The Saga of Preempting Prescription Drug Design Defect Claims

Not so long ago in a Circuit not so far away, the issue of whether design defect claims against branded prescription drug manufacturers are preempted was joined. Much like the origins of the Jedi or the major end-of-year holidays as we know them, one would expect a clearer published record of how this came to be. There can be a tendency to read back from recent experience and imbue our past selves with more knowledge or foresight that we actually had. For preemption of design defect claims against branded prescription drug manufacturers, we know we have been arguing for it for years and we are not quite sure why it took so long for a Circuit Court to adopt it. As we noted a few weeks ago, Yates v. Ortho-McNeil-Janssen Pharms., Inc., No. 15-3104, 2015 U.S. App. LEXIS 21428 (6th Cir. Dec. 11, 2015), did find preemption, and did it pretty definitively. So definitively that it took our spot as top decision of 2015. Along the way, the court declined to follow a prior decision of the same court, Wimbush v. Wyeth, 619 F.3d 632 (6th Cir. 2010), which itself reversed decisions of the trial court in Longs v. Wyeth, 536 F. Supp. 2d 843 (N.D. Ohio 2008) (granting summary judgment), and Longs v. Wyeth, 621 F. Supp. 2d 504 (N.D. Ohio 2009) (denying motion to alter judgment), each of which included the holding that pre-approval design defect and negligence claims were preempted. It is with the Longs/Wimbush decisions where our story starts, subject to some back story and with a healthy dose of links to past posts.

We first note, however, that it has long been our view, expressed in many posts and elsewhere, that design defect does not make much sense as a theory of liability for a prescription drug. In most cases, what the plaintiff alleges made the drug excessively risky and thus defectively designed cannot possibly be changed without making it a different drug. One of the principles of pharmacology is that changes to the chemical compound will typically affect both the desired and undesired effects in the body–or as the Supreme Court observed in Bartlett, “because of [a drug’s] simple composition, [it] is chemically incapable of being redesigned.” 133 S. Ct. 2466, 2475. Rarely, a true change to the “design” of the active compound can be identified—maybe chop off this ethyl group or change it from a racemic mixture to a stereoisomer—that will plausibly reduce the pertinent risk, while maintaining benefits and avoiding new risks. Even where that kind of proposed design change exists, the change would make the drug a different product, not a better version of the same product, which is what design defect is supposed to be about. There may be some cases where a plaintiff claims that a different balance of a combined drug’s ingredients, or an inactive ingredient, or the delivery system should be changed to reduce the risk without making it a different drug. Even those cases, though, seem better suited to warnings-based claims.

We continue to maintain that prescription drug cases should be about the adequacy of the information provided through professional labeling, which is consistent with the focus of the regulatory scheme. Because all drugs have risks, and individualized physician decisions to prescribe drugs weigh the risks and benefits of various options (including not prescribing any drug) in light of the likely impact of patient counseling, it makes much more sense to impose liability based on inadequate disclosure of the risks than on the presence of the risks. As set out in our comment k survey, a bunch of jurisdictions agree with us, including where they have adopted a statutory scheme that makes warnings an issue in design defect cases. When there is a reason to look at the design itself, negligent design is arguably a better framework for evaluating liability than strict liability design defect, as the former puts the focus on the defendant’s knowledge and conduct whereas the latter puts it on the product itself. With all due respect to mechanical engineers and with some interest in shortening our discursive detour, prescription drugs are not lawnmowers.

