Sixth Circuit Affirms Exclusion Of General Causation Expert in Onglyza MDL

The MDL and state court proceedings involving saxagliptin-based diabetes drugs (such as Onglyza and Kombiglyze) strike us as the mass tort that never should have been.  These proceedings initially followed a familiar model—a publication identified a signal of a risk (albeit an exceptionally weak signal), and plaintiffs’ lawyers took their cue to collect their inventories in the usual ways and file lawsuits.  Those lawsuits, however, have largely crashed and burned in both state and federal court, and for the same reason.  The plaintiffs were unable to find competent expert opinion that saxagliptin is capable of causing heart failure.  We gave you our take on the MDL judge’s order excluding the plaintiff’s expert here, and the California appellate opinion affirming a similar order came in as our fifth best drug/device case of 2023. 

The Sixth Circuit has now joined the chorus.  In In re Onglyza & Kombiglyze Products Liability Litigation, No. 22-6078, 2024 WL 577372 (6th Cir. Feb. 13, 2024) (to be published in F.4th), the Sixth Circuit held that the MDL judge correctly excluded the plaintiff’s general causation expert and granted summary judgment, and it also upheld the district judge’s refusal to allow the plaintiffs a second chance. 

To recap, the plaintiffs in Onglyza alleged that the medication caused heart failure and related cardiac conditions.  The FDA had asked the manufacturers to conduct a randomized controlled trial with multiple cardiac endpoints, which they did.  The result was that there were no statistically significant differences between the drug and placebo for any primary endpoint, which included cardiac death, heart attack, and ischemic stroke.