Thursday, August 19, 2010

On Second-Guessing the FDA

Remember what Wyeth v. Levine, 129 S. Ct. 1187 (2009), supposedly wasn’t about?  We addressed this issue here shortly after the Levine decision.  One thing that Levine wasn’t about (according to the majority, anyway) was a claim that the defendant failed to contraindicate an FDA approved use of the drug in question (Phenergan, in case you’ve forgotten).  We observed:


In particular, the majority [in Levine] held that through the miracles of post-verdict review, the plaintiff’s trial attacks on the FDA and its approval of the IV-push indication disappeared.  [129 S. Ct. at 1194.]  It no longer mattered that the plaintiff had expressly argued that the drug should never be used for that indication (id.: plaintiff “offered evidence that the IV-push method should be contraindicated and that [the drug] should never be administered intravenously”), since the Vermont Supreme Court later declared that the verdict “established only” that the “warning was insufficient.”  There were thus “any number of ways” – all unstated – for Wyeth to change its warnings.  Id.  In this way, Levine was no longer the “failure to contraindicate” case that we always considered it to be.  Id.  Allowing that kind of a claim couldn’t support a majority, so the plaintiff's claim became something else.
We thought that change was significant, because the failure-to-contraindicate issue had been prominent in both the briefing and at oral argument.  See here, here, and here.  That the Levine majority went out its way to emphasize that it was not addressing a failure to contraindicate claim suggested – and still suggests – that there was no majority on the Court for a holding that preemption would not apply to a claim that essentially posited that an FDA-approved use should never be allowed.

Instead, as everyone following preemption knows, the Court held that a failure to warn claim didn’t present a conflict because manufacturers could revise their warnings immediately to take account of new information without FDA prior approval.  Levine, 129 S. Ct. at 1197-98.

To put in starker terms, the majority in Levine shied away from allowing claims that state law could say “no” to a use of a drug to which the FDA had said “yes.”  Except where FDA regulations (allegedly) by their own terms allowed modifications, there wasn’t a majority on the Court for letting the common law go there.

We’re now a little more than a year into the post-Levine era, and already it looks like the other side (definitely) and the courts (all too often) are ignoring the implicit limits of Levine.  We’re seeing more and more claims that seek to second-guess and overturn FDA decisions in the same way that the erstwhile failure-to-contraindicate claim in Levine would have done.

Most notably, and most recently, is yesterday’s decision by the Sixth Circuit in Wimbush v. Wyeth, slip op. , ___ F.3d ___, 2010 WL 3256029 (6th Cir. Aug. 18, 2010).  That’s the case that reversed Longs v. Wyeth, 621 F. Supp.2d 504 (N.D. Ohio 2009) – the name difference having to do with a change in ethe state administrator who is the nominal plaintiff in the case.

Wimbush/Longs involves a claim that went way beyond a change in warnings – in fact, warnings weren’t even involved.  2010 WL 3256029, at *5 (“plaintiff dismissed her failure-to-warn claims and cannot show that Wyeth is strictly liable for defective design under [state law], her only remaining claims are for negligence”).  The Sixth Circuit cloaked that claim in legalistic jargon, calling it a “pre-approval negligence claim.”  Id. at *7.  However, stripping away the jargon, what the plaintiff asserted was quite clear:  that the defendant was “negligent” for ever having submitted the drug to the FDA in the first place, despite the fact that the FDA went ahead and approved the drug for marketing.  See Id. at *6 (“Defendants breached their duty by . . . putting [the drug] on the market in 1996, when [they] knew [it] was unreasonably dangerous”) (quoting complaint); (“claims regarding [defendant’s] acts and omissions, prior to the FDA’s approval, in bringing the drug to market”) (court’s characterization).

Note that plaintiff did not, and could not, allege that the defendants withheld information from the FDA.  That would be a preempted fraud on the FDA claim under Buckman Co. v. Plaintiffs’ Legal Committee, 531 U.S. 341 (2001).  Thus, the claim is that the defendants were “negligent” in ever marketing the drug based upon the same information that the FDA had when it approved it.  In other words, that’s a direct second-guess of the FDA’s decision to approve the drug for the use that was prescribed for the plaintiff’s decedent.  This aspect of the case – that although the FDA said “yes” to the use in question, state common law could say “no” – is no different than the failure-to-contraindicate claim that the Levine majority refused to touch with a ten-foot pole.

Nonetheless, the court in Wimbush relied on Levine to hold that it was OK (or, at least, not preempted) for the plaintiff to argue that an FDA-approved drug should never have been marketed at all – or, as the court put it, “negligent for bringing [the drug] to market at all.”  2010 WL 3256029, at *7.

