By Philip Goldberg and Christopher Gramling, partners with Shook, Hardy & Bacon L.L.P.

After the Supreme Court issued its 2009 ruling in Wyeth v. Levine,1 there was widespread confusion as to when the federal Food & Drug Administration’s (FDA) approval of a drug’s labeling would preempt state failure-to-warn claims against the drug’s manufacturer. The court had rejected the preemption defense in the case, leading some to prematurely conclude that preemption in these cases was no longer viable, particularly against manufacturers of brand-name drugs. But the Court also stated that preemption could be found in those situations where there was “clear evidence” the FDA would not have approved the labeling change the plaintiff claims was needed to prevent his or her injury. However, it was unclear what “clear evidence” was or when it existed.

The Court filled in these important gaps in 2019 when it issued its ruling in Merck Sharp & Dohme Corp. v. Albrecht. The Court held that such clear evidence exists when either (1) the plaintiff presents no new evidence of a causal association between the drug and injury alleged warranting a labeling change, or (2) the FDA was “fully informed” of the justification for the labeling change and communicated it would not have approved that change.2 In these situations, federal drug law preempts state failure-to-warn claims. In the past few years, federal courts have developed a strong body of case law for when these failure-to-warn claims are preempted.

We reviewed this jurisprudence for an article published in the Rutgers Journal of Law and Public Policy. See Philip Goldberg, Christopher Gramling & Sarah O’Rourke, A Prescription for Pharmaceutical Preemption, 20 Rutgers J.L. & Pub. Pol’y 209 (2023).3 These cases revealed that, although preemption is a demanding defense, it is certainly attainable. Further, the courts did not establish bright-line rules for when preemption exists. Each drug’s story is different, and the trial judges dove into the substance of the cases to determine whether preemption was appropriate based on the facts, medical science, and circumstances of the specific drug at issue.

Here’s a short guide to what we found:

Burden-shifting Framework

Most courts followed the burden-shifting approach the Second Circuit developed in Gibbons v. Bristol-Myers Squibb Co.4 Under this regime, the court first assesses whether the brand-name manufacturer could have used the FDA’s Changes Being Effected (CBE) process to change the labeling on its own volition and avoid the injuries the plaintiff alleges. Under the CBE process, a label change can be made only if there is “newly acquired information” of a “causal association” between the drug at issue and the alleged injury.

If the plaintiff’s allegations do not meet the threshold standard for when a labeling change is permitted under FDA regulations, the manufacturer would not have had federal authority to change the labeling to be consistent with the plaintiff’s theory of the case. In preemption terms, it would be impossible for the manufacturer to adhere to both the federal FDA regulations that prohibit label changes and the state tort ruling requiring it to change the label. That’s the definition of impossibility preemption. This burden of proof is on the plaintiff.

Second, if the plaintiff shows the CBE process was available to the manufacturer, then the burden shifts to the manufacturer to show the FDA was “fully informed” of this new information and there is “clear evidence” the FDA would not have approved the labeling change. These questions are for judges, not juries, to answer.

Newly Acquired Information

To start the inquiry, the plaintiff must show that he or she has “newly acquired information” that existed at the time of the alleged injury such that the manufacturer could have changed the drug’s labeling under the CBE or other authorized process. This information can include new data, as well as new analyses of previously submitted data. It need not be in the form of a final, peer-reviewed publication.

As the Fourth Circuit explained, the trial court must rigorously assess the scientific evidence at the time and determine whether the information is, indeed, new and material.5 For example, if the FDA had the information before it approved the labeling, it would not be “new.” Similarly, studies published after the alleged injuries could not have provided a basis for a labeling change that would have avoided the plaintiffs’ alleged injuries.

Following this approach, the court in McGrath v. Bayer Healthcare Pharms., Inc., found for preemption because the “new” studies were inconclusive on risks and adverse effects of the drug, meaning the manufacturer could not use the CBE process to change the labeling.6 The court noted that “the FDA prefers a more cautious approach” to labeling that does not include warning of risks that are “not well-grounded in scientific evidence.”7

For the same reason, adverse event reports, themselves, do not constitute newly acquired information,8 and neither do conclusions that are “litigation-driven and unsupported by any published research.”9 In short, the information must meet the scientific rigor the FDA would require of the manufacturer in seeking a label change.