So, why did all that matter for preemption? As we have said when questioning the reasoning of a number of decisions, it will typically be appropriate to ask whether an asserted cause of action has been recognized by state law and whether the plaintiff can make out a prima facie case for it before looking at whether the claim is preempted. Even where a plaintiff did assert a strict liability design defect claim against a prescription drug manufacturer, one would have expected these issues to get sorted out on motions for summary judgment rather than on motions to dismiss, at least until the law became fairly well settled. That means that there were a number of reasons why a prescription drug case would not have produced a ruling on whether a design defect claim for a prescription drug is preempted, let alone one that got published. We can add to that the traditional focus on the larger issue of whether warnings claims—like we said, logically and historically, the main theory of liability for prescription drug cases—would be preempted. In drug preemption’s formative period, there was not much litigation against generic drug manufacturers (and innovator liability was practically unheard of). In branded prescription drug cases, manufacturers were hesitant to raise preemption broadly, particularly while Levine was on its way up to the Supreme Court. That started to change with Buckman Co. v. Plaintiffs Legal Committee, 531 U.S. 341 (2001), and the year after Buckman saw the first published decision using Buckman to exclude evidence that information was withheld from FDA. SeeBouchard v. American Home Products Corp., 213 F. Supp.2d 802 (N.D. Ohio 2002) (Katz, J.).

Even in litigations where warnings-based claims are the plaintiffs’ main theory, there will be cases where the prescribing physician gives clear testimony that she knew the relevant risk and a different warning would not have changed the prescription decision, or that she never read the manufacturer’s warnings at all. Rather than simply accede to summary judgment on warnings-based claims—keeping in mind that manufacturing defect claims with prescription drugs are almost always a dead end—some plaintiffs began to proceed solely on a claim that the drug should not have been available for any physician to prescribe it to any patient, arguably making the prescriber’s knowledge and lack of proximate cause for failure to warn irrelevant to the manufacturer’s liability. While this “stop selling” or “never sell” theory has been labeled in various ways (including “fraud on the FDA”) to meet the plaintiff’s particular needs, it seems to us like a re-framing of design defect, albeit without pesky requirements like a feasible alternative design, no liability with adequate (or without inadequate) warnings, or proximate cause.

We saw three such prescription diet drug cases in fairly short order, each of which shifted gears after the collapse of the warnings claim that other plaintiffs had featured. (Why all three involved residents of northern Ohio is something of a mystery, like why Rey has a British accent.) The first one was resolved before comprehensive rulings. The second one was Lance, where the trial court granted summary judgment based on Pennsylvania law not recognizing a cause of action for “you should not have sold it” under strict liability or negligence, without getting to whether such a hypothetical claim would have been preempted. Two appellate courts expanded the state’s negligence law to accommodate the new theory with the latter dancing around the impact of the intervening decisions in Mensing and Bartlett.The third case—this time asserting Ohio claims and having them heard in an Ohio federal court—was called Buchanan v. Wyeth. When the plaintiff died, the case became known as Longs after the representative of the estate. On appeal, the case became Wimbush after the initial representative died and was replaced. Like in Lance, that the plaintiff’s decedent was alleged to have died from the condition that the prescribing physician testified was a known and acceptable risk in prescribing the drug to the plaintiff might have garnered some extra sympathy from the court. The role of sympathies when the plaintiff’s lawyers abandon the bread-and-butter claim (warnings) and instead argue for a novel reading of the state Product Liability Act is harder to predict.

This is basically how it went down in Longs v. Wyeth, 536 F. Supp. 2d 843 (N.D. Ohio 2008), although we are not mimicking the decision’s sequence. After defendant moved for complete summary judgment on lack of proximate cause and partial summary judgment on preemption, the plaintiff withdrew her warnings claim in favor of her previously asserted statutory design defect and common law negligence claim. (We will ignore the punitive damages aspect of the case.) In opposing summary judgment on her remaining claims, plaintiff elected to rest principally on legal arguments rather than offer evidence. At argument, she maintained that the drug never should have been sold based on the evidence of its risks available by the time it was approved and launched. Under the Ohio statutory design defect provision, modeled on comment k, there is no design defect if the manufacturer “provides an adequate warning . . . concerning that unavoidably unsafe aspect.” Based on the evidence in the record, the defendant received summary judgment because the Redux warnings for primary pulmonary hypertension were adequate as a matter of law and plaintiff did not carry her burden of proving that they were inadequate. Id. at 854. That left the common law negligence claims, assuming they were not abrogated by enactment of the Ohio Product Liability Act. The court held that the OPLA (for the time when the plaintiff’s cause of action arose) abrogated “general negligence” claims but left “negligent design” claims. Id. at 855.

