We’ve been following the Levaquin litigation mostly from afar(for which we should be grateful), and often in white-knuckled silence, as one problematic ruling after another was handed down. The not-all-that-surprising result was that a test case called Schedin (under which still more adverse rulings may be found) produ ced a finding of not just liability, but liability for punitive damages.

Well, recently, the Eighth Circuit got a shot at the Levaquin [fill in descriptive noun of your choice]. Unfortunately, while it addressed some of the most egregious errors, it didn’t repair anywhere near all the damage. See In re Levaquin Products Liability Litigation, ___ F.3d ___, 2012 WL 5971181 (8th Cir. Nov. 30, 2012).

Most importantly, the court of appeals held as a matter of law that there wasn’t enough evidence to justify the imposition of punitive damages (which reduced the size of the verdict by well over half):

As a matter of law, the record evidence failed to establish [that the defendant] deliberately disregarded the risk of tendon injuries in elderly patients taking corticosteroids, as required for punitive damages under Minnesota law. By warning of that risk in its package insert, [defendant] actively sought ways to prevent the dangers associated with its product. The 2001 warning also was published in the PDR, a reference widely used by physicians. Regardless of [defendant’s other] alleged actions . . ., we cannot characterize [it] as hiding information it openly published. The 2001 warning was in [the prescriber’s] physical possession and was specific and clear if read. For drug warnings to succeed in protecting patients, doctors must order their practice and their continuing medical education so as to find time to learn about new and updated warnings for the drugs the doctor is prescribing.

Id. at *7 (emphasis added). That’s good, we like cases holding that punitive damages are barred by warnings, even if inadequate. Moreover, because the plaintiffs undoubtedly threw all the mud in their possession against the wall in this bellwether trial, this ruling probably kills punitive damages for the rest of the Levaquin MDL. It’s pretty hard to conceal deliberately something that you specifically warn about.

Still, since the District Court construed the Minnesota punitive damages bifurcation statute narrowly and allowed in all sorts of prejudicial evidence only “tangentially” related to the compensatory claims, In re Levaquin Products Liability Litigation, 2010 WL 4867588, at *3 (D. Minn. Nov. 23, 2010), we think that the entire verdict should have been reversed. We know how punitive damages work in practical terms, and we have no doubt that admission of evidence purportedly relevant only to “motive and intent” id., seriously prejudiced the presentation of the entire case before the jury.

Then there’s failure to warn. The plaintiffs were allowed to proceed on two theories, including a bizarre failure to “include comparative . . . toxicity information in the package insert” claim. 2010 WL 4867588, at *3. The Eighth Circuit pointedly did not endorse the comparative labeling claim. Id.(“we need not address whether the district court erred in denying [defendant's] motions based upon [plaintiff’s] comparative toxicity theory”). The court found “harmless error” because the plaintiff could recover on the other, more standard, warning theory. Id.

We think that’s a cop-out. In the first place, the FDA strictly regulates when a manufacturer can make product comparisons. See 21 C.F.R. §§201.57(c)(2)(iii); 201.80(c)(3)(v) (requiring “substantial evidence derived from adequate and well-controlled studies”). In the second place, before Levaquin, no state anywhere had held that the pharmaceutical duty to warn included an obligation to recommend somebody else’s product as “safer.” Comparative warning claims were rejected in Baycol litigation. In re Baycol Products Litigation, 532 F. Supp.2d 1029, 1040-43 (D. Minn. 2007). This issue thus provokes one of our largest pet peeves – federal courts exercising diversity jurisdiction have no power to “predict” novel and expansive theories of tort liability. Yes, even in the Eighth Circuit. E.g.,Leonard v. Dorsey & Whitney LLP, 553 F.3d 609, 612 (8th Cir. 2009) (“[o]ur duty is to conscientiously ascertain and apply state law, not to formulate new law based on our own notions of what is the better rule”).

We have the same objection to allowing liability on the other, more normal warning claim – because the defendant actually did warn. It “changed the tendon warning in the package insert” to make it stronger. Levaquin, 2010 WL 4867588, at *4. Nonetheless, liability was affirmed:

Courts disagree about whether simply changing the package insert warnings insulates a drug manufacturer from failure-to-warn liability, and Minnesota courts have not decided this issue. Many courts considering the question have held a properly worded package insert is a sufficient warning as a matter of law, at least when it is combined with an entry in the PDR.

Id. That should have been the end of it. The plaintiff should have gone home empty-handed. Since Minnesota state courts admittedly have not recognized such liability, and it’s certainly not the majority rule elsewhere, a federal court supposedly applying state law can’t go making things up. Leonard, supra. Only by characterizing the question as whether warnings “insulate[]” the defendant, 2010 WL 4867588, at *4, rather than whether plaintiff could maintain a claim on the first place for inadequate warnings in the presence of an adequate package insert, could the court pretend that it was declining to make new law. Of course, it was really making new – and quite bad – law.

We make the same observation about causation. The prescriber never testified that he relied on Dear Doctor letters, and the court conceded that it was a “stretch” to base warning causation upon speculation that he might have “relied on his colleagues’ comments about particular drugs.” Levaquin, 2010 WL 4867588, at *6. We reiterate. That should have been the end of it. The opinion cites no Minnesota law allowing causation to be premised on either warning letters or through some sort of gestaldt from colleagues. Federal courts should not create new grounds for liability from whole cloth.

So overall, we’re quite disappointed in the outcome. Assuming that Schedinwas the MDL plaintiffs’ preferred bellwether case, we have to conclude that these are very weak cases, indeed – a conclusion supported by plaintiff losses in two subsequent Levaquin trials. The courts should not be bending over backwards to encourage liability where none should exist.