The Day after Levine: Analyzing the Supreme Court's Recent Ruling that FDA Approval of Label Warnings Does Not Preempt State Tort Law
|By ANTHONY J. SEBOK
|Tuesday, March 17, 2009|
The Supreme Court's decision in Wyeth v. Levine was eagerly awaited, and it sorely disappointed the pharmaceutical industry. By a vote of 6-3, the Court held that a Vermont products liability suit was not preempted by the fact that the FDA had approved the label that a jury in that state had deemed inadequate. (For details of the underlying lawsuit see my prior column on the case.) In this column, I will consider Levine's ruling and ramifications.
Levine's Holding and Facts
The holding of Levine is pretty straightforward. The drug in question, injectable Phenergan, was approved by the FDA for sale under the Food, Drug, and Cosmetic Act ("FDCA") because it was deemed "safe and effective" by the agency in 1955. When a drug is approved under the FDCA, not only the drug itself is approved, but so is the exact language of the label and any material included with the drug's packaging. Over the subsequent years, the label for Phenergan was modified with the FDA's approval. One of the modifications it approved was a warning that stated that extreme caution should be used when directly injecting Phenergan (a method known as "IV push," as opposed to using an IV drip) since Phenergan can cause gangrene when accidentally injected into an artery.
The plaintiff in the case, Diana Levine, who lost an arm due to the inadvertent injection of Phenergan into her artery, was awarded damages because she convinced a jury in Vermont that the warnings relating to the risk of intra-arterial injection of Phenergan were inadequate.
Wyeth, the manufacturer of Phenergan, denied that the warnings were inadequate. The company also took the position that the entire question about the adequacy of the warnings was preempted by the fact that the FDA had approved the warnings in question. The Supreme Court held that Wyeth was wrong.
The Stakes of Levine: Federal Preemption As a Way to Cut Off State Litigation
What is at stake? Preemption has become an important way for parties who distrust tort litigation to cut it off at the very root. If a question of liability is "preempted" by federal law, that means that the only remedy that a victim of an alleged wrong has is under the federal statute. State tort law is cut off. Since federal statutes often omit federal statutory damages remedies, and federal courts are loath to imply rights of action, preemption often operates as an immunity.
Preemption can come in various forms. The easiest to understand is explicit preemption—when Congress says that no state may impose requirements on an activity. If Congress wants to control precisely what is said on cigarette packages, it may explicitly bar the states from requiring language different from that approved by the Federal Trade Commission, as it did with the Cigarette Labeling Acts. The Supreme Court held that in such cases of explicit preemption, a state cannot get into labeling business through the back door by having a jury impose liability when a cigarette package did not have a label that would meet the demands of the state's tort law, even if it met the demands of the federal labeling act.
There is also implicit preemption. Congress may not explicitly take a matter away from the states, but it may do so by indirection. Implicit preemption, it is said, often takes one of two forms—impossibility and frustration. Each came into play in Levine.
The Impossibility and Frustration Arguments For Federal Preemption That Were Raised in Levine
At the beginning of the battle over whether the Vermont tort suit was preempted, Wyeth relied heavily on an implied preemption argument based on impossibility. Wyeth argued that, since the FDCA prohibits the distribution of a drug with labels that not approved by the FDA, Wyeth could not have complied with Vermont tort law as expressed by the jury in the Levine verdict, since it could not have legally altered its labels.
If that sounds a little suspicious, that's because it is. The FDA is not crazy. It allows drug manufacturers to change their labels if new reasons emerge that, in the eyes of reasonable persons, would require a change to protect the public's safety. If Wyeth had a good reason to alter its labels to reflect new warnings that would have protected Levine, it could have made those changes, and then applied to the FDA later for those changes to be approved.
In my opinion, if you had asked Wyeth why it did not change the label on Phenergan before Levine's injury, it would have made a few replies, all of them sincere. The first is that it did not believe that the label approved by the FDA was inadequate. The second was that, although it knew of cases in which Phenergan had been inadvertently injected into arteries since the 1970's, there was no new data or research that told it anything more than what it already knew—that there was a slight risk of injury by gangrene when Phenergan was injected. Third, and finally, it would say that it could not imagine how any additional language, or alternative language, about this risk could have helped patients like Levine. In other words, the only way to eliminate the risk would be to eliminate the injection of Phenergan as a treatment option, which neither Wyeth nor the FDA felt was in the best interest of patients.