The Wimbush court, we think, wrongfooted its argument from the beginning, characterizing this tort claim as being within “a field which the States have traditionally occupied,” and therefore applying a presumption against preemption.  Id. at *8 (citing Levine).  Whatever our view of whether the warning claim in Levine should have been so characterized, the claim in Wimbush – whether the drug should have been “brought to market at all” – does not (at least since the FDA has existed) reside in a field occupied by state law.  Rather, the FDA decides what drugs are marketed.  Even cases denying preemption in other contexts recognize that:

All prescription drugs marketed in this country must first receive FDA approval.  Manufacturers of new drugs must submit a new drug application (NDA) to the FDA that demonstrates the drug’s effectiveness and safety for its intended use.  The 1962 Food, Drug and Cosmetics Act (FDCA) established this avenue . . . with the core objective of ensuring that drugs are both safe and effective. . . .  New drug approval requires, among other deliverables, the results of successful clinical trials and labeling that accurately portrays the benefits and risks of the drug, as indicated by those trials and other data.
Demahy v. Actavis, Inc., 593 F.3d 428, 431-32 (5th Cir. 2010).

So the first (mis)step on the road to allowing common-law plaintiffs to attack the mere presence on the market of an FDA-approved drug is recognition of a presumption against preemption on the ground that it’s “traditional” for states to make yes/no decisions on whether drugs are safe enough to be sold.

Guided by this misapplied presumption, Wimbush held that it was not “impossible” for the defendant to comply with federal and state law.  But see the rabbit go into the hat.  Its analysis ignores the FDA approval altogether, and focuses instead on the “the process leading up to placing a drug on the market.”  2010 WL 3256029, at *9.  Obviously, if FDA approval itself were considered the relevant factor, then there would be impossibility, because the FDA said "yes" to marketing.  But, due largely to the presumption, the court kicks the process back a step and puts state law in the position of supervising a defendant’s submission of a product for FDA approval:

Simply because tort liability “parallel[s] federal safety requirements” does not mean that liability is preempted.  Allowing a state tort claim under the instant circumstances would not “interfere[] with the methods by which the federal statute was designed to reach its goal.”
2010 WL 3256029, at *10.

With all due respect, we find that quite hard to believe.  How does a claim that a defendant was “negligent” in the “process leading up to placing a drug on the market” play out”?  Is it that the defendant should have done another study or analyzed an existing study differently?  But the FDA’s approval means that it did not think any additional studies or analysis was required.  So the “process” claim is really an attack on the end result, that being the FDA’s application of its criteria for approval.  It’s also a blatant second-guessing of the adequacy of the information made available to the FDA – which brings into play the same policy factors that undergird the Buckman decision:

Conversely, [the] claims would also cause applicants to fear that their disclosures to the FDA, although deemed appropriate by the Administration, will later be judged insufficient in state court. Applicants would then have an incentive to submit a deluge of information that the Administration neither wants nor needs, resulting in additional burdens on the FDA's evaluation of an application.
531 U.S. at 351.  No two ways about it, the "process" claim in Wimbush has to involve some allegation that the submission, and therefore the FDA's approval decision, was based upon "insufficient" information a la Buckman.

Oddly, the Wimbush opinion doesn’t even cite Buckman (we can't believe that the defendant would have failed to argue it).  Instead, the court goes on to distinguish Geier v. American Honda Motor Co., 529 U.S. 861 (2000), which didn’t involve any challenge to the validity of the government’s decision that the automobiles in question met federal standards and could be sold, and Riegel v. Medtronic, Inc., 552 U.S. 312 (2008), which was an express preemption case and thus quite different.  See 2010 WL 3256029, at *10.  In a footnote of all things, Wimbush asserts that there’s no difference between “complementary” warning claims as in Levine and a claim that an FDA approved drug should never have been marketed:

Both the district court, and [the defendant] seek to distinguish Levine . . . on the basis that Levine involved a state law inadequate warning claim whereas the instant case involves a state law negligent-bringing-to-market claim.  [This] is indeed true. . . .  However, we find this to be a distinction without a difference, as the rationale in Levine applies with equal force to this claim as it did to the inadequate warning claim.  Just as state tort law on adequacy of warnings can be seen as “complementary,” to the FDA’s labeling regulation, so too can state law duties regarding the decision to bring a product to market be seen as complementary to the FDA’s function of approving a drug for market.  This makes sense, as whether the FDA approves a drug for market depends, in very large part, upon the results of the manufacturer's investigation and testing prior to seeking FDA approval.  If the manufacturer is negligent in this investigation, then the entire FDA approval process is tainted from the outset.
2010 WL 3256029, at *11 n.7 (various citations omitted) (emphasis added).

