By Daniel J. Feith, counsel to Sidley Austin LLP and former Deputy Assistant Attorney General, Consumer Protection Branch, U.S. Department of Justice.

On June 1, 2021, a Massachusetts federal district court ruled that federal law preempted claims that GlaxoSmithKline LLC (GSK) failed to adequately warn of risks associated with taking the anti-nausea drug Zofran during pregnancy.  In re: Zofran (Ondansetron) Prods. Liability Litig., 2021 WL 2209871 (D. Mass. June 1, 2021).  This ruling, which brought an end to multi-district litigation over Zofran, demonstrates the ongoing benefit to drug makers of the Supreme Court’s decision in Merck Sharp & Dohme Corp. v. Albrecht, 139 S. Ct. 1668 (2019), that judges, not juries, must decide whether prescription drug failure-to-warn claims are preempted.  In addition, in basing its preemption determination on FDA actions that GSK elicited after the litigation had begun, the ruling provides drug makers with a valuable lesson: post-Albrecht, the best litigation defense may be a smart regulatory offense.


Zofran (ondansetron) is an anti-nausea drug approved in 1991 for the prevention of nausea and vomiting induced by surgery, chemotherapy, or radiation.  For years, it has also been prescribed off-label for preventing nausea and vomiting during pregnancy.  In 2015, plaintiffs began bringing state-law claims against GSK—which manufactured Zofran until March 2015, when it sold the rights to Novartis—alleging that it failed to warn of the risk of birth defects from taking the drug during pregnancy.  Later that year, 425 suits involving such claims were consolidated into a multi-district litigation, In re: Zofran (Ondansetron) Products Liability Litigation, No. 15-md-2657 (D. Mass.).

As a defense, GSK asserted that these claims were preempted because federal law prohibited it from adding the warnings that state law allegedly required.  This defense turned on whether GSK could have unilaterally revised Zofran’s labeling to add the warnings plaintiffs sought.  Under the Federal Food, Drug, and Cosmetic Act (FDCA), manufacturers generally may not change a drug’s labeling without FDA approval.  In 2009, however, the Supreme Court held that state-law failure-to-warn claims were not preempted if a drug manufacturer could unilaterally change labeling as state law required under the “Changes Being Effected” (CBE) regulation, which permits manufacturers to alter risk information in approved labeling based on “newly acquired information.”  See Wyeth v. Levine, 555 U.S. 555, 570-72 (2009) (citing 21 C.F.R. § 314.70(c)(6)(iii)(A)).  At the same time, the Court recognized that preemption would still apply “when there is ‘clear evidence’ that the FDA would not have approved the warning that state law requires.”  Albrecht, 139 S. Ct. at 1676 (quoting Wyeth, 555 U.S. at 571).

In 2018, GSK moved for summary judgment on its preemption defense.  It contended that it had provided the FDA with all relevant information and that the information the plaintiffs relied on was immaterial.  In support of these arguments, GSK relied on the FDA’s inaction in 2011 regarding Zofran’s labeling after receiving supplemental information from GSK regarding Zofran’s pregnancy-related risks, the FDA’s denial of a 2013 citizen petition to add a pregnancy-related warning to Zofran’s labeling, and the FDA’s denial of a 2015 application by Novartis to add pregnancy-related warnings to the labeling.  The plaintiffs responded that GSK had not fully informed the FDA of the basis for pregnancy-related warnings because GSK had allegedly withheld several categories of relevant information.  In February 2019, the district court denied GSK’s motion, holding that whether GSK had fully met its disclosure obligations, and whether the information GSK allegedly withheld was material, were factual questions that must be submitted to the jury.

Three months after the district court’s decision, the Supreme Court decided Albrecht, which elaborated on Wyeth’s “clear evidence” standard in two important respects.  First, it held that judges, rather than juries, must decide whether the FDA would have rejected the proposed warning, even if that determination required resolving “contested brute facts.”  139 S. Ct. at 1680.  Second, the Court explained that to establish clear evidence of the FDA’s disapproval, a manufacturer must “show that it fully informed the FDA of the justifications for the warning required by state law and that the FDA, in turn, informed the drug manufacturer that the FDA would not approve changing the drug’s label to include that warning.”  Id. at 1678.  In dicta, the Court also noted that the FDA’s disapproval must take the form of “agency action carrying the force of law.”  Id. at 1679.

Following Albrecht, GSK renewed its motion for summary judgment.  In addition, GSK and Novartis followed the path mapped in Albrecht to obtain “clear evidence” of preemption.  First, in 2019, GSK filed a citizen petition asking the FDA to review all the information that the plaintiffs claimed GSK had withheld and to determine whether it warranted adding pregnancy-related warnings to Zofran’s labeling.  GSK also attached the various reports, studies, and analyses that the FDA supposedly had not yet seen.  The FDA subsequently met with counsel for both GSK and the plaintiffs, and the plaintiffs submitted additional exhibits, including expert reports from the MDL.  See In re: Zofran, 2021 WL 2209871, at *10-11.  Second, in 2020, Novartis submitted a prior approval supplement (PAS) to the FDA proposing labeling changes.  Those changes included adding statements that “use of ondansetron in pregnancy is not recommended” and that “[i]t is possible that ZOFRAN can cause fetal harm,” and removing a statement that “[a]vailable data do not reliably inform the association of ZOFRAN and adverse fetal outcomes.”  Novartis argued that these changes were justified by new epidemiological studies.  See id. at *12-13.

The FDA rejected both GSK’s and Novartis’s requests.  It denied GSK’s petition on procedural grounds, “without comment on the relevance, if any, of [the submitted] information to ondansetron product labeling.”  Id. at *11-12.  It also rejected Novartis’s proposed warning against use of Zofran during pregnancy on the ground that the available data do not support the warning given inconsistencies in the published findings and limitations in the design of the epidemiological studies.  Novartis then proposed several other changes, including adding a statement that, “based on the available data,” certain adverse fetal outcomes “cannot be ruled out.”  In March 2021, the FDA rejected that proposal as well, instead agreeing only to a new statement that “available data … preclude an assessment of drug-associated risk of adverse fetal outcomes due to important methodological limitations.”  See id. at *13-15.

The District Court’s Decision

On June 1, following the FDA’s decisions, the district court granted GSK’s renewed motion for summary judgment.  At the outset, the court assumed, without deciding, that GSK could have unilaterally changed Zofran’s labeling through the CBE process prior to 2015.  The court therefore did not consider GSK’s arguments that the information on which the plaintiffs relied was not “newly acquired information,” as required by the CBE regulation.  See id. *28.

The court instead focused on whether the FDA was fully informed of the justifications for the proposed warning by the time of its 2021 decision on Novartis’s PAS.  The court found that by the time of that decision, GSK and Novartis not only had provided the relevant information to the FDA but also had made the agency aware of the plaintiffs’ contentions regarding the significance of that information.  Accordingly, the court concluded that “all of the information concerning the safety of Zofran that plaintiffs allege was withheld from the FDA had been provided to it by the time of the 2020 Novartis PAS,” and “the FDA was ‘fully informed’ of the justifications for the warning label that plaintiffs contend was required by state law.”  Id. at *29-30.

Next, the court examined whether the FDA had informed the manufacturer that it would not approve the proposed warning.  The court noted that Novartis had “specifically requested warnings concerning the use of Zofran during pregnancy, based on the possibility of fetal injury,” including warnings that “‘[t]he use of ondansetron in pregnancy is not recommended’” and that “‘[i]t is possible that ZOFRAN can cause fetal harm when used during pregnancy.’”  Id. at *30-31.  But, the court explained, the FDA “rejected that proposal” as unsupported by “available data” and instead approved labeling stating that “‘[a]vailable postmarketing data have not identified a drug-associated risk of miscarriage or adverse maternal outcomes.’”  Id. at *31.  Thus, the court concluded, the FDA had “approved a label that contains language that is directly contrary to the language proposed by plaintiffs.”  Id.  Furthermore, the court held, the FDA’s disapproval in a PAS decision clearly carried the force of law.  See id.

Finally, the court responded to several arguments the plaintiffs made to avoid the preemptive effects of the decision on Novartis’s PAS.  First, the plaintiffs argued that “there is insufficiently clear evidence that the FDA would have rejected a label with enhanced pregnancy warnings related to animal studies” because Novartis did not propose any changes to the portion of the labeling stating that animal studies did not show any fetal risk from use during pregnancy.  The court, however, explained that even though the FDA was not asked to change that portion of the labeling, it was fully informed of and asked to consider the relevant studies cited by the plaintiffs.  See id.

The court also rejected the notion that the FDA had not considered the information simply because Novartis had not called into question the animal-studies portion of the labeling.  That notion, the court stated, “assumes that the FDA was not following the statutory requirement that it consider ‘all’ relevant information in evaluating the PAS.”  Id. at *32 (citing 21 C.F.R. § 201.57(c)(9)(i)(B)).  The court rebuffed the assumption that the FDA “blatantly ignored” its duties, finding it “highly unlikely” that the agency “turned a blind eye to evidence that Zofran causes birth defects.”  Id.  The court also observed that several other courts had found preemption “where the manufacturer had not requested the precise warning sought by the plaintiffs when the FDA had nonetheless made it clear that it would not accept that label change.”  Id.

Second, the court held that it was irrelevant that the FDA’s disapproval was directed to Novartis rather than GSK.  The court explained that, under Supreme Court precedent, the “essential question … is whether a manufacturer would be permitted to add a warning proposed by a plaintiff to a drug’s label.”  Id. at 33.  Preemption therefore “does not depend on whether the defendant manufacturer is the one who asked for the changes, or to which the FDA explicitly communicated its decision.”  Id.