But had plaintiff asserted the same negligent design claim that Ohio courts had recognized before the OPLA? (The analysis of abrogation really should have looked at whether common law had recognized the duty that plaintiff sought to impose liability on the defendant for breaching. A good preemption analysis should ask similar question about state law.) Well, “Plaintiff had not clearly stated the nature of her negligence claim.” Id. at 854. She had alleged a bunch of things about the drug and the defendant, but they all flowed into “Defendants breached their duty by . . . putting Redux on the market in 1996” or “they never should have marketed Redux.” Id. And she had offered no evidence supporting any breach: “Plaintiff has not put forth any evidence that Defendants did not take appropriate steps in providing adequate warnings about Redux, that Defendants should have taken Redux off the market sooner, or that Defendants’ alleged failure to take Redux off the market sooner proximately caused Buchanan’s injury and/or death.” Id. at 856. She also offered nothing on the drug’s design—other than it had risks—or how to lessen them—other than to never sell it. So, plaintiff lost summary judgment on the merits as to whatever it was her negligence claim was supposed to assert, meaning she had no claims left. But the court never actually determined that she was asserting a negligent design claim, as Ohio law had recognized it, as opposed to something that would have imposed a previously unrecognized duty on a manufacturer.

This led to a somewhat confused decision on preemption, which distinguished between pre-approval and post-approval claims. We say “led,” but the opinion discussed preemption first, before it had determined that plaintiff had no viable claims. The conflict analysis for preemption of whatever pre-approval claims plaintiff was trying to make was fairly straightforward:

In other words, the FDA is responsible for regulating which drugs are on the market and the warnings such drugs must provide. As such, Plaintiffs strict liability and negligence claims that Redux was an “unreasonably dangerous” drug for which no warning would have been adequate directly conflicts with the FDA’s authority to determine which drugs are sufficiently safe and effective to be, marketed. Although Plaintiff asserts that she alleges only that Defendants should not have marketed Redux, and that she does not argue that the FDA did anything wrong, the court finds that her claim that Redux should never have been placed on the market interferes with the FDA’s objectives. Consequently, all claims relating to pre-FDA approval are preempted by the FDA. In addition, to the extent that Plaintiff alleges fraud-on-the-FDA or that Defendants concealed or misrepresented information to the FDA, these claims are preempted, as well.

Id. at 847 (emphasis added). Perhaps because plaintiffs’ purported post-approval claims were so poorly described, the analysis of why they were not preempted was not too probing:

In In re Diet Drugs Prods. Liab. Litig. (“Mingus v. Wyeth,” or “Mingus”), No. 04-23744, 2006 WL 1071545, at *3 (E.D. Pa. Apr. 21, 2006), the court held that the plaintiff’s claims that “Wyeth designed a defective product and was negligent in not taking Redux off the market sooner tha[n] it did. . . . [i]n essence, . . . that Wyeth failed in its post-approval duties” were not preempted. Furthermore, Defendants have cited nothing in the FDCA, FDA regulations, or case law that suggests that the FDA intended to preclude all design defect claims. Therefore, the court finds that Plaintiffs post-FDA approval design defect claims, under strict liability and negligence, are not preempted.

Id. at 848. Mingus was the first diet drug case that we mentioned above. Without going into the details, it was not a final or published decision and preemption was hardly its focus. The larger issue with finding that some sort of post-approval design claim was not preempted was that plaintiff never articulated such a claim or what evidence would support it distinct from the evidence that would have supported a (preempted) pre-approval design defect claim. This was not like Levine where the drug was marketed for decades before the plaintiff took it. Rather, although plaintiff had no prescription or pharmacy records—just to make things more complicated—she claimed to have used the drug shortly after it hit the market. In other words, this was not a case where the evidence of “you should have stopped selling it” was different than the evidence of “you never should have sold it,” so a distinction of pre-approval and post-approval claims did not make much sense.

