURT OF VERMONT
BRIEF FOR THE UNITED STATES AS AMICUS CURIAE SUPPORTING PETITIONER
INTEREST OF THE UNITED STATES
This case concerns the extent to which state law may hold a drug manufacturer liable for using labeling ap proved by the federal Food and Drug Administration (FDA). FDA administers the approval process for new drugs and monitors the safety of approved drugs after they have been marketed. At the Court's invitation, the United States filed a brief as amicus curiae at the peti tion stage of this case.
1. Under the Federal Food, Drug, and Cosmetic Act (FDCA or Act), 21 U.S.C. 301 et seq., a drug manufac turer may not market a new drug unless it has submit ted a new drug application to the Food and Drug Admin istration (FDA) and received the agency's approval. 21 U.S.C. 355(a). An application must contain, among other things, "the labeling proposed to be used for such drug," 21 U.S.C. 355(b)(1)(F) (Supp. V 2005); see 21 C.F.R. 314.50(c)(2)(i) and (e)(2)(ii); "full reports of investiga tions which have been made to show whether or not such drug is safe for use and whether such drug is effective in use," 21 U.S.C. 355(b)(1)(A) (Supp. V 2005); and "a dis cussion of why the benefits exceed the risks [of the drug] under the conditions stated in the labeling," 21 C.F.R. 314.50(d)(5)(viii); see 21 C.F.R. 314.50(c)(2)(ix).
The FDCA also requires that drugs not be misbran ded. 21 U.S.C. 331(a) and (b). A drug is misbranded if, among other things, its "labeling is false or misleading in any particular;" the labeling does not provide "ade quate directions for use" or certain "adequate warn ings;" or the drug "is dangerous to health when used in the dosage or manner, or with the frequency or duration prescribed, recommended, or suggested in the labeling thereof." 21 U.S.C. 352(a), (f) and (j). FDA has estab lished specific requirements for drug labeling. 21 C.F.R. Pt. 201.
FDA will approve a new drug application only if it finds, among other things, that (i) the drug is "safe for use under the conditions prescribed, recommended, or suggested in the proposed labeling thereof," (ii) there is "substantial evidence that the drug will have the effect it purports or is represented to have under the condi tions of use prescribed, recommended, or suggested in the proposed labeling thereof," and (iii) the proposed labeling is not "false or misleading in any particular." 21 U.S.C. 355(d).
After a drug has been approved and marketed, the manufacturer must investigate and report to FDA any adverse events associated with the use of the drug in humans, 21 C.F.R. 314.80, and must periodically submit any new information that may affect FDA's previous conclusions about the safety, effectiveness, or labeling of the drug, 21 C.F.R. 314.81. See 21 U.S.C. 355(k); Food and Drug Administration Amendments Act of 2007 (FDAAA), Pub. L. No. 110-85, § 901, 121 Stat. 922 (en hancing FDA's authority to require postmarket clinical studies, clinical trials, and surveillance). FDA "shall" withdraw its approval of an application if it finds, among other things, that the drug is not safe or effective under the conditions of use specified in the drug's labeling. 21 U.S.C. 355(e).
Following FDA's approval of an application, the man ufacturer generally may not make changes to the drug, including "[c]hanges in labeling," without first submit ting a supplemental application to FDA and securing the agency's prior approval for the change. 21 C.F.R. 314.70(b)(2)(v)(A). A manufacturer must submit such a supplemental application "to include a warning about a clinically significant hazard as soon as there is reason able evidence of a causal association with a drug." 21 C.F.R. 201.57(c)(6). "An applicant may ask FDA to ex pedite its review of a supplement for public health rea sons." 21 C.F.R. 314.70(b)(4). A manufacturer may, however, change a drug's labeling after its supplemental application is received by FDA, without waiting for the agency's approval of the change, if, among other things, the change "add[s] or strengthen[s]" a warning or a statement about administration of the drug in order to promote safety. 21 C.F.R. 314.70(c)(6)(iii)(A) and (C).1 FDA interprets that regulation to permit changes with out prior approval only to address "newly discovered risks" for which there is sufficient evidence of causal association with the drug. See 47 Fed. Reg. 46,623 (1982); 73 Fed. Reg. 2848 (2008) (proposing to codify that interpretation). If a manufacturer makes such a change before receiving FDA's approval, the agency may later disapprove the change and order the manufac turer to cease distribution of the changed product. 21 C.F.R. 314.70(c)(7).
2. After FDA approved petitioner's new drug appli cation for the anti-nausea drug Phenergan, petitioner informed FDA of adverse events in which Phenergan apparently was inadvertently injected intra-arterially, resulting in gangrene and amputation. See, e.g., Pet. App. 139a-140a (1967 report). Over the ensuing years, FDA and petitioner engaged in back-and-forth commu nications concerning the appropriate labeling to address the risks presented by inadvertent intra-arterial injec tion. See, e.g., id. at 141a-166a. As part of its delibera tions, FDA convened an expert advisory committee to consider that question. Id. at 144a, 147a-148a. FDA was thus fully aware of the risk of an inadvertent intra-arte rial injection, and the labeling or revised labeling it ap proved uniformly contained warnings to address that risk. See, e.g., id. at 142a-143a, 151a-154a, 162a, 165a; J.A. 271, 276, 277, 282, 283, 311-312, 356, 359, 374, 382, 390-391.
As of 2000 (when the events giving rise to this suit occurred), the FDA-approved labeling stated, in part, that "[u]nder no circumstances should Phenergan Injec tion be given by intra-arterial injection due to the likeli hood of severe arteriospasm and the possibility of resul tant gangrene." Pet. App. 167a. The labeling went on to explain that the "preferred" method of administering the drug is "by deep intramuscular injection," because intravenous administration can result, in some circum stances, in inadvertent intra-arterial injection. Ibid. For circumstances in which the drug is injected intrave nously, the labeling described in detail how such injec tion should be done, in order "to avoid * * * inadver tent intra-arterial injection." Ibid.; see id. at 4a-5a n.1.
