This post is from the non-Reed Smith side of the blog, only.

Last week the Third Circuit overturned the dismissal of thousands of cases in In re: Fosamax Products Liability Litigation, — F.3d –, 2017 WL 1075047 (3rd Cir. Mar. 22, 2017), finding that the trial court had improperly granted summary judgement on preemption grounds by misapplying the Supreme Court’s “clear evidence” standard announced in Wyeth v. Levine. It should come as no surprise that we firmly believe the trial court got it right and have lauded its decision in the past. So, it should be equally unsurprising that the Third Circuit’s abysmal “clear evidence” preemption decision is a front-runner for our list of the ten worst cases for 2017.

Regrettably, since Wyeth v. Levine, successful applications of preemption against warning claims asserted in litigation involving branded drugs have been few and far between. That is due in large measure to the both dauntingly high and utterly ill-defined preemption standard set by the Court in Levine, which requires an intensive fact-based analysis and a powerful regulatory case. So when the standard has been met, it is usually with facts that can hardly be described as equivocal. The district court in In re Fosamax, found as much and therefore a somewhat detailed recitation of the regulatory history is warranted.

Fosamax was approved by the FDA in 1995 for the treatment of osteoporosis in postmenopausal women. In re Fosamax, 2017 WL 1075047 at *4. Over the next 15 years, scientific studies were published concerning a possible link between long-term Fosamax use and atypical femur fractures (“AFF”). Merck kept the FDA aware of these studies and in June 2008, FDA asked Merck and other manufacturers to submit whatever data they had on the issue. Id. at *4-5. In September 2008, Merck submitted a “Prior Approval Supplement” (“PAS”) – the same thing as the “Changes Being Effected” (“CBE”) process discussed in Levine, only involving prior FDA review − seeking FDA approval to add information about AFF to both the Warnings & Precautions and Adverse Reactions sections of the Fosamax label. Because it is important to the arguments and the court’s conclusion, we quote some specific language from the requested change and the FDA’s response. Merck’s proposed language included:

Low-energy fractures of the subtrochanteric and proximal femoral shaft have been reported in a small number of bisphosphonate-treated patients. Some were stress fractures (also known as insufficiency fractures) occurring in the absence of trauma. . . . Patients with suspected stress fractures should be evaluated, including evaluation for known causes and risk factors . . . and receive appropriate orthopedic care. Interruption of bisphosphonate therapy in patients with stress fractures should be considered, pending evaluation of the patient, based on individual benefit/risk assessment.

Id. at *5. In April 2009, the FDA responded that Merck could add the information to the Adverse Reactions section of the label, but as to the Warnings/Precautions section, “more time [was] need[ed] for FDA to formulate a formal opinion on the issue of a precaution around these data.” Id. at *5. In its formal response the FDA wrote:

Identification of “stress fractures” may not be clearly related to the atypical subtrochanteric fractures that have been reported in the literature. Discussion of the risk factors for stress fractures is not warranted and is not adequately supported by the available literature and post-marketing adverse event reporting.

Id. at *6.

Almost a year later, the FDA issued a public statement that the data to date “had not shown a clear connection” between Fosamax and a risk of AFF. The FDA commissioned a task force, which after another six months (September 2010) reported a “relationship” between AFF and the class of bisphosphonate drugs, but not anything “proven to be causal.” Id. The FDA responded to the task force’s report with a statement also indicating that causation “is not clear,” and that the Agency was “considering label revisions.” Id. After another month, the FDA announced revised labeling for AFF for this class of drug. Causation was described as “not clear.” Id. at *7. The FDA’s proposed label change stated “[c]ausality has not been established.” Id. In response, Merck proposed adding additional language advising doctors to rule out stress fractures. The FDA ultimately rejected the addition of the term “stress fractures,” finding that the term “represents a minor fracture and this would contradict the seriousness of the atypical femur fractures associated with bisphosphonate use.” Id. The FDA’s approved language, including lack of causality, was added to the Fosamax label in January 2011. Id.

Against this regulatory background, the district court found that the claims of all MDL plaintiffs with injury dates before September 14, 2010 – when the FDA last reaffirmed its position that it did not have enough evidence of a causal association to allow a warning change – were preempted. Id. at *8. The district court ruled that the regulatory history “provided clear evidence that the FDA would have rejected a stronger Precautions warnings because the FDA did reject a stronger Precautions warning.” Id. at *9.

The Third Circuit clearly saw things differently. On appeal, it considered two questions: What is clear evidence? And who should determine whether clear evidence exists? Id.

As to the first question, and perhaps most significantly, this is the first case formally to make the leap from Levine’s amorphous “clear evidence” requirement for implied preemption of warning claims involving branded drugs, to a “clear and convincing” evidence standard for proof of preemption, which required a “highly probable” result. Our problems with Levine are many, and among them has always been this ill-defined standard. Coupled with the Court’s describing impossibility preemption as a “demanding defense,” district courts knew that the standard was high, but just how high was left completely open-ended and has resulted in a lineage of case law that is highly fact-specific with no clearly established definitions or guidelines.

Despite 8 years of court rulings all over the country using the facts of Levine as a benchmark against which to assess the facts of a given case (if stronger facts than Levine then preemption might be possible, if less than Levine, no preemption), the Third Circuit retroactively undertook to read the Supreme Court’s mind, deciding that what Levine really meant when it used the term “clear evidence” was to announce a standard of proof. Id. at *11. The court parsed the ruling in Levine – that absent clear evidence that the FDA would not have approved the label change, it cannot be ruled that it is impossible for a manufacturer to comply with both federal and state law – into three parts. First, it established the legal rule — a manufacturer is “absolved of state-law liability” where it is impossible to comply with both federal and state law simultaneously (impossibility preemption). Second, it established the factual showing needed to satisfy the legal rule – the FDA would not have approved the label change. Third, it established the standard of proof specifying “how convincing the factual showing must be” – by clear evidence. Id. So, according to the Third Circuit:

The term “clear evidence” therefore does not refer directly to the type of facts that a manufacturer must show, or to the circumstances in which preemption will be appropriate. Rather, it specifies how difficult it will be for the manufacturer to convince the factfinder that the FDA would have rejected a proposed label change. The manufacturer must prove that the FDA would have rejected a warning not simply by a preponderance of the evidence, as in most civil cases, but by “clear evidence.”


The principal problem with this aspect of Fosamax is that it flies in the face of other, binding United States Supreme Court precedent. Because the Third Circuit reached its conclusion without even mentioning the Supreme Court precedent that explicitly rejects any “special burden” in implied preemption cases, one wonders if the Third Circuit went off on its own adventure, beyond the scope of the parties’ briefing. The controlling case is Geier v. American Honda Motor Co., 529 U.S. 861, 872-873 (2000), which held that imposition of any “special burden” on the implied preemption defense both lacks “basis . . . in this Court’s precedents” and creates “practical difficulty by further complicating well-established pre-emption principles.” Geier not only rejected imposition of purported “special burdens” in the implied preemption context, but did so on a stronger set of facts for creating such a burden, where Congress had expressly spoken to preemption with both statutory preemption and savings clauses. In the FDCA, by contrast, Congress was merely silent.

Fosamax thus did just what Geier held was not the law. The Third Circuit determined that when the Court said “clear evidence” it really meant “clear and convincing evidence” as a heightened standard of proof requiring a defendant asserting a preemption defense to prove that it is “highly probable” or “reasonably certain” that the FDA would have rejected a label change. Id. at *12.

The court then moved on to the question of whether deciding what the FDA would have done was a question of law or fact. Disregarding uniform precedent since Levine that has treated the issue as a question of law, the Third Circuit determined it was a question of fact. The court found that none of those other cases explicitly addressed the issue and so could not be the basis for concluding that the “clear evidence” test is a legal question. Id. at *13. But if it sounds like a duck, and walks like a duck . . . . In so ruling, Fosamax violated 3d Cir. I.O.P. 9.1 on a truly epic scale. IOP 9.1 provides:

It is the tradition of this court that the holding of a panel in a precedential opinion is binding on subsequent panels. Thus no subsequent panel overrules the holding in a precedential opinion of a previous panel.