Tainted?  Does that mean that a plaintiff gets to ask a jury to ignore the FDA approval?  That's a pretty raw conflict.  Maybe it somehow got lost due to the fact that in this case (but probably not the next) the drug in question was removed from the market.  There's no conceivable analogy to Levine there.  It is, however, exactly the conflict underlying preemption of fraud on the FDA claims.

There are any number of other ways in which this strained analogy also fails.  First, allowing a claim that a label should have been updated sooner than it was doesn’t implicate the FDA’s original approval of the label.  Second, the manufacturer’s reaction to new information in the labeling context doesn’t implicate the adequacy of information previously sent to the FDA, and thus doesn’t implicate the Buckman factors.  Third, “categorical liability” – that a legal product should never have been sold at all – is not a traditional tort claim, whereas an inadequate warning claim is.  E.g., Restatement (Third) of Torts, Products Liability §2, reporter’s notes to comment d, part IV(D)) (1998).  Fourth, the contention that “the entire FDA approval process is tainted from the outset” is identical to the allegation underlying a preempted fraud on the FDA claim.  Fifth, causation is much different.  A warning claim is subject to the usual strictures and limits imposed on drug claims by the learned intermediary rule.  Causation in a claim that a drug should never have been marketed at all, like a fraud on the FDA claim, skips the prescriber entirely, and can be argued to turn on the mere fact that the drug was available for sale.

And most fundamentally, for the precise reason that the majority in Levine stayed away from the failure-to-contraindicate claim in that case, the claim that Wimbush waved through the preemption gate is a direct second-guess of the FDA.  Indeed, the claim in Wimbush is worse than the claim Levine avoided because, while the Levine claim challenged only one approved use of the drug (administration by IV push), the claim in Wimbush is that the FDA should never have approved that drug for any purpose at all (the use in question being off-label).  That type of claim strikes at the core purpose of the FDA, which is to determine what drugs can and can't be sold in this country.

If the Sixth Circuit in Wimbush said “bring ‘em on” to tort claims second-guessing FDA decisions, Judge Posner said “no way” to another claim of the same nature.  In Robinson v. McNeil Consumer Healthcare, slip op., 2010 WL 3156548 (7th Cir. Aug. 11, 2010) (previously discussed here and here), the plaintiff claimed that the drug in question, which the FDA allowed to be sold over the counter without a prescription, should have been a prescription-only drug. Again, as with the Levine failure-to-contrainticate non-claim, and the Wimbush should-never-be sold-at-all claim, the claim in Robinson amounted to state law saying “no” to something (over the counter sale) about which the FDA had said “yes.”

Judge Posner, at least, was having none of it.  Deciding whether a drug should be generally available OTC or restrictively available by prescription only was the job of the FDA, not a common-law jury:

Although the plaintiff contends that it's unreasonably dangerous, or defective in design, either contention implying (if accepted) that the drug should be taken off the market, at the oral argument of the appeal her lawyer said “we’re not saying take it off the market . . . we’re not saying it should be banned.”  So maybe the plaintiff just wants it sold by prescription only.  But this would increase the drug’s cost (because of the prescribing physician's fee and the time required by the patient to obtain and fill the prescription) and thus reduce its availability . . . .

The decision whether to permit a drug to be sold over the counter rather than just by prescription is for the FDA to make.  The agency bases its decision on whether the drug is safe and effective for use without a doctor’s permission, and it has decided not to require that drugs containing [the active ingredient] be sold by prescription only.

Granted, that doesn’t bar a court from holding that state law requires warnings on the label of an over-the-counter drug beyond what the FDA has required.  [But] the FDA stated: “We believe that the overall benefit versus risk profile for [the drug] remains very favorable when they are used according to the labeled instructions. It is in the interest of the public health to maintain . . . a range of therapeutic options for the short-term relief of pain.
2010 WL 3156548, at *5-7 (lots of citations omitted – it’s Posner, after all).

So that’s two claims second-guessing the FDA addressed by appellate courts barely a week apart.  Clearly, the other side’s feeling emboldened to bring the very types of claims that the Supreme Court ran away from in Levine.  Here's a third.  In the Bartlett litigation that we’ve been following (here and here), there’s a “design defect” claim being litigated.  The defendant in Bartlett doesn’t seem to have raised preemption specifically against that claim – only making the stock generic preemption argument (and getting its head handed to it, see Bartlett v. Mutual Pharmaceutical Co., 659 F. Supp.2d 279 (D.N.H. Sept. 30, 2009), reconsideration denied, 2010 WL 2765358 (D.N.H. July 12, 2010).  We commented before on the potential for such claims to devolve into second-guessing the grounds for an FDA’s approval decision that a drug was safe and effective and should be marketed:

It will be interesting to see what Judge Laplante ultimately makes of a purely “design defect” trial involving a prescription drug.  Will he allow the plaintiff to argue that – contrary to the FDA’s decision not to force NSAIDs off the market due to the SJS risk – a civil jury can conclude that the drug’s risks outweigh its benefits and the drug should not be sold for an approved indication?  Most “design defect” claims involving drugs devolve either into:  (1) “FDA was wrong” claims; or (2) “failure to recall” claims that, as we’ve discussed before aren’t recognized in any state.  Maybe we’ll have a Bartlett III to look forward to where Judge Laplante delves characteristically deeply into these questions.
It turns out, there have been a bunch of Bartlett decisions in the mean time, and – sure enough – this “design defect” claim is shaping up as a common-law second-guessing of the FDA’s weighing of the risks and benefits of the drug.  See Bartlett v. Mutual Pharmaceutical Co., 2010 WL 3210724, at *1 (D.N.H. Aug. 10, 2010) (“a product’s risks and benefits must be evaluated with reference to the public as a whole, not [plaintiff] in particular”); Bartlett v. Mutual Pharmaceutical Co., 2010 WL 3092649, at *6 n.7 (D.N.H. Aug . 2, 2010) (discussing “risk-utility analysis” in the context of “various alternative drugs”); Bartlett v. Mutual Pharmaceutical Co., 2010 WL 3239247, at *2-4 (D.N.H. July 30, 2010) (“a product is defective if its risks outweigh its benefits so as to make it unreasonably dangerous”; “proof of an alternative design” is not “an essential element”); Bartlett v. Mutual Pharmaceutical Co., 2010 WL 2889114, at *8 (D.N.H. July 22, 2010) (“other NSAID side effects that [plaintiff] never suffered . . .[are] is highly relevant to determining whether [the drug’s] safety risks outweigh its medical benefits”).

Since weighing the risks and benefits of drugs is precisely what the FDA does during its approval process, a common-law cause of action predicated on “risk/benefit” analysis as the basis for finding a drug defective also amounts to second-guessing the FDA’s approval decision. While the defendant in Bartlett doesn’t seem to have put that proposition squarely before the court, if the case goes in that direction, it’s another example of plaintiffs ignoring Levine and arguing that the jury can say “no” where the FDA said “yes.”

Finally, there’s the recent decision in Lance v. Wyeth, slip op., 2010 WL 2991597 (Pa. Super Aug. 2, 2010), (discussed here), which permitted a “negligent design” claim in a prescription drug case under Pennsylvania law.  That’s a variant of the claim that was held not preempted in Wimbush (they’re both fen-phen cases), but Lance addressed state law issues that Wimbush let slide (and, conversely, didn't have preemption before it).

This, too, could turn into a second-guessing of the FDA (plaintiff can be relied upon to try), but fortunately it’s less likely in Pennsylvania than in Wimbush or Bartlett, for a number of reasons.  First, and foremost, Lance specifically held that Pennsylvania law “defers” to the FDA in decisions whether a prescription drug should be on the market.  2010 WL 2991597, at *7 (“we defer to the federal regulatory scheme and the FDA's decision as to whether a drug should lawfully remain on the market”).  Second, the plaintiff (unlike in Bartlett) in Pennsylvania darn well better have an alternative design – pure risk/utility balancing isn’t allowed.  “The determination of whether a product was negligently designed turns on whether an alternative, feasible, safer design would have lessened or eliminated the injury plaintiff suffered.”  Berrier v. Simplicity Manufacturing, Inc., 563 F.3d 38, 64 (3d Cir. 2009); see Aaron v. Wyeth, 2010 WL 653984, at *11 (W.D. Pa. Feb. 19, 2010) (dismissing negligent design claim in prescription drug case for lack of alternative design).  Third, defendants can argue that compliance with governmental regulations – that is, obtaining FDA approval, establishes due care.  Brogley v. Chambersburg Engineering Co., 452 A.2d 743, 745 (Pa. Super. 1982).

However, the mere fact that claims second-guessing the FDA are proliferating is in itself troublesome, since there’s no basis in Levine (whatever Wimbush might assert) for allowing them.  The other side won Levine only by kicking to the curb claims like failure-to-contraindicate that would have said “no” to uses – and modes of distribution, and products, and designs – to which the FDA has said “yes.”  In that respect, we’d like to see Wimbush appealed, but with the Supreme Court having heard four FDA-related preemption cases over the past three years, who knows?  Either they're extremely interested in the subject or just about sick of it.  But we won't know until someone tries.

No comments: