The order of operations can matter.  Back in elementary school, you may have learned a mnemonic about somebody’s aunt to help you remember the right order for doing certain math problems.  In computer programming, engineering, auto repair, surgery, and a myriad of other endeavors, you can get very different results if you take the same steps in a different order.  Some ordering may be obvious—anesthesia before incision in surgery—but the importance of proceeding in a specific sequence may only become apparent with the benefit of experience.  When it comes to pleadings and preemption, we have been beating the order of operations drum since some of our first posts, noting the pandemonium that can result when a court analyzes issues out of order.  (We have decried a similar dynamic with novel claims.)

The allegations and issues in Evans v. Gilead Sciences, Inc., No. 20-cv-00123-DKW-KJM, 2020 WL 5189995 (D. Haw. Aug. 31, 2020), are very similar to those in a decision we lauded last month for its directness and brevity.  The order of analysis, however, was different, resulting in a longer, less than a full win, and a chance for a third strike for the plaintiff.  It is worth noting that the Evans plaintiff was pro se and proceeding in forma pauperis.  Aside from the Latin lingo, this means that the court had already dismissed plaintiff’s first complaint and the motion to dismiss was on his second attempt.  It also meant, although there is no legal requirement for this, that the plaintiff was cut some extra slack.  We suppose that a represented party, had it done like the plaintiff here—filed a complaint that “lacks focus” with claims that are “difficult to deconstruct” and responded to a motion to dismiss with a cut-and-paste from “an antitrust class action complaint . . . which is irrelevant to Evans’ product liability claims and the arguments Gilead raises for dismissal”—would have gotten the boot without much analysis.  Instead, the court undertook a fairly detailed, if out of order, analysis.

The basic facts are that plaintiff claimed joint pain from his use of the prescription medication Truvada for either Pre-Exposure Prophylaxis or treatment of HIV.  He was prescribed the drug in August 2018, more than a decade after Truvada was approved and a mere three months after FDA had approved an update to labeling.  The physician and patient labeling had extensive disclosures about bone mineral density and kidney risks, recommending monitoring for certain patients.  Plaintiff made a design claim based on the idea that one of the active ingredients in Truvada should have been a different compound that Gilead also had approved (swapping in TAF for TDF in the lingo of the case), a warnings claim based on a mishmash of allegations, and boilerplate allegations about fraud and warranty.  Gilead moved to dismiss on the adequacy of pleadings and preemption.

After recounting the basics on these issues and Hawaii product liability law—of which our only confusion was the court’s reference to the “trilogy of preemption cases the Supreme Court has decided involving drug manufacturers” when it discussed Albrecht along with Levine, Mensing, and Bartlett—the court launched into its preemption analysis first.  As the court later pointed out, preemption is an affirmative defense and a constitutional one at that.  The pleadings challenge should have been considered first.  Ultimately, the claims for fraud—which requires heightened pleading—express warranty, and implied warranty (without specifying whether the merchantability or fitness variant was being pursued) were easy to dismiss, even though plaintiff was offered a third chance to plead them.  Warnings was a little more complicated, although it met the same fate.  Hawaii follows the learned intermediary doctrine, so plaintiff had to allege facts supporting that his prescribing physician would have changed his behavior and avoided plaintiff’s alleged injury had he been provided with additional or different warnings.  Plaintiff did not, so he failed to plead proximate cause and his warnings claim was dismissed, again with a third chance afforded to fix it.

To get there, however, the court walked through the allegations about inadequate warnings, which included alleged inadequacies in the patient warnings—which should have been irrelevant because of the learned intermediary doctrine—and the physician labeling, both as to bone risks and kidney risks—which should have been irrelevant because plaintiff did not allege a kidney injury.  Ultimately, the court, relying largely on California law, strung together an argument about inadequacies in the extensive physician labeling particularly as to the phrasing of the disclosure of joint pain and the bone density monitoring recommendation and concluded that “a reasonable jury could conclude that these warnings were inadequate.”  Setting aside whether we think that was correct, it clearly went nowhere given the lack of allegation of proximate cause.  And, perhaps more importantly, this analysis of the adequacy of the warnings allegations was not used to focus the analysis of whether a warnings claim—if a proper one were alleged—would have been preempted.  The court had already decided, in a rather abstract consideration of the issues, that it was not.

That is our main problem with the decision.  Recall that the prescription here—the one plaintiff did not allege would have changed with different warnings—was made three months after FDA approved the labeling that plaintiff was challenging.  In this context, given Levine and Albrecht, the inquiry should have been whether there was an allegation of new information that could have triggered the need for a CBE submission about the allegedly inadequate warnings—that is, the phrasing of the disclosure of joint pain and the bone density monitoring recommendations—during that narrow window.  Given that many of the cases the court cited acknowledged the narrow application of CBEs—something Levine butchered—the absence of such a specific allegation in the complaint should have been the end of it, assuming the court had decided to address preemption at all given warnings had not been pleaded adequately.

Instead, the court offered two justifications for finding no preemption on this record.  First, the court said plaintiff did not have to plead facts suggesting that a CBE submission could have been used because, as we noted before, preemption is an affirmative defense so plaintiff did not have plead around it.  Except that the CBE is a narrow exception to the general rule that a drug with an approved label, like this one had, must be marketed with the approved label.  To do otherwise is to violate FDA requirements.  So, without pleading something suggesting a CBE could have been used post-approval of the label being challenged, the plaintiff’s allegations did “admit[] all the ingredients of an impenetrable defense.”  Second, the court required Gilead to have proved by “clear evidence” that the FDA would not have approved a change to the drug’s warnings.  Given that plaintiff had not articulated a grounds for changing the label at all or what it should have said different, this was not the right inquiry.  We also submit that detailed labeling approved three months earlier with no allegation of new evidence of risk since is pretty clear evidence of the futility of trying to change the label for no reason.

Before the warnings analyses, the court addressed whether the design defect claim was preempted.  It never addressed whether plaintiff satisfied his pleadings requirements or whether Hawaii law would recognize a claim based on the core allegation that a medication should have utilized a different active ingredient, which it also sold.  (We have been making noise about the latter issue for a long time, too.)  Rather, it went straight to preemption, generously analyzing both the possibility of a pre-approval design claim and a post-approval design claim from the same muddled complaint.  This analysis was thorough and correct.

As for pre-approval, Mensing forecloses the idea that Gilead could have independently brought Truvada to market with TAF instead of TDF as one of its active ingredients.  Getting a different NDA approved cannot be assumed.  The court noted outlier decisions and followed the Supreme Court’s guidance from Bartlett and MensingBartlett (and Yates) rejected that the defendant could be liable for just not selling the drug.  Mensing rejected that the suggestion preemption does not apply where the defendant could have tried to get a different drug on the market.  Citing Utts I (but not Utts II), the court held broadly that “pre-approval design defect claims against brand-name drug manufacturers are preempted by the FDCA.”

Post-approval fared no better.  Bartlett makes it clear that FDA has to approve a labeling change describing a change in the formulation.  So, a CBE is not available and the approval of the label in May 2018 listing the formulation of Truvada meant that a design defect claim based on changing the formulation between May 2018 and August 2018 would be preempted.  Pretty cut and dry, except for the cognitive dissonance over why the analysis of preemption of the warnings claim was so different.  And plaintiff will not get a chance to try to plead around preemption on his purported design claims, an exercise the court labeled “futile.”  When plaintiff does come back with a third try on his other claims, it will be interesting to see if the defendant gets a second chance with preemption.