Fosamax conceded that its finding that preemption was a matter of fact for jury consideration was contrary to a number of prior Third Circuit statements, which the panel dismissed as “off-hand.” 2017 WL 1075047 at *12. That is not a legitimate characterization. Relegated to an “e.g.” citation in footnote 106 was the standard of review that controlled in In re Federal-Mogul Global Inc., 684 F.3d 355, 364 n.16 (3d Cir. 2012): “The scope of preemption presents a pure question of law, which we review de novo”; as well as the prior FDCA preemption case, Horn v. Thoratec Corp., 376 F.3d 163, 166 (3d Cir. 2004), which likewise held that preemption was a matter for “plenary review” because “it is a question of law.” But Fosamax ignored a lot more precedent than that, including the en banc decision in Orson, Inc. v. Miramax Film Corp., 189 F.3d 377 (3d Cir. 1999), which held specifically that, because it was not bound by “prior decisions, the court was “free to consider the preemption issue as a matter of law.” Id. at 380. Preemption being a matter of law was plainly the holding of the court in Taj Mahal Travel, Inc. v. Delta Airlines, Inc.:

[Plaintiff] argues that [defendant] waived the preemption defense by not raising it until it moved to dismiss [a different] claim, which occurred [later]. . . . However, the preemption defense is a pure issue of law . . . and could be dispositive. Since the parties have briefed and argued preemption on appeal, we will consider it.

164 F.3d 186, 190 (3d Cir. 1998). See also: South Jersey Sanitation Co. v. Applied Underwriters Captive Risk Assurance Co., 840 F.3d 138, 143 (3d Cir. 2016) (“preemption determinations are questions of law”); Roth v. Norfalco LLC, 651 F.3d 367, 374 (3d Cir. 2011) (“preemption … determinations [a]re based on questions of law”); Elassaad v. Independence Air, Inc., 613 F.3d 119, 124 (3d Cir. 2010) (“[w]e also exercise de novo review of a preemption determination, as it is a question of law”); Deweese v. National Railroad Passenger Corp., 590 F.3d 239, 244 n.8 (3d Cir. 2009) (“a preemption determination . . . is a question of law”; Travitz v. Northeast Dep’t ILGWU Health & Welfare Fund, 13 F.3d 704, 708 (3d Cir. 1994) (“[t]he issue of preemption is essentially legal”); Pennsylvania Medical Society v. Marconis, 942 F.2d 842, 846 (3d Cir. 1991) (“[t]he question of preemption involves an issue of law”); Ayers v. Philadelphia Housing Auth., 908 F.2d 1184, 1188 (3d Cir. 1990) (“we determine that as a matter of law the federal regulations . . . preempt the application of [state law]”) (citations and footnote omitted); Pokorny v. Ford Motor Co., 902 F.2d 1116, 1119 (3d Cir. 1990) (“the question of whether [plaintiff’s] action is pre-empted . . . involves a pure issue of law”). These are not “off-hand” statements or dicta of any sort. Some of them are even West case headnotes. In deciding, after decades of contrary precedent, to convert preemption into a factual issue, Fosamax blatantly disregarded the “tradition” of the Third Circuit and did what panels are forbidden to do by IOP 9.1. We have said many times before that strange things happen in tort preemption cases, but Fosamax goes beyond “strange.” If precedent means anything, this aspect of Fosamax is a lawless decision.