3. In April 2000, respondent sought treatment at a health center for headache and nausea. Pet. App. 2a. The health center's staff first administered Phenergan to respondent by intra-muscular injection. Ibid. When respondent's nausea continued, the staff administered an additional dose of Phenergan later the same day by intravenous injection into her arm. Ibid. The intrave nous injection was made by a procedure the parties re fer to as IV push, whereby the staff did not drip the Phenergan solution through a free-flowing bag into a tube already inserted into respondent's arm, but instead sought to inject it directly into a vein in her arm. See id. at 2a, 52a. The IV push apparently resulted in inadver tent arterial injection, which caused gangrene and re quired amputation of respondent's hand and forearm. Id. at 2a.
Respondent brought and settled an action against the health center where she had received the injection of Phenergan. Pet. App. 50a. She also sued petitioner in a Vermont state court, asserting negligence and fail ure-to-warn claims premised on alleged inadequacies in the drug's labeling. Id. at 3a. Respondent asserted that "the label should not have allowed IV push as a means of administration, as it was safer to use other available options, such as intramuscular injection or administra tion through the tubing of a hanging IV bag." Ibid. Af ter the trial court rejected petitioner's preemption de fense, id. at 51a-65a, the jury found in respondent's fa vor, and the trial court entered judgment in the amount of $6,774,000, id. at 3a.
4. a. The Supreme Court of Vermont affirmed. Pet. App. 1a-34a. It interpreted 21 C.F.R. 314.70(c) "to allow unilateral changes to drug labels whenever the manufac turer believes it will make the product safer." Id. at 13a. The court viewed that section as crucial to its preemp tion analysis: "While specific federal labeling require ments and state common-law duties might otherwise leave drug manufacturers with conflicting obligations, [Section] 314.70(c) allows manufacturers to avoid state failure-to-warn claims without violating federal law" by making unilateral changes to FDA-approved labeling. Id. at 11a.
The Vermont Supreme Court also relied on a provi sion in the 1962 amendments to the FDCA that states that "[n]othing in the amendments * * * shall be con strued as invalidating any provision of State law * * * unless there is a direct and positive conflict between such amendments and such provision of State law." Drug Amendments of 1962, Pub. L. No. 87-781, § 202, 76 Stat. 793. The court construed that provision to limit preemption to circumstances in which it would be physi cally impossible for a manufacturer to comply with both federal and state law. Pet. App. 21a. Here, the court determined, there was no such impossibility because the record did not affirmatively show that FDA would have rejected a supplemental application seeking to streng then the warning. Id. at 17a.
b. Chief Judge Reiber dissented. Pet. App. 35a-48a. He explained that respondent's state-law claims conflict with federal law because, while "FDA concluded that the drug-with its approved methods of administration and as labeled-was both safe and effective," the "jury con cluded that the same drug-with its approved methods of administration and as labeled-was 'unreasonably dangerous.'" Id. at 35a (quoting Town of Bridport v. Sterling Clark Lurton Corp., 693 A.2d 701, 704 (Vt. 1997)). Supporting that conclusion, in the Chief Judge's view, was the fact that FDA does not merely establish minimum safety standards, but instead "balances its assessment of a drug's safety against concerns for the drug's efficacy, taking into account that a safer but less effective drug is not necessarily best for the public health overall." Id. at 47a. With respect to drug labels, the Chief Judge explained, "FDA considers not only what information to include, but also what to exclude," in part because overwarning can do more harm than good. Ibid.
The Chief Judge also took issue with the majority's understanding of Section 314.70(c). Pet. App. 39a-41a. He explained that the regulation "allow[s] manufactur ers to address newly-discovered risks," but "does not allow manufacturers to simply reassess and draw differ ent conclusions regarding the same risks and benefits already balanced by the FDA." Id. at 40a.
SUMMARY OF ARGUMENT
Respondent's claims are preempted because they challenge labeling that FDA approved after being in formed of the relevant risk.
A. FDA approves new drugs based on a thorough evaluation of their safety, efficacy, and labeling. The agency's consideration of safety and effectiveness is di rectly tied to its consideration of proposed labeling, be cause a drug's safety and effectiveness depend on the conditions under which it is used (such as its dosage, method of administration, and intended use). Many drugs can be dangerous if not used as directed, and la beling can help ameliorate risks of misuse. As part of the new-drug approval process, FDA considers and ap proves specific labeling for a drug, and the drug manu facturer is generally barred from making unilateral changes to the FDA-approved labeling.
In deciding whether to approve a drug, FDA does not merely establish minimum standards of safety. Instead, as with the Class III medical devices at issue in Riegel v. Medtronic, Inc., 128 S. Ct. 999 (2008), FDA weighs a drug's health benefits against its health risks, and "gen erally considers a drug safe when the expected thera peutic gain justifies the risk entailed by its use." United States v. Rutherford, 442 U.S. 544, 555 (1979). FDA also balances the health benefits and detriments of particular labeling, in part because labeling must strike a balance between notifying users of potential dangers and not unnecessarily deterring beneficial uses through over warning. If FDA concludes that a drug's benefits out weigh its risks only under certain conditions, the agency may require appropriate labeling to reflect that determi nation.
B. Because FDA's approval strikes a balance be tween competing considerations, state laws that strike a different balance conflict with FDA's determination and are impliedly preempted. In Riegel, this Court de termined that "[s]tate tort law that requires a manufac turer's [devices] to be safer, but hence less effective, than the model the FDA has approved disrupts the fed eral scheme." 128 S. Ct. at 1008. So too here, state tort law that required a manufacturer to use different label ing than that approved by FDA would disrupt the agency's balancing of health risks and benefits. The FDA-approved labeling gives specific instructions on how to inject Phenergan intravenously; respondent, in contrast, would impose further limits on such injec tion-limits that might harm some patients' health by restricting their physicians' treatment options. If a state regulatory agency directed drug manufacturers not to use FDA-approved labeling, the conflict with fed eral law would be manifest. As in Riegel, the fact that juries instead of an expert agency would second-guess FDA's judgments in individual cases only exacerbates the conflict.
C. The Vermont Supreme Court opined that an FDA regulation, 21 C.F.R. 314.70(c), "allow[s] unilateral changes to drug labels whenever the manufacturer be lieves [the changes] will make the product safer." Pet. App. 13a. That interpretation of the regulation is wrong, because Section 314.70(c) permits manufacturers to make changes to the labeling, subject to FDA's subse quent review and approval, based only on newly avail able information, not based on information that was pre viously submitted to FDA-let alone whenever the man ufacturer believes a different label "will make the prod uct safer." If manufacturers were free to make unilat eral changes to labeling the day after FDA's approval, based on information that was previously available to FDA, the approval process would be undermined and the agency's careful balancing of risks and benefits as reflected in the labeling would be thwarted. Here, there is no question that petitioner informed FDA of the rele vant risk, and FDA determined that Phenergan was safe and effective under the conditions set forth in the label ing, including intravenous administration.
