We don’t agree with Wyeth v. Levine.  We particularly don’t agree when courts extend Wyeth v. Levine beyond its actual holding to reach questionable results.  That is what recently happened in the Actos MDL, where the district court denied what seems in our view to have been a solid motion for summary judgment asserting that the plaintiffs’ failure to warn claims were preempted.  See In re Actos (Pioglitazone) Prods. Liab. Litig., No. 6:11-md-2299, 2014 U.S. Dist. LEXIS 1749 (W.D. La. Jan. 7, 2014).

Our readers are familiar with Levine, the 2009 Supreme Court opinion holding that FDA approval of prescription drug labeling (including its warnings) does not alone result in implied federal preemption of state law failure-to-warn claims.  555 U.S. 555, 570 (2009).  The opinion does not foreclose implied preemption in all prescription drug claims, but there has to be “clear evidence” (which Levine did not bother to define) of effective FDA regulation with which state-law obligations would conflict.  As the Supreme Court said, “absent clear evidence that the FDA would not have approved a change to [the drug’s] labeling, we will not conclude that it was impossible for [the manufacturer] to comply with both federal and state requirements.”  Id.

After Levine, implied preemption in prescription drug cases has become a tough row to hoe, one that requires “clear evidence that the FDA would not have approved” a different warning.  The Supreme Court did not, mind you, hold that FDA regulation could never preempt state law failure-to-warn claims, nor did the Supreme Court hold that the FDA has to have actually rejected the plaintiffs’ proposed warning for preemption to apply.  As we read it, if the evidence shows that the FDA would have rejected the proposed warning, a state law duty mandating that same warning conflicts with federal law and is preempted.

We are fully aware that cases rarely turn out that way, and this is where we get to the recent Actos MDL order, where the court applied a version of preemption that is decidedly more restrictive than Levine.  The plaintiffs alleged that the defendants failed adequately to warn about bladder cancer. We are not sure exactly what the plaintiffs contended the label ought to have said; the order does not disclose that information, which makes us suspect that the plaintiffs did not disclose that information either.  We do know, however, that carcinogenicity, bladder tumors, and bladder cancer—and information on these topics set forth in the product’s labeling—were topics of active and specific FDA regulation over something like ten years.  2014 U.S. Dist. LEXIS 1749, at **16-21.

The district court was aware of that history, and by all appearances the court took great care in setting forth the undisputed facts.  Our beef is with the court’s application of Levine, starting with the court’s opening observation that “[w]hile the defendants spend a majority of their briefing arguing the specific facts of this case and detailing the numerous discussions and negotiations between [the manufacturer] and the FDA concerning the labeling . . . the parties appear to agree on one important aspect of the preemption issue, that is if the Supreme Court case of Wyeth v. Levine . . . applies, the plaintiffs’ claims are not preempted.”  Id. at *28.

Let’s unpack that statement.  First, we doubt that the defendants agreed that they would lose if Levine applied.  Levine obviously applied because it is Supreme Court authority on implied preemption in prescription drug cases. We think it more likely that the defendants argued that Levine applied and that the claims were preempted because the extensive regulatory history on precisely the topic of bladder cancer warnings demonstrated that the FDA would have rejected a different warning. That is what Levine says, and we expect that is what the defendants said, too.

Second, the district court seems dismissive of the defendants’ arguing “the specific facts of this case,” including the “numerous discussions and negotiations” with the FDA.  Maybe we are discerning a tone that is not really there, but regardless, we cannot fault the defendants for arguing the specific facts of their case.  Without a clear definition of “clear evidence,” what else can defendants do?  Levine makes “discussions and negotiations” with the FDA not only relevant, but also potentially dispositive if those interactions show that the FDA would have rejected the plaintiffs’ proposed warning.  We think defendants were right on the money to argue the specific facts of the case, and we do not agree, as the court later found, that “the defendants appear to misunderstand their federal-law duties as they might relate to preemption.”  Id. at **32-33.

The part with which we take most exception is the district court’s ruling that the plaintiffs’ failure-to-warn claims were not preempted because “the defendants offer no evidence that a stronger warning was proposed by [the manufacturer] and rejected by the FDA.”  Id. at **34-35.  That might have been true as a matter of fact, but Levine does not require that the FDA actually reject a stronger warning.  The standard is whether the FDA would have rejected the plaintiffs’ proposed warning, and the difference is significant.  Levine correctly permits an inference from regulatory history that the FDA would have rejected a different warning if one had been proposed.  By contrast, the Actos MDL court appears to be requiring that a manufacturer actually propose a stronger warning and that the FDA actually reject it.  That is a more restrictive reading of implied preemption than Levine requires or should allow.

The district court’s view of Levine explains why it was not moved by the defendants’ detailed recitation of the product’s history and the “numerous discussions and negotiations” with the FDA.  Based on the court’s holding, it only matters whether the manufacturer requested the stronger warning and the FDA rejected it. Nothing else makes a difference.  We believe in a more nuanced approach to implied preemption, one that is true to Levine. The touchstone is whether a defendant is unable to obey both state and federal law, and a defendant who has a state-law duty to give a warning that the FDA would reject is damned if it does and damned if it doesn’t.  That is conflict preemption.