Having decided that preemption doesn’t necessarily have to be a question of law, the court’s next awful conclusion was that juries are to determine what the FDA might have done with a different warning. The court looked at the specific question to be answered under Levine – would the FDA have rejected the label change plaintiffs argue was required – and determined it was the type of fact question that could be decided by a jury. The court’s reasoning was threefold: (1) the question asks about the probability of a future event; (2) the question requires the decision maker to weigh conflicting evidence and draw inferences; and (3) the question requires predicting FDA’s actions which requires assessments of FDA official’s motives and thought processes. All of which the court says are functions left to juries. 2017 WL 1075047 at *15-16. Which leads to the court’s ultimate conclusion that “[a] state-law failure-to-warn claim will only be preempted if a jury concludes it is highly probable that the FDA would not have approved a label change.” Id. at *18. Thus the decision opens the door to dueling FDA experts speculating about how the FDA would have viewed some hypothetical proposed label change, while discovery from actual FDA employees is mostly prohibited. See United States ex rel. Touhy v. Ragen, 340 U.S. 462, 468 (1951); Giza v. HHS, 628 F.2d 748, 751-52 (1st Cir. 1980).

This kind of jury second-guessing creates the kind of practical problems that drove the preemption decision in Buckman Co. v. Plaintiffs Legal Committee, 531 U.S. 341 (2001) (also not cited in Fosamax). It “will dramatically increase the burdens facing potential applicants-burdens not contemplated by Congress,” id. at 350, by encouraging plaintiffs to nit-pick FDA decisions – such as those involving bisphosphonates that we discussed above – without regard to the FDA’s actual decision-making process. It will expose “would-be applicants” to unpredictable civil liability,” id. for doing what the FDA directed them to do. Looking at how the plaintiffs in Fosamax discounted the FDA’s “don’t say that” decision, it is likewise plain that allowing such claims to survive preemption will “also cause applicants to fear that their [interactions with] the FDA, although deemed appropriate by the Administration, will later be judged insufficient in state court.” Id. at 351. Allowing jury speculation about what the FDA might have done, had it received a different submission from a regulated company is, as a practical matter, no different than allowing the same jury speculation about what the FDA might have done, had it not been “defrauded” in Buckman.

Since the court in Fosamax was no longer addressing a question of law, its role at the summary judgment stage was “therefore limited to determining whether there are genuine issues of material fact that preclude judgment as a matter of law.” Id. at *13. Remember, the court is imposing a heightened standard of proof, so to affirm summary judgment the court concluded it would have to find “that no reasonable juror could conclude that it is anything less than highly probable that the FDA would have rejected Plaintiff’s proposed atypical-fracture warning had Merck proposed it to the FDA in September 2010.” Id. at *19. Sure sounds like a “special burden” on preemption to us.

What does that mean in the context of Fosamax? Despite the FDA’s rejection of the defendant’s proposal to warn about the particular risk, the court, enforcing the clear and convincing standard against the defendant (preemption being an affirmative defense), holds that plaintiffs can get to the jury on their contention that the FDA’s rejection, after 10 months of review and back-and-forth, was merely a semantic quibble that the defendant could have fixed. See id. at *21 (“a reasonable jury could also conclude that the FDA rejected [defendant’s] proposed warning about femoral fractures in 2009 not because it denied the existence of a causal link between Fosamax and fractures, but because [defendant] repeatedly characterized the fractures at issue as “stress fractures.”). And when you read the court’s view of what Levine requires to get summary judgment, it leaves us very concerned that such a standard will ever be met: “Because the [Levine] test requires the factfinder to speculate about hypothetical scenarios using inferences drawn from historical facts, reasonable jurors could reach a broad range of conclusions.” Id. at *22.

White the Third Circuit claims its adoption of a clear and convincing standard of proof is less than proof beyond a reasonable doubt, id. at *12, the piling of a special burden of proof atop the standard for summary judgment certainly sounds very close to indisputable proof. We used to be concerned that courts would only find “clear evidence” in the context of an outright rejection of a label change. But even that wasn’t enough in Fosamax to allow the court to find that jurors might not differ in their interpretation of that rejection. And it is those differences which bring us back to the fact that this should be a question for the court, not a jury. For a defense like preemption that is equally operative for all plaintiffs, there should be some element of consistency – which will be eradicated if the decision is left to multiple juries to decide. Moreover, deciding federal preemption is an analytical process where legal training adds value and where testing credibility, something usually reserved for a jury, is not as material.

There remain many questions about how this decision and the Fosamax MDL will play out and we’ll be watching it all closely. Definitely more commentary to come.