Wyeth's real preemption argument was that the FDA knew each of the three points made in the prior paragraph, and that the FDA's decision to approve the label was its implicit approval of Wyeth's position that no further warnings were practicable and that injectable Phenergan was a worthwhile product that should be available to the public. Wyeth's argument was that Vermont, by imposing liability on Wyeth for providing Phenergan with the label approved by the FDA, would frustrate the federal government's goal of allowing Americans to have access to injectable Phenergan.
Justice Stevens's Majority Opinion in Levine, and How It Dealt with Wyeth's Arguments for Preemption
From this perspective, Justice Stevens's majority opinion is Levine seems reasonable. He held, on behalf of the majority, that the evidence was insufficient to show that the FDA would have prohibited Wyeth from changing its label.
Justice Stevens conceded that if the FDA had rejected "the kind of warning required by the Vermont jury," then Wyeth would have faced a classic "impossibility" dilemma, no different than the dilemma it would have confronted if Vermont had passed a statute requiring warning language on Phenergan's label that would have been rejected by the FDA.
But, Justice Stevens pointed out, there the FDA had not rejected "the kind of warning required by the Vermont jury." So the impossibility argument raised by Wyeth had to depend on a counterfactual: What would the FDA have done if Wyeth had proposed to strengthen the warning on the label?
Proving a counterfactual is hard. In fact, there is evidence that supports Wyeth's argument that the language demanded by Levine would have been rejected. In 1988, in response to concerns expressed by the FDA about the risks of intra-arterial injection of Phenergan, Wyeth proposed new language that for the label that described how to decrease the risk of "inadvertently injecting [Phenergan] intra-arterially." The FDA did not respond to this request until 1996, and then, without comment, did not approve this new language.
According to Justice Stevens, the FDA's failure to approve the language submitted by Wyeth in 1988 does not prove that it would have rejected "the kind of warning required by the Vermont jury," or that it would have punished Wyeth for putting such language on its label once Wyeth became convinced that such language was necessary.
Two Possible Approaches: Looking to Actual FDA Decisions Alone, Or Including Background Reasons for FDA Decisions as Well
Stevens's impossibility argument could take one of two forms. In one version, Wyeth would have been able to claim preemption only if it had submitted to the FDA exactly the warning required by the Vermont jury, and was rejected by the FDA. I call this the "actual warning" test. In the second version, if Wyeth had proven that the FDA rejected the reasoning behind the warning required by the Vermont jury, then it could claim preemption. I call this the "background reasons" test.
The actual warning test for impossibility is unrealistic for two reasons. First, it is very hard for a pharmaceutical company to anticipate which warnings will be the subject of litigation in the future. Second, as Justice Stevens himself noted in another portion of his opinion, "the jury verdict did not mandate a particular replacement warning"—it only held that the one approved by the FDA was inadequate. So it turns out that the reason Wyeth never offered the FDA "the kind of warning required by the Vermont jury" isbecause there was no "Vermont warning" before Levine was injured. Wyeth would have, at best, had to guess what Vermont would have deemed safe.
A Third Circuit Case Exemplifies the "Background Reasons" Test in Practice
In other words, the actual warning test for impossibility preemption proposed by Justice Stevens in Levine is virtually impossible to satisfy (pun intended). On the other hand, Justice Stevens may be interpreted to have meant to adopt the background reasons test. To see how the background reasons test might work, consider the case of Colaciccio v. Apotex.
In Colaciccio, Judge Sloviter of the U.S. Court of Appeals for the Third Circuit had written an opinion finding that lawsuits against manufacturers of antidepressants such as Paxil, Zoloft and their generic equivalents were preempted because the label had been approved by the FDA. These drugs are forms of selective serotonin reuptake inhibitors ("SSRIs") and the plaintiffs in these suits alleged that they increased the risk of suicide and that the warnings on the labels of these drugs concerning the increased risk of suicide were insufficient.
Judge Sloviter held that the approval by the FDA of the warnings on the labels of the SSRIs implicitly preempted the suits. Interestingly, she did not say that the basis of the implicit preemption was impossibility. Judge Sloviter, as we shall see below, found the impossibility test a non-starter in the typical post-approval labeling case.