D. The Vermont Supreme Court also erred in con cluding that normal conflict-preemption principles do not apply to the FDCA. The 1962 amendments to the FDCA provide that they should not "be construed as invalidating any provision of State law * * * unless there is a direct and positive conflict between such amendments and such provision of State law." Pub. L. No. 87-781, § 202, 76 Stat. 793 (emphasis added). That provision means that the relevant amendments do not give rise to field preemption; it does not express any intent to preserve state laws that conflict with federal law, as Vermont tort law does in this case.
THE FDCA PREEMPTS TORT CLAIMS THAT WOULD IM POSE LIABILITY FOR THE USE OF LABELING THAT THE FOOD AND DRUG ADMINISTRATION APPROVED AFTER BEING INFORMED OF THE RELEVANT RISK
Federal law preempts state laws that conflict with federal law, including state laws that either "make it 'im possible' for private parties to comply with both state and federal law," Geier v. American Honda Motor Co., 529 U.S. 861, 873 (2000), or that "stand as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress," Hines v. Davidowitz, 312 U.S. 52, 67 (1941). Because respondent's claims chal lenge labeling that FDA approved after being informed of the relevant risk (and that petitioner was not free to change in the manner urged by respondent without FDA's prior approval), they conflict with FDA's ap proval of the labeling and are preempted.
A. FDA's Approval Of A Drug, Including Its Labeling, Re flects The Agency's Expert Weighing Of The Health Risks And Benefits Of The Drug As Labeled
1. FDA may approve a new drug application only if it determines, among other things, that (i) the drug is "safe for use under the conditions prescribed, recom mended, or suggested in the proposed labeling thereof," (ii) there is "substantial evidence that the drug will have the effect it purports or is represented to have under the conditions of use prescribed, recommended, or sug gested in the proposed labeling thereof," and (iii) the proposed labeling is not "false or misleading in any par ticular." 21 U.S.C. 355(d). The agency's consideration of safety and effectiveness is directly tied to its consid eration of "the proposed labeling," ibid., because a drug's safety and effectiveness usually depend on the conditions under which it is used (such as its dosage, its method of administration, and its intended use). Thus, "[d]rug labeling serves as the standard under which FDA determines whether a product is safe and effec tive." 50 Fed. Reg. 7470 (1985). Labeling is "[t]he cen terpiece of risk management," as it "communicates to health care practitioners the agency's formal, authorita tive conclusions regarding the conditions under which the product can be used safely and effectively." 71 Fed. Reg. 3934 (2006); cf. Riegel v. Medtronic, Inc., 128 S. Ct. 999, 1004 (2008) ("FDA evaluates safety and effective ness [of Class III medical devices] under the conditions of use set forth on the label.").
FDA's review of a new drug application is similar to its premarket approval process for Class III medical de vices, see 60 Fed. Reg. 39,180 (1995), which this Court has described as "rigorous," Riegel, 128 S. Ct. at 1004; see id. at 1018 (Ginsburg, J., dissenting); Medtronic, Inc. v. Lohr, 518 U.S. 470, 477 (1996). "Under the FDCA, a drugmaker must submit research data to FDA at two general stages of new-drug development." Merck KGaA v. Integra Lifesciences I, Ltd., 545 U.S. 193, 196 (2005). A manufacturer first submits an investigational new drug application seeking authorization to conduct clinical trials (i.e., trials on humans) in order to investi gate the safety and effectiveness of the drug. See 21 U.S.C. 355(i) (2000 & Supp. V 2005); 21 C.F.R. 312.20. In determining whether to permit clinical trials to pro ceed, FDA considers whether "the drug involved repre sents an unreasonable risk to the safety of the persons who are the subjects of the clinical investigation." 21 U.S.C. 355(i)(3)(B)(i). The investigational new drug ap plication must be supported by pre-clinical research re garding the safety and efficacy of the drug, including "pharmacological and toxicological studies of the drug involving laboratory animals or in vitro." 21 C.F.R. 312.23(a)(8); see generally 21 C.F.R. 312.23(a).
If clinical trials demonstrate safety and efficacy, a manufacturer may submit a new drug application seek ing approval to market the drug. See 21 U.S.C. 355(b) (2000 & Supp. V 2005). The applicant must submit "the labeling proposed to be used for such drug," 21 U.S.C. 355(b)(1)(F) (Supp. V 2005), as well as extensive infor mation about the composition, manufacture, and specifi cation of the drug, the pre-clinical and clinical studies, and "any other data or information relevant to an evalu ation of the safety and effectiveness of the drug product obtained or otherwise received by the applicant from any source." 21 C.F.R. 314.50(d)(5)(iv); see 21 U.S.C. 355(b)(1)(A) (Supp. V 2005). FDA has issued numerous guidance documents that describe, in detail, how to pre pare a new drug application. See FDA, Drug Applica tions, New Drug Application (NDA) Process (last modi fied Dec. 28, 2007) <http://www.fda.gov/cder/ regulatory/applications/NDA.htm> (listing the guidance documents).
FDA must deny any application that does not "in clude adequate tests by all methods reasonably applica ble to show whether or not such drug is safe for use un der the conditions prescribed, recommended, or sug gested in the proposed labeling." 21 U.S.C. 355(d)(1). There must also be "substantial evidence" of the drug's effectiveness, with "substantial evidence" defined to be "adequate and well-controlled investigations, including clinical investigations, by experts qualified by scientific training and experience to evaluate the effectiveness of the drug involved, on the basis of which it could fairly and responsibly be concluded by such experts that the drug will have the effect it purports or is represented to have." 21 U.S.C. 355(d).