Clear Evidence the FDA Would Not Have Approved the Labeling

If the plaintiff meets this burden, for the claims to be preempted, the defendant must show: (1) the FDA was “fully informed” of the justifications for the warnings required by the lawsuit, and (2) the FDA communicated it would not have approved changing the labeling to include the new warning. Thus, these questions look at what actions the manufacturer and FDA needed to take for the district court to conclude the FDA would not have approved the warnings at issue.

The first inquiry is whether the FDA had the information necessary to approve or reject the warnings proposed by the failure-to-warn claims in the litigation.10 As several courts have held, the manufacturer does not need to be the entity that informed the FDA of these risks. A citizen’s petition raising these risks is sufficient.11 The key is that the FDA was fully informed of the concerns at issue in the case.

If so, the court looks at whether or how the FDA communicated that it would have rejected the labeling change. Plaintiffs have argued that only a final agency action rejecting the specific wording at issue should carry this weight. But courts have rejected any such formalistic approach. They have explained that a decision on whether to submit a proposed or revised warning often reflects ongoing discussions between the manufacturer and the FDA12 and companies should not submit CBE amendments “merely to preserve [a] preemption defense.”13 The court needs to look at the communications between the manufacturer and FDA as a whole in determining whether the FDA conveyed it would have rejected the warning.14 There are no magic words or phrases.

Following this approach, numerous courts over the past few years have found for preemption based on a variety of FDA actions, including the release of an FDA risk assessment, rejection of a citizen petition raising the risks, and inaction with respect to requiring the warnings in light of “extensive and ongoing evaluation of the issue.”15 In reaching these rulings, courts performed in-depth assessments of the substance and science.


By following these guideposts, district judges have issued thoughtful and thorough analyses of what the manufacturers and FDA knew, when they knew it, what they did with that knowledge, and, ultimately, what the FDA decided was the proper public health care decision about the drug’s labeling.  There is no doubt that preemption remains a demanding defense, but, as these cases show, it is attainable.


  1. Wyeth v. Levine, 555 U.S. 555, 580 (2009).
  2. Merck Sharp & Dohme Corp. v. Albrecht, 139 S. Ct. 1668, 1672 (2019).
  3. Sarah O’Rourke worked at Shook, Hardy & Bacon L.L.P. at the time of the journal article’s publication but has since left the law firm for a corporate counsel position.
  4. 919 F.3d 699 (2d Cir. 2019).
  5. See Knight v. Boehringer Ingelheim Pharms, Inc., 984 F.3d 329, 340 (4th Cir. 2021).
  6. 393 F. Supp. 3d 161, 167-68 (E.D.N.Y. 2019).
  7. Id. at 169 (internal citation omitted).
  8. Gayle v. Pfizer, Inc., 452 F. Supp. 3d 78, 88 (S.D.N.Y. 2020).
  9. R.S.B. ex rel. Hammar v. Merck & Co., Inc., No. 20-C-1402, 2022 WL 3927868, at *4 (E.D. Wis. Aug. 31, 2022).
  10. In re Fosamax (Alendronate Sodium) Prod. Liab. Litig., 593 F. Supp. 3d 96, 120 (D.N.J. 2022).
  11. See, e.g., Pfaff v. Merck & Co., No. 12-MD-02331 (BMC), 2022 WL 4121406, *7 (E.D.N.Y. Sept. 9, 2022); In re Incretin-Based Therapies Prod. Liab. Litig., 524 F. Supp. 3d 1007 (S.D. Cal. 2021).
  12. See In re Fosamax, 593 F. Supp. 3d at 126.
  13. Id. at 144.
  14. Id.
  15. In re Incretin-Based Therapies Prod. Liab. Litig., 524 F. Supp. 3d at 1033.