Ultimately, the court concluded that the FDA had been “fully informed of the justifications for the warning proposed by plaintiffs,” and that “there is no doubt that the FDA would not approve the changes to the warning label proposed by plaintiffs.”  Id. at *3.  Given this clear evidence of FDA disapproval, the court held that federal law preempted the plaintiffs’ claims.  On July 1, the plaintiffs appealed the district court’s decision to the First Circuit.


The In re: Zofran decision demonstrates Albrecht’s pivotal impact on prescription drug products liability litigation, which primarily entails failure-to-warn claims.  Before Albrecht, the Zofran MDL was headed to trial, where a jury would have been tasked with determining whether GSK fully complied with its disclosure obligations and whether the information GSK allegedly withheld would have been material to FDA’s labeling decisions—two legally and scientifically complex questions that the jury would have had to decide after likely having had its views colored by the plaintiffs’ evidence of birth defects associated with use of Zofran during pregnancy.  Instead, as a result of Albrecht, the district court decided these issues before trial in a clear, thorough opinion focused only on the evidence relevant to preemption.  The district court’s opinion also demonstrated an understanding, often lacked by juries, of the delicate balance the FDA must strike in determining the appropriate content and format of drug labeling.  Whereas juries see only the costs of omitting a given warning, cf. Riegel v. Medtronic Inc., 552 U.S. 312, 325 (2008), the district court recognized that the FDA “is concerned not only with avoiding insufficient warnings (that is, failing to warn against risks), but also avoiding over-warning (that is, warning against risks that are unduly speculative, hypothetical, or not adequately supported by science),” which might deter beneficial uses of a drug.

In re: Zofran also demonstrates how, post-Albrecht, a strong litigation defense can depend on a smart regulatory offense.  Instead of relying on the existing record of FDA decision making when it renewed its summary judgment motion, GSK, together with Novartis, expanded that record by filing a citizen petition and PAS with the FDA, respectively.  Those actions, together with the FDA’s responses, were critical to the district court’s determination that the FDA had disapproved the warnings proposed by the plaintiffs after having been fully informed of their justifications.  Drug makers potentially could even forestall litigation entirely through similar regulatory engagement strategies.

Furthermore, In re: Zofran suggests a way drug makers can persuade courts to accept the lengthy delays associated with engaging with FDA in the midst of litigation.  Here, FDA took ten months to fully resolve Novartis’s PAS, and fourteen months to resolve GSK’s citizen petition.  Judges may be reluctant to accept such lengthy delays to allow FDA to act.  In re: Zofran, however, suggests that they arguably should as a matter of primary jurisdiction.  Without explicitly saying so, the decision recognized that by reframing the “clear evidence” inquiry as a question of whether FDA actually disapproved a proposed change after having been fully informed of its justifications, Albrecht made preemption turn on matters within FDA’s “special competence.”  United States v. W. Pac. R. Co., 352 U.S. 59, 64 (1956); see 21 C.F.R. § 10.25(b) (recognizing the FDA’s “primary jurisdiction to make the initial determination on issues within its statutory mandate”).  As In re: Zofran demonstrates, labeling decisions satisfy the three factors that traditionally warrant a primary jurisdiction referral: they are at the heart of the task the FDCA assigns FDA, depend on FDA’s expertise, and would materially aid courts in resolving pharmaceutical failure-to-warn claims.  See Arroyo-Melecio v. Puerto Rican Am. Ins. Co., 398 F.3d 56, 73-74 (1st Cir. 2005).  Accordingly, In re: Zofran supports permitting drug makers to present FDA with warnings proposed by plaintiffs in products liability cases even after litigation has begun.

Finally, In re: Zofran adds to a growing body of case law answering, in a manner favoring preemption, questions left open by Albrecht.  In particular, In re: Zofran sheds light on the significance, if any, that should be ascribed to FDA inaction.  In Albrecht, the majority “noted” in dicta that only agency action carrying the force of law can have preemptive effect, Albrecht, 139 S. Ct. at 1679, but Justice Alito observed in a concurrence that FDA’s decision “not to act” on new information is also relevant to preemption given FDA’s duty under 21 U.S.C. § 355(o)(4) to initiate a labeling change if it becomes aware of safety information that should be included in the labeling, id. at 1684 (Alito, J., concurring in the judgment).  In re: Zofran endorses, at least to an extent, Justice Alito’s broader view.  Relying on FDA’s duty to consider “all” information regarding pregnancy-related risks, see 21 C.F.R. § 201.57(c)(9)(i)(B), the district court refused to narrowly construe FDA’s PAS decision as preempting only the specific warnings Novartis proposed, and instead read it broadly to disapprove other potential labeling changes addressed by the information before FDA.  In re: Zofran thus suggests that, at a minimum, FDA inaction in response to new information can influence how broadly a court construes affirmative FDA actions, such as PAS decisions.  See also In re Incretin-Based Therapies, 2021 WL 880316, at *16-17 (applying § 355(o)(4) in a similar manner).