A few weeks after summary judgment, plaintiff filed motions to Vacate Order and Judgment and to Alter Judgment Pursuant to Federal Rule of Civil Procedure 59(e). On preemption, she argued that Reigel and Kent somehow undercut the finding of preemption for pre-approval claims. The court waited until Levine was decided and addressed it too:

The instant case is distinguishable from Wyeth as this case does not involve a failure to warn claim, which served as the basis for the Supreme Court’s determination. Furthermore, the court, in holding that a failure to warn claim is not preempted, focused on the fact that the claim arises out of the actions of the manufacturer post-FDA approval. The Court found that after the FDA approves a product for the market, manufacturers maintain a duty to update warning labels because of their superior knowledge of new risks concerning their products. See id. This post-FDA approval duty is distinguishable from a manufacturer’s duty prior to approval by the FDA, a circumstance that the Court does not explicitly address. While Wyeth may stand for the proposition that post-FDA approval claims are preempted, it does not purport to hold that the same is true for pre-FDA approval claims. Thus, Plaintiff fails to demonstrate that this court erred in its holding that Plaintiffs pre-FDA approval claims are preempted.

621 F. Supp. 2d at 509. Having won on all claims—the plaintiff’s other arguments for undoing summary judgment were rejected as well—the defendant had no reason to try to expand the preemption ruling to cover the unsupported post-approval claims.

On appeal, of course, everything was up for grabs again. While the Sixth Circuit affirmed on the OPLA design defect claim and the post-approval negligence claims, it reversed the pre-approval negligence claims with a preemption analysis that netted the case the honor of being our worst decision of 2010. To get there, the court had to get past abrogation and the one-sided record evidence below. The abrogation analysis did not look at whether the negligence theory plaintiff espoused had ever been adopted as part of the pre-OPLA common law, just whether all negligence claims had been abrogated. 619 F.3d at 639. On the merits, the appellate court started with a curious narrowing of what the district court had held. In its summary judgment decision, the district court found that the walkover on evidence properly in the record precluded plaintiff from making out a prima facie case for any negligence claim. 536 F. Supp. 2d at 855-56. This was emphasized in the denial of the motions to vacate: “Plaintiff's failure to produce evidence regarding proximate cause was fatal to any claim of negligence.” 621 F. Supp. 2d at 512-13 (emphasis added). While appellate court agreed that the record evidence below meant that the defendant should have won summary judgment on any asserted post-approval negligence claim, it wrongly limited the district court’s holding to “Buchanan failed to present evidence demonstrating proximate cause between Wyeth's alleged negligence after the FDA approved Redux and Buchanan's injury or death sufficient to survive a motion for summary judgment.” 619 F.3d at 641 (emphasis added). So, instead of evaluating whether plaintiff made out a prima facie for any negligence claim to which preemption might apply, the court jumped to preemption in the abstract. Seeid. at 645-46 (“For these reasons, we hold that the district court erred in granting summary judgment to Wyeth on preemption grounds on Buchanan's pre-approval common law negligence claims. In so holding, however, we do not pass upon whether there may be alternative bases for adjudicating these claims short of trial. Neither the parties nor the district court have asserted any such alternative, so the issue is not before us.”). It did not even evaluate whether Ohio law imposed the duty that could have created the conflict with federal law, as essential part of a proper conflict preemption analysis: “This discussion assumes the viability of such a claim under the Ohio common law. We do not address whether such a claim actually exists; instead, we address only whether FDA approval would preempt such a claim if it does exist.” Id. at 641 n.6.