Instead, Judge Sloviter took a more conventional route. She observed that implied preemption can occur when state "requirements"—including state tort law—"frustrate" the ends of federal law without necessarily making compliance with both state and federal law impossible. This familiar idea, which was given its fullest articulation in Justice Breyer's opinion in Geier, is good law and survives Levine.
There is no evidence that the FDA rejected applications by the SSRI pharmaceutical companies to change the labels to reflect the warnings that would be required by state tort law. But, as Judge Sloviter noted, there is a raft of evidence that the FDA "rejected the scientific basis for the warnings" the Colacicco plaintiffs argued were required by state law.
What was this evidence? An internal FDA advisory committee had concluded that no warning was necessary on Paxil, an SSRI similar to those which were the subject of the lawsuit. In the face of organized public concern over the risks of SSRIs, the FDA publicly rejected the arguments made by citizens' groups. In litigation around the country, the FDA filed amicus briefs arguing that in its opinion, the risk of "under-use" of SSRIs "based on dissemination of unsubstantiated warnings may deprive patients of efficacious and possibly lifesaving treatment."
One might say that all this is well and good, but what of Justice Stevens's test: Why not require the pharmaceuticals to find out whether the warnings they might be asked to adopt in the states are truly rejected by the FDA? Judge Sloviter anticipated this argument, and I think her treatment of it deserves to be reproduced in its entirety:
"Plaintiffs argue, however, that the FDA's actions were insufficient to manifest such a rejection here. They ask us to overlook the FDA's various public statements rejecting the existence of an association between SSRIs and adult suicidality because they were not made in the context of the FDA's formal rejection of a [labeling] supplement submitted by one of the defendant pharmaceutical companies.
We agree that a court could more easily determine the preemption issue if the FDA had formally rejected such a [labeling] supplement, but we cannot compel the defendant companies to suggest a [labeling] supplement that they believe is unnecessary. Nor do we favor encouraging regulated parties to submit [labeling] supplements for the sole purpose of insulating themselves from potential liability. . . . Thus, we reject the notion that, in order to rise to the level of a conflict in this situation, the FDA's rejection of a warning must be imbued with the formality proposed by the plaintiffs" (emphasis supplied).
Preemption in Colaciccio was plainly based on the background reasons test, not the actual warning test. The question today is whether the background reasons test can survive Levine. The Supreme Court remanded Colaciccio back to the Third Circuit a few days after Levine, for reconsideration in light of Justice Stevens's opinion. I hope that the Third Circuit distinguishes Colaciccio and Levine on their facts. Although Wyeth may disagree, one must admit that the SSRI pharmaceuticals engaged the FDA more deeply and more decisively than was true in the case of Phenergan.
The plaintiffs' bar may respond that the FDCA empowers the FDA to accept or reject proposed labeling language, not reasoning, and if they are right, then the actual warning test will be the only implied preemption test currently available to pharmaceutical companies in post-approval warning cases. If this is the situation, then the FDA should consider codifying the background reasons test in a regulation.
In his cryptic concurrence in Levine, Justice Breyer emphasized that Justice Stevens's argument against preemption does not touch upon the question of implied preemption resulting from regulation. Geier, after all, which was authored by Justice Breyer, involved a regulation promulgated by a federal agency, not a "mere" approval. How would Justice Breyer treat a preemption argument based on a regulation issued by the FDA asserting that the adoption of background reasons concerning the sufficiency of warnings reflected a federal policy regarding the use of pharmaceuticals? Much would depend on how the regulation was written, and one might argue that, unlike the regulation concerning passive restraints for car passengers which was at issue in Geier, such a regulation would be so vague as to fail some basic test of administrative law.
After Levine, Big Pharma is faced with a choice. If it can convince the Supreme Court to adopt the background reasons test set out by Judge Sloviter in Colaciccio, and Big Pharma can do the kind of work that was done by the SSRI manufacturers to produce a record upon which a court can predict agency "rejection" as I have defined it, then Levine may turn out to have produced a good result. If the background reasons test is rejected, then Big Pharma's best hope is to ask the FDA to issue regulations that incorporate its logic, and to hope that they can convince Justice Breyer to take up their cause once a preemption case under the new regulation is challenged.