FDA's "rigorous evaluation process * * * scruti nizes everything about the drug-from the design of clinical trials to the severity of side effects to the condi tions under which the drug is manufactured." FDA, The FDA's Drug Review Process: Ensuring Drugs Are Safe and Effective (Review Process) (visited June 2, 2008) <http://www.fda.gov/fdac/special/testtubetopatient/ drugreview.html>. "[A]n FDA review team-medical doctors, chemists, statisticians, microbiologists, pharma cologists, and other experts-evaluates whether the studies the sponsor submitted show that the drug is safe and effective for its proposed use." Ibid. In order to ensure that FDA has all the information it needs, the agency "usually communicates often with sponsors about scientific, medical, and procedural issues that arise dur ing the review process." FDA, The CDER Handbook 24 (1998) <http://www.fda.gov/cder/handbook/handbook. pdf>; see 21 C.F.R. 314.50(f)(4), 314.102(a), (c), (d), and (e). FDA may also consult with independent panels of scientific experts. 21 U.S.C. 355(n). FDA is more likely to consult with such an independent advisory committee if a drug is the first in its class. Review Process, supra.
After a drug is approved, the manufacturer must investigate and report adverse events to FDA. 21 C.F.R. 314.80. If such an event is serious and unex pected, the report must be made "as soon as possible but in no case later than 15 calendar days of initial receipt of the information by the applicant." 21 C.F.R. 314.80(c)(1)(i). Certain other information must be sub mitted to FDA within three working days. 21 C.F.R. 314.81(b)(1). A variety of other postmarketing reports are filed periodically, including quarterly and annual reports analyzing adverse events, 21 C.F.R. 314.80(c)(2), 314.81, and annual reports disclosing, among other things, all "significant new information * * * that might affect the safety, effectiveness, or labeling of the drug," including any new studies and data, 21 C.F.R. 314.81(b)(2)(i), (v), (vi). FDA also receives reports from health professionals and members of the public. FDA, Postmarketing Surveillance Programs (last modified Apr. 9, 2004) <http://www.fda.gov/cder/regulatory/ap plications/Postmarketing/surveillancepost.htm>.
Following its approval of a drug, FDA "monitors ad verse events" from various reports, and "uses this infor mation to update drug labeling." Postmarketing Sur veillance Programs, supra. In addition, FDA "shall" withdraw its approval of an application if it finds, among other things, that the drug is not safe or effective under the conditions of use specified in the drug's labeling. 21 U.S.C. 355(e).
2. In determining whether to grant or continue its approval of a new drug application, FDA does not merely impose minimum standards of safety, as the Ver mont Supreme Court concluded. See Pet. App. 19a. "No drug is absolutely safe; all drugs have side effects." Re view Process, supra; see FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 142 (2000) ("[V]irtually ev ery drug or device poses dangers under certain condi tions."). Thus, FDA weighs health benefits against health risks. See 71 Fed. Reg. at 3934; 60 Fed. Reg. at 39,180; 47 Fed. Reg. 39,149 (1982). And FDA "generally considers a drug safe when the expected therapeutic gain justifies the risk entailed by its use." United States v. Rutherford, 442 U.S. 544, 555 (1979); accord Brown & Williamson, 529 U.S. at 140; Review Process, supra. The agency has, for example, approved cancer treat ments that are highly toxic and thus not "safe" as that term is ordinarily used, but that are nonetheless safe in the relevant sense because the potential benefits to the health of cancer patients outweigh the risks. 61 Fed. Reg. 44,413 (1996); see Brown & Williamson, 529 U.S. at 142.
FDA's balancing of a drug's risks and benefits is sim ilar to the balancing it undertakes in the analogous con text of Class III medical devices. As this Court recently explained, FDA weighs the benefits of a Class III device against its risks, and FDA "may * * * approve devices that present great risks if they nonetheless offer great benefits in light of available alternatives." Riegel, 128 S. Ct. at 1004. So too with drugs. FDA's risk-benefit balancing looks in part to the availability of more effec tive or less risky alternatives. If similar, safer products are already on the market, the agency may require a heightened health benefit to justify the heightened risk. For example, FDA determined in 2005 that a drug prod uct called Bextra should be withdrawn from the market because it presented greater safety risks than other drugs for the same indication with comparable efficacy, and the manufacturer withdrew it. See FDA, Alert for Healthcare Professionals, Valdecoxib (marketed as Bextra) (Apr. 7, 2005) <http://www.fda.gov/cder/drug/ InfoSheets/HCP/valdecoxibHCP.pdf>; Memorandum from John K. Jenkins & Paul J. Seligman through Ste ven Galson to NDA files 20-998, 21-156, 21-341, 21-042 at 17 (Apr. 6, 2005) <http://www.fda.gov/cder/drug/ infopage/COX2/NSAIDdecisionMemo.pdf>. Similarly, FDA withdrew its approval of terfenadine when a safer alternative became available because the drug's risks no longer outweighed its benefits in light of the alternative. See 62 Fed. Reg. 1889 (1997); 63 Fed. Reg. 53,444 (1998). On the other hand, FDA decided not to with draw its approval of erythromycin estolate despite the availability of alternatives with lower risks, because the drug's higher risks were "offset by" its greater efficacy in certain circumstances. 47 Fed. Reg. at 22,547-22,548.
3. In the course of weighing health risks and bene fits, FDA considers the overall health consequences of including particular instructions or warnings in a drug's labeling. As explained above, a drug's safety and effec tiveness are not determined in the abstract, divorced from its labeling. See 71 Fed. Reg. at 3934. Rather, FDA requires each new drug application to contain "a discussion of why the benefits exceed the risks under the conditions stated in the labeling." 21 C.F.R. 314.50(d)(5)(viii) (emphasis added); see 21 U.S.C. 355(d); 21 C.F.R. 314.50(c)(2)(ix). When FDA concludes that a drug's benefits outweigh its risks only under certain conditions, the agency requires appropriate labeling to reflect that determination. See, e.g., 21 C.F.R. 314.110(a).
Moreover, labeling must strike a balance between notifying users of potential dangers and not unnecessar ily deterring beneficial uses through overwarning. 71 Fed. Reg. at 3935. "Exaggeration of risk could discour age appropriate use of a beneficial drug," and thereby harm the public health. Ibid. In addition, excessive warnings can cause more meaningful risk information to "lose its significance." 44 Fed. Reg. 37,447 (1979); ac cord 71 Fed. Reg. at 3935; 65 Fed. Reg. 81,083 (2000). "Warnings about dangers with less basis in science or fewer hazards could take attention away from those that present confirmed, higher risks." Brooks v. Howme dica, Inc., 273 F.3d 785, 796 (8th Cir. 2001), cert. denied, 535 U.S. 1056 (2002). Thus, as the dissent in the Ver mont Supreme Court explained, there are "a number of sound reasons why the FDA may prefer to limit warn ings on product labels." Pet. App. 47a (quoting Brooks, 273 F.3d at 796).
B. FDA's Approval Of A Drug Preempts Claims Challeng ing The FDA-Approved Design Or Labeling When FDA Has Been Made Aware Of The Relevant Risk
Respondent's claims are preempted because they challenge labeling that FDA approved after being in formed of the relevant risk.
1. When federal law merely seeks to impose mini mum safety standards, state laws that impose more stringent safety standards ordinarily are not preempted because such standards do not ordinarily frustrate the federal government's objective to ensure minimum lev els of safety. See, e.g., Sprietsma v. Mercury Marine, 537 U.S. 51, 57 n.6, 64-68 (2002); cf. Lohr, 518 U.S. at 501 (emphasizing, in holding that FDA's substantial-equiva lence determination for a grandfathered medical device did not expressly preempt state tort claims, that FDA had not "weighed the competing interests relevant to the particular requirement in question").
Where, however, federal regulation is designed to strike a balance between competing considerations, state laws that strike a different balance are impliedly preempted because they interfere with the federal bal ancing. Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 152 (1989); Chicago & N.W. Transp. Co. v. Kalo Brick & Tile Co., 450 U.S. 311, 321, 326-327, 330 (1981). In Geier, for example, an agency did not merely impose minimum safety standards. 529 U.S. at 874-875. Instead, it determined that, in light of competing consid erations, public safety was best served by affording manufacturers the choice to install a variety of different passive restraint systems in their vehicles. Id. at 881. The Court held that a state suit seeking to impose liabil ity for a manufacturer's decision not to use a particular type of restraint system would stand as an obstacle to the federal agency's decision. Id. at 881-883; see also, e.g., Buckman Co. v. Plaintiffs' Legal Comm., 531 U.S. 341, 348 (2001) (fraud-on-FDA claim preempted because it would interfere with FDA's ability to strike "a some what delicate balance of statutory objectives"); Interna tional Paper Co. v. Ouellette, 479 U.S. 481, 494 (1987) (state nuisance law preempted because it would "upset the balance of public and private interests so carefully addressed by" the federal permitting regime for water pollution).
2. So too here, the jury's imposition of liability based on petitioner's use of FDA-approved labeling would in terfere with FDA's expert weighing of risks and bene fits. As discussed above, FDA approves labeling for a new drug based on its determination that the labeling strikes the appropriate balance between health risks and benefits. See pp. 16-17, supra. Overwarning can both deter beneficial uses of a drug and "limit physician appreciation of potentially far more significant" risks. 71 Fed. Reg. at 3935; see 65 Fed. Reg. at 81,083. Ac cordingly, FDA's approval of a new drug under the FDCA and its implementing regulations "establish[es] both a 'floor' and a 'ceiling'" with respect to drug label ing. 71 Fed. Reg. at 3935.
That conclusion is confirmed by the fact that liability under state law turns on whether a drug, as labeled, is "unreasonably dangerous." Pet. App. 35a (Rieber, C.J., dissenting) (quoting Town of Bridport v. Sterling Clark Lurton Corp., 693 A.2d 701, 704 (Vt. 1997)); J.A. 15 (re spondent's complaint); J.A. 219 (jury instructions). Any such finding would directly conflict with FDA's determi nation that the drug is safe and effective under the con ditions prescribed, recommended, or suggested in the labeling. Pet. App. 35a-36a. Indeed, respondent specifi cally urged the jury to second-guess FDA and reject the agency's expert judgment. See J.A. 82, 85, 98, 211, 212, 249.
If a state regulatory agency directed manufacturers not to use FDA-approved labeling, but instead to pro vide different or additional warnings, the conflict with federal law would be manifest. As in Riegel, the conflict is exacerbated, rather than ameliorated, by the fact that juries would make those determinations in individual tort suits. As the Riegel Court explained, a jury tends to focus on the risk of a particular design or labeling that arguably contributed to a particular plaintiff's in jury, not on the overall benefits of that design or label ing; "the patients who reaped those benefits are not rep resented in court." 128 S. Ct. at 1008. In contrast, FDA's drug-approval determinations consider the inter ests of all potential users of a drug, including "those who would suffer without new medical [products]" if juries in all 50 States were free to second-guess FDA's expert determinations. Id. at 1009; see Pet. App. 48a (Reiber, C.J., dissenting).
Thus, just as "[s]tate tort law that requires a manu facturer's [Class III medical devices] to be safer, but hence less effective, than the model FDA has approved disrupts the federal scheme," Riegel, 128 S. Ct. at 1008, state tort law that requires a manufacturer to use differ ent labeling than that approved by FDA would disrupt the federal balance.
Here, there is no question that FDA was presented with extensive information about the dangers of acciden tal intra-arterial injection from intravenous administra tion of the drug. Indeed, the agency approved labeling that explained how to inject the drug intravenously so as "to avoid * * * inadvertent intra-arterial injection," and thereby ensured that the drug was safe and effec tive under the stated conditions of use. Pet. App. 167a; see pp. 4-5, supra. Nor did the Vermont Supreme Court point to any marked change in the number or type of reported cases of accidental intra-arterial injection from intravenous administration establishing that the risk was of a distinct type or substantially greater magnitude than the risks of which FDA had been made aware. Thus, the state supreme court's decision sanctioned what amounts to a frontal assault on FDA's approval of the labeling in question.
C. Federal Law Does Not Permit Manufacturers To Make Unilateral Changes To FDA-Approved Labeling Based On Previously Available Information
In holding that respondent's claims are not pre empted, the Vermont Supreme Court relied in large part on its view that an FDA regulation, 21 C.F.R. 314.70(c), "allow[s] unilateral changes to drug labels whenever the manufacturer believes it will make the product safer." Pet. App. 13a. That is incorrect. Petitioner was not free to disregard FDA's judgment concerning previously known risks.
1. As discussed above, the FDCA requires a manu facturer to receive FDA's approval for a new drug's la beling. 21 U.S.C. 355(a) and (d). Because FDA's ap proval strikes an important balance between, among other things, warning of risks and not overdeterring beneficial uses, manufacturers ordinarily may not mod ify designs or labeling approved by FDA without first obtaining FDA's approval for the change. See 21 C.F.R. 314.70; cf. Riegel, 128 S. Ct. at 1005, 1007 (discussing similar requirement for Class III devices). Here, for example, FDA instructed petitioner that the "final printed labeling * * * must be identical" to the ap proved labeling. Pet. App. 165a. Indeed, a unilateral modification of the labeling, absent special circumstan ces, can open a manufacturer to liability for misbranding the drug. See 21 U.S.C. 352(a); 21 U.S.C. 352(f) (Supp. V 2005); 21 C.F.R. 201.100(c)(1) and (d); see also 21 U.S.C. 355(a). If manufacturers were free to make uni lateral changes to labeling the day after FDA's approval based on information that was previously available to the agency, the approval process would be greatly un dermined and the agency's careful balancing of risks and benefits thwarted. The Vermont Supreme Court's view that "FDA approval of a drug label" is nothing more than "a first step," id. at 15a, is therefore a fundamental misconception of the federal regulatory framework.
Under FDA's regulations, a manufacturer ordinarily must submit a supplemental application before making any changes to an approved drug, including changes in labeling. 21 C.F.R. 314.70(b)(2)(v). As a general rule, the manufacturer must obtain prior approval by FDA before making such changes. Section 314.70(c)(6), on which the court below relied, provides only a limited exception to that rule permitting "the holder of an ap proved [new drug] application [to] commence distribu tion of the [changed] drug product involved upon receipt by the agency of a supplement for the change" if, among other requirements, the change "add[s] or streng then[s]" a warning or a statement about administration of the drug in order to promote safety. 21 C.F.R. 314.70(c)(6)(iii)(A) and (C). And even that limited excep tion requires submission of a supplemental new drug application to, albeit not prior approval by, the agency.
As FDA explained when it proposed that regulation in 1982, however, substantive changes may be made without prior FDA approval only "to correct concerns about newly discovered risks from the use of the drug." 47 Fed. Reg. at 46,623 (emphasis added). FDA deter mined that, "[a]lthough most changes in labeling would require the applicant to submit a supplement and obtain FDA approval before making a change," some changes that "would make available important new information about the safe use of a drug product" could be made upon submission of a supplemental application. Id. at 46,635 (emphasis added); see also FDA, Guidance for Industry, Changes to an Approved NDA or ANDA 25 (Nov. 1999) <http://www.fda.gov/cder/guidance/2766fnl. pdf> (explaining that changes may be made without prior FDA approval to add "a precaution arising out of a postmarketing study") (emphasis added).
In a proposed rule issued in January 2008, FDA con firmed that interpretation of Section 314.70(c). 73 Fed. Reg. 2848 (2008). The proposed rule would "reaffirm [FDA's] longstanding position that a supplemental ap plication * * * is appropriate to amend the labeling for an approved product only to reflect newly acquired in formation," and may "add or strengthen a contraindica tion, warning, precaution, or adverse reaction only if there is sufficient evidence of a causal association with the drug, biologic, or device." Ibid. (emphasis added). FDA explained that "[a]llowing sponsors to unilaterally amend the labeling for approved products without limi tation-even if done to add new warnings-would under mine the FDA approval process required by Congress" and "disrupt FDA's careful balancing of how the risks and benefits of the product should be communicated." Id. at 2849. Thus, the supplemental-application process is "primarily designed to provide information to FDA so that the agency," not the manufacturer, "can decide when safety information should be included in the label ing for a product." Ibid.
Even when a manufacturer may make a change prior to FDA's approval under Section 314.70(c), the supple mental application must "give a full explanation of the basis for the change." 21 C.F.R. 314.70(c)(3). The agency may then reject the change based on its own bal ancing of the relevant health risks and benefits. See 21 C.F.R. 314.70(c)(7). If FDA rejects the change, it may order the manufacturer to cease further distribution of the changed product. Ibid. Thus, whether to authorize a change remains "squarely and solely FDA's" decision. 71 Fed. Reg. at 3934. Moreover, products distributed with a unilaterally changed label "remain subject to enforcement action" if FDA finds that the unilateral change rendered "the labeling false or misleading." Ibid.; see 21 U.S.C. 352 (2000 & Supp. V 2005). In prac tice, therefore, manufacturers typically consult with FDA before making any labeling changes. See 73 Fed. Reg. at 2849; 71 Fed. Reg. at 3934.
As the dissent in the Vermont Supreme Court cor rectly explained, Section 314.70(c) does not "allow manu facturers to simply reassess and draw different conclu sions regarding the same risks and benefits already bal anced by the FDA." Pet. App. 40a. Instead, any changes to a drug's labeling without prior FDA approval must be the subject of a supplemental application, which FDA can approve or reject, and must be based on new information establishing that risks arising from use of the drug are of a different type or greater severity than the risks of which FDA had previously been made aware-not cumulative new information that does not add to the information that was previously available to the agency. FDA's interpretation of its own regulation is entitled to significant deference. See, e.g., Auer v. Robbins, 519 U.S. 452, 461 (1997).
For that reason, a state law premising liability on petitioner's failure to depart from the FDA-approved labeling concerning intravenous injection of the drug would not only "stand as an obstacle to the accomplish ment and execution of the full purposes and objectives of Congress," Hines, 312 U.S. at 67, it would also "make it 'impossible' for private parties to comply with both state and federal law," Geier, 529 U.S. at 873. See 73 Fed. Reg. at 2853. Under respondent's theory, peti tioner would have had to change the FDA-approved la beling, but petitioner could not have done so without prior FDA approval.
2. The parties have disputed whether FDA ex pressly rejected the precise warning that respondent asserts should have been included in the labeling. See, e.g., Br. in Opp. 15-17. That factual dispute is irrelevant, because it is FDA's approval decision-which, in this case, followed the disclosure of the relevant risk-not the specifics of the agency's deliberations or speculation about whether the agency might have modified the label in some manner not expressly proposed at the time of the approval, that gives rise to preemption. Any state- law liability would be premised on a re-weighing of the same risks and benefits that FDA already considered in deciding to approve the precise labeling at issue, and federal law would preclude the applicant from making the state-mandated change without prior FDA approval. The agency could not reasonably be expected to ex pressly reject every possible variant of approved label ing as part of its decisional process. Indeed, it would underestimate the post hoc imagination of lawyers to think such an exhaustion of potential variants by the manufacturer or the agency is even possible. More to the point, such express rejection of a precise proposal is not necessary for preemption.
Moreover, any inquiry into the specifics of FDA's decisionmaking process would pose serious practical concerns. Here, FDA expressly rejected one particular labeling change proposed by petitioner, see Pet. App. 17a-18a, but it appears that FDA viewed the change as non-substantive and rejected it for formatting reasons, cf. id. at 18a. With the passage of time, however, it would be increasingly difficult to reconstruct the agen cy's decisionmaking process. If preemption turned on the details of the agency's deliberations, preemption analysis would devolve into an intrusive, and potentially inconclusive, second-guessing of the agency's decisional process. Such an intrusion could also impose unreason able discovery demands on the agency to explain the details of its deliberative process. Cf. U.S. Br. at 21-23, Warner-Lambert v. Kent, 128 S. Ct. 1168 (2008); U.S. Br. at 28-30, Buckman, supra.
3. As in Riegel, this Court need not rely on defer ence to FDA's views to conclude that respondent's claims challenging FDA's weighing of health risks and benefits are preempted. See Riegel, 128 S. Ct. at 1009. But FDA's interpretation of its own regulations is enti tled to significant deference. See Auer, 519 U.S. at 461. And FDA's further views on preemption are also enti tled to some weight under Skidmore v. Swift & Co., 323 U.S. 134 (1944). Cf. Riegel, 128 S. Ct. at 1009. FDA's role in administering its own drug approval process makes it "uniquely qualified to determine whether a particular form of state law 'stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress.'" Lohr, 518 U.S. at 495-496 (quoting Hines, 312 U.S. at 67); see Geier, 529 U.S. at 883 (explaining that an agency has a "thorough under standing of its own regulation and its objectives and is 'uniquely qualified' to comprehend the likely impact of state requirements") (quoting Lohr, 518 U.S. at 496).
In the preamble to a January 2006 rule concerning the labeling of drugs, FDA explained that the govern ment's "long standing view" is that "FDA approval of labeling under the [FDCA] * * * preempts conflicting or contrary State law," especially considering that "FDA interprets the [FDCA] to establish both a 'floor' and a 'ceiling'" for labeling. 71 Fed. Reg. at 3934, 3935. The agency also "recognize[d] that FDA's regulation of drug labeling will not preempt all State law actions." Id. at 3936. FDA then provided some specific examples of circumstances in which state laws are preempted, though it did not attempt to exhaust such circumstances. See id. at 3935-3936 (noting that "at least" those exam ples would be preempted). In this brief, the government has articulated a more generally applicable rule of deci sion, consistent with those examples, based on the pre amble's explanation that (i) the labeling requirements are not a mere minimum safety standard, but rather strike a balance between risks and benefits, and (ii) FDA's regulations permit changes in labeling without prior approval only in narrow circumstances. See id. at 3934-3935; see also 73 Fed. Reg. at 2853; Testimony of Deputy FDA Commissioner Randall Lutter Before The House Comm. on Oversight and Government Reform 1- 2 (May 14, 2008) <http://oversight.house.gov/documents /20080514142253.pdf>.2
Respondent suggests (Supp. Pet. Stage Br. 2 n.1; Br. in Opp. 19 n.6) that the government took a different po sition in a district court filing in Perry v. Novartis, Civ. No. 05-5350 (E.D. Pa. Sept. 21, 2005). The Perry brief argued, however, that failure-to-warn claims "premised on scientific information known to and considered by FDA as part of the approval process, would * * * be preempted." U.S. Br. at 12, Perry, supra. Respondent quotes (Br. in Opp. 19 n.6) a portion of the brief stating that FDA approval does not preempt all state-law label ing claims, because manufacturers may make labeling changes without prior FDA approval in some circum stances. The brief went on to explain, however, that such changes may "be made to warn of new hazards or cautions." U.S. Br. at 11, Perry, supra (emphasis added). Thus, the Perry brief is one of several amicus filings in which the government addressed FDCA pre emption issues on a more fact-specific basis, without articulating a more general rule of decision.
D. Neither The 1962 Nor The 2007 Amendments To The FDCA Displaced The Operation Of Ordinary Conflict- Preemption Principles
1. The Vermont Supreme Court mistakenly thought that Section 202 of the 1962 amendments to the FDCA precludes the application of ordinary preemption princi ples. See Pet. App. 21a-23a. That provision states as follows:
Nothing in the amendments made by this Act to the Federal Food, Drug, and Cosmetic Act shall be con strued as invalidating any provision of State law * * * unless there is a direct and positive conflict between such amendments and such provision of State law.
76 Stat. 793.
At the outset, it is not clear to what extent Section 202 applies here. It is limited to "the amendments made by" the 1962 legislation. § 202, 76 Stat. 793. While those amendments broadened the scope of FDA's new drug approval process by requiring the agency to consider the efficacy as well as the safety of a drug, see § 102(b), 76 Stat. 781, FDA's new drug approval process predated the amendments, see 21 U.S.C. 355(a) and (d) (1958). Indeed, FDA approved Phenergan before 1962. See Pet. 6; Br. in Opp. 23 n.8.
Assuming arguendo that Section 202 is relevant in this case, that provision means only that the 1962 amendments do not preempt the field of drug regulation; it does not manifest an intent to displace ordinary prin ciples of conflict preemption. 71 Fed. Reg. at 3935 n.8. To the contrary, Section 202 expressly contemplates preemption in circumstances involving "a direct and pos itive conflict." 76 Stat. 793. The Vermont Supreme Court read that phrase to refer only to situations in which it would be impossible to comply with both federal and state law, as distinguished from situations in which state law would frustrate the purpose of the federal scheme. Pet. App. 21a-23a. Any such distinction is ir relevant here because, as discussed above, the impossi bility standard is satisfied in this case. See pp. 21-25, supra.
In any event, at the time of the 1962 amendments, the phrase "direct and positive conflict" had long been understood to refer to conflict preemption generally, not to a mere subset of such preemption. See, e.g., United Constr. Workers v. Laburnum Constr. Corp., 347 U.S. 656, 663 n.5 (1954); Sinnot v. Davenport, 63 U.S. (22 How.) 227, 243 (1859). This Court has long contrasted "direct and positive" conflict preemption with "field" preemption, rather than using it as a byword for the impossibility variant of conflict preemption. E.g., Kelly v. Washington ex rel. Foss Co., 302 U.S. 1, 9-10 (1937). Indeed, the Court found a "direct and positive conflict" in circumstances analogous to this one, where federal law imposed some conditions and state law purported to impose an additional one that would frustrate Con gress's objective. Sinnot, 63 U.S. (22 How.) at 241-243. More generally, this Court has never "driven a legal wedge-only a terminological one-between 'conflicts' that prevent or frustrate the accomplishment of a fed eral objective and 'conflicts' that make it 'impossible' for private parties to comply with both state and federal law." Geier, 529 U.S. at 873. A sponsor of Section 202 confirmed that the phrase "direct and positive conflict" takes its ordinary meaning by explaining that the amendment "would merely say that this Food and Drug Act shall not be construed as the intent of Congress to abolish all State laws on the same subject where they are not in conflict with the Federal law." 108 Cong. Rec. 21,083 (1962) (emphases added).
Not surprisingly then, "[t]he Court has * * * re fused to read general 'saving' provisions to tolerate ac tual conflict both in cases involving impossibility and in 'frustration-of-purpose' cases." Geier, 529 U.S. at 873- 874 (citation omitted). Especially considering that the Constitution itself, via the Supremacy Clause (U.S. Const. Art. VI, Cl. 2), subordinates state law to federal law, the courts should not lightly assume that federal law is so self-negating as to authorize state law to frus trate its objectives. Thus, even when a statute contained a savings clause providing that "[c]ompliance with" a federal safety standard "does not exempt any person from any liability under common law," 15 U.S.C. 1397(k) (1988), this Court held that the clause did not preclude the application of ordinary conflict preemption princi ples, including frustration-of-purpose principles. Geier, 529 U.S. at 868, 873-874. The savings clause here, which expressly provides for conflict preemption, likewise does not displace ordinary conflict preemption principles.
Nor is it material that the FDCA lacks an express preemption provision for drugs like the one at issue in Riegel. When Congress enacted a premarket approval process for Class III medical devices in 1976, it ex pressly preempted state requirements that are "differ ent from, or in addition to," certain federal requirements (21 U.S.C. 360k(a); see Riegel, 128 S. Ct. at 1003)-i.e., state-law provisions that conflict most directly with the federal regime. The enactment of that provision in 1976 does not suggest that FDA's new drug approval process has any less preemptive effect, because the device amendments were enacted years later by a subsequent Congress. See Gomez-Perez v. Potter, No. 06-1321, slip op. 10 (May 27, 2008); California Div. of Labor Stan dards Enforcement v. Dillingham Constr., N.A., Inc., 519 U.S. 316, 331 n.8 (1997). Rather, the presence of an express preemption provision for devices but not drugs is explained by the fact that Congress legislated in 1976 against the backdrop of a then-existing state premarket approval requirement for devices, whereas States do not appear to have had similar requirements for drugs in 1938, when the FDCA was enacted. See Riegel, 128 S. Ct. at 1017-1018 (Ginsburg, J., dissenting); see also H.R. Rep. No. 853, 94th Cong., 2d Sess. 45 (1976). Moreover, this Court has looked to conflict preemption principles in determining whether a federal requirement applica ble to a device is different from, or in addition to, a state requirement, and thus expressly preempted by Section 360k(a). See Riegel, 128 S. Ct. at 1008; Lohr, 518 U.S. at 500; id. at 508 (Breyer, J., concurring in part and in the judgment).
2. Respondent has suggested (Supp. Pet. Stage Br. 8-9) that recent amendments to the FDCA bear on the question presented. But those amendments do not re flect any intent to limit the FDCA's preemptive effect. In 2007, Congress enacted Section 901(a) of the FDAAA, Pub. L. No. 110-85, 121 Stat. 922, to enhance FDA's au thority to require applicants to undertake postmarket ing actions, including additional clinical studies, clinical trials, and safety labeling changes. That provision speci fies that it "shall not be construed to affect the responsi bility of the responsible person or the holder of the ap proved application * * * to maintain its label in accor dance with existing requirements, including subpart B of part 201 and sections 314.70 and 601.12 of title 21, Code of Federal Regulations (or any successor regula tions)." 121 Stat. 925-926 (to be codified at 21 U.S.C. 355(o)(4)(I) (Supp. I 2007)). That simply means that the relevant amendments do not affect obligations under other federal laws. It does not manifest any intent to depart from the application of ordinary principles gov erning the preemption of conflicting state laws.
Respondent selectively quotes floor statements of some individual legislators suggesting that, in their view, the FDCA does not preempt state-law claims. Supp. Pet. Stage Br. 8-9. At least as many other legisla tors, however, opined that the FDCA would continue to have broad preemptive effect. See, e.g., 153 Cong. Rec. S11,940 (daily ed. Sept. 21, 2007) (Sen. Gregg); id. at S11,839-S11,840 (Sen. Coburn); id. at S11,939 (Sen. Enzi); id. at S12,050 (daily ed. Sept. 25, 2007) (Sen. Al exander). And as noted, the text of the rule of construc tion that Congress actually enacted, which is limited to the effect of Section 901, itself preserves complemen tary federal requirements without evincing any intent to protect conflicting state laws.
The judgment of the Supreme Court of Vermont should be reversed and the case remanded for further proceedings.
PAUL D. CLEMENT
GREGORY G. KATSAS
Acting Assistant Attorney
EDWIN S. KNEEDLER
Deputy Solicitor General
Assistant to the Solicitor
DOUGLAS N. LETTER
PETER R. MAIER
THOMAS R. BARKER
Acting General Counsel
GERALD F. MASOUDI
Associate General Counsel
WENDY S. VICENTE
Department of Health and
1 The FDCA itself does not directly address changes to a drug's labeling after a new drug application has been approved. That subject is instead left to FDA to address by regulation. Compare 21 U.S.C. 360e(d)(6) (addressing certain changes to an approved medical device).
2 Respondent's reliance (Br. in Opp. 8, 28) on snippets from various earlier Federal Register notices is misplaced because those notices did not squarely address the preemption question here. See 65 Fed. Reg. at 81,103 (stating that proposed changes to existing labeling rules would not have federalism implications); 63 Fed. Reg. at 66,384 (response to comments concerning Medication Guides for "a small number of pro ducts," id. at 66,379); 44 Fed. Reg. at 37,437 (responding to comment that FDA should use different administrative procedures).