Given this path, it is not surprising that the preemption analysis went as it did, relying heavily on Levine’s presumption against preemption. Id. at 642-43. Our original post on Wimbush lays out detailed criticisms of the reasoning, which we will not repeat here. The crux on lack of conflict was this:

In this case, as a general proposition, we can discern no physical impossibility between complying with a state law duty to exercise reasonable care in the process leading up to placing a drug on the market and complying with the federal government's process for approving drugs. This is not to say that such a physical impossibility could never exist, for instance if a state duty required that the manufacturer do something that the FDA forbade or vice versa. But such a situation would, we think, be the exception to the rule. Thus, we are not persuaded that it is always impossible to comply with both state law duties and FDA regulations in the process leading up to FDA approval.

Id. at 643 (emphasis added). Of course, the “state law duty to exercise reasonable care” is a vague characterization of the duty at the center of a claim that the manufacturer should not have sold the drug because of its risks—risks that FDA had considered acceptable in light of its benefits. There was also the lingering issue of whether drugs that have been recalled, as Redux was within eighteen months of its approval, get a different set of rules when it comes to preemption.

By contrast to the complicated history that led to the tortured Wimbush decision, Yates presented a much more direct preemption challenge to design defect claims. Yates was part of the Ortho Evra MDL, overseen by the same Judge Katz who issued the Bouchard decision mentioned above. We have posted previously on a number of summary judgment decisions from the litigation centered around the lack of viable warnings claims under any state’s law. Like here and here. These decisions made sense because the risks at issue in these cases had been the subject of detailed warnings for Ortho Evra and other estrogen-based contraceptives for a long time and physicians are well aware of them. Along the way, various other claims fell, including design defect based on lack of affirmative evidence and on preemption. Yates was at the end of the line of these decisions and knocked out all claims in two separate summary judgment decisions. It ended up as first to be decided on appeal and presents a nice study of a complete defense win in a fundamentally questionable case brought over a risk the prescriber understood, discussed with the plaintiff, and shared by the other drugs that were considered for the plaintiff. We will focus on the preemption of the purported New York law design defect claim given the same court’s prior decision in Wimbush.

It would be easy to say that the explanation for the difference between the result in Wimbush and Yates was the intervening decision in Bartlett, particular its oft-cited statement that “Once a drug—whether generic or brand-name—is approved, the manufacturer is prohibited from making any major changes to the ‘qualitative or quantitative formulation of the drug product, including active ingredients, or in the specifications provided in the approved indication.’” 2015 U.S. App. LEXIS 21428, *27 (quoting Bartlett, 133 S. Ct. 2466, 2471 (2013)). Yates considered this to be dicta and found Bartlett “did not reach the sweeping conclusion that all design defect claims regarding a generic drug and clarified that preemption cannot be avoided if the only way a manufacturer can comply with both federal and state law is to exit the market.” Id. Nor did it read Bartlett as limited to generic drugs. Instead, the court applied an impossibility (conflict) preemption analysis based on Levine, Mensing and Bartlett--including Levine’s “clear evidence” standard (not repeated in Mensing or Bartlett) but not any presumption against preemption. The start of the analysis, as we have been saying all along, is to determine what duties are actually imposed by the state law upon which plaintiff seeks to impose liability. New York law follows the same standard for design defect claims, whether labeled as negligent design or strict liability, requiring plaintiffs to prove a feasible alternative design and (consistent with comment k) not imposing liability on manufacturers of prescription drugs that establish the warnings were adequate. Continuing the approach from prior cases, Yates analyzed preemption for pre-approval and post-approval claims separately, starting with the latter.

It was not a close call. “Yates’s post-approval design defect claim is clearly preempted by federal law.” Id. at *32. Unlike most purported design defect cases, including the diet drug cases that first articulated a novel “stop selling” theory, the plaintiff in Yates actually came up with a proposed change to the design of the drug: reduce the estrogen dosage by 20%. She probably had nothing more than speculation to say that this change would have avoided plaintiff’s stroke or would not have reduced the drug’s efficacy, both of which she would have been required to prove. For impossibility preemption, however, the question is whether the manufacturer could have changed the dosage on its own after approval. Based on a straightforward read of the regulations, “We think it clear that changing the dosage level of the active of Ortho Evra constitutes a ‘major change,’ such that prior FDA approval is necessary.” Id. at **33-34. Thus, any post-approval liability based on failure to reduce the estrogen dosage was preempted.

Pre-approval was a little harder, even without Wimbush, because of the facile appeal of the argument that “no federal law [] restricts a brand-name drug manufacturer from designing a reasonably safe product prior to FDA approval.” Id. at *35. But the Yates court looked back at what plaintiff would need to prove to impose liability and found no duty that could be reconciled with federal law.

But Yates's argument regarding defendants’ pre-approval duty is too attenuated. To imagine such a pre-approval duty exists, we would have to speculate that had defendants designed Ortho Evra differently, the FDA would have approved the alternate design. Next, we would have to assume that Yates would have selected this method of birth control. Further yet, we would have to suppose that this alternate design would not have caused Yates to suffer a stroke. This is several steps too far. Even if New York law requires defendants to produce and market a different design, the ultimate availability to Yates is contingent upon whether the FDA would approve the alternate design in the first place.

Id. at **35-36. This walked right into the Mensingindependence principle. “Defendants could not have complied with whatever pre-approval duty might exist without ultimately seeking the FDA’s approval prior to marketing Ortho Evra, and certainly prior to Yates’s use of the drug.” Id. at *37.

Wimbush did not dictate a different result. Whereas Wimbush held it was not always impossible to comply simultaneously with FDA requirement and with “a state law duty to exercise reasonable care in the process leading up to placing a drug on the market”—ignoring that the plaintiff had never articulated what the defendant should have done besides not try to get the drug approved—the vagueness of the claim in Yates worked against her. “However, Yates has not explained precisely what a pre-approval claim would look like in her case. And we are unable to conceive of any coherent pre-approval duty that defendants would have owed to Yates when it was developing Ortho Evra.” Id. at **37-38. Wimbush was also distinguishable because Redux was withdrawn and Ortho Evra is still fairly widely prescribed. To us, we do not see a basis for a general rule for preempting design defect claims for drugs that are marketed at the time the case is pending but not for those that have been withdrawn or discontinued for some reason.

The Yates court also noted that Bartlett’s rejection of the “stop selling” rationale as a way to get around conflict preemption also supports its finding of preemption for pre-approval design defect claims. “In contending that defendants’ pre-approval duty would have resulted in a birth control patch with a different formulation, Yates essentially argues that defendants should never have sold the FDA-approved formulation of Ortho Evra in the first place. We reject this never-start selling rationale for the same reasons the Supreme Court in Bartlett rejected the stop-selling rationale of the First Circuit.” Id. at *39. This is surely correct, but it suggests that Wimbush’s preemption decision should not be “good law” anymore. Id. at *38. Recall, the plaintiff in Longs/Wimbush withdrew her warnings claims to proceed solely on the theory—however labeled—that the drug never should have been sold. Wimbush did not “presume that an actor seeking to satisfy both his federal- and state-law obligations is not required to cease acting altogether in order to avoid liability,” as Bartlett requires. It accepted the opposite premise. Going forward, we hope and expect that the preemption of design defect claims for branded prescription drugs will expand as the conflict of poor reasoning like in Wimbush and Lance cannot be reconciled with the principles outlined in Mensing and Bartlett.

Do we have some tie back to the Star Wars trilogy of trilogies? Do we have some prediction about which circuit will be first to follow Yates? Do we have some final imbedded joke to amuse the readers (or ourselves)? No, no and no. We do, however, have a wish for a Happy New Year to whoever made it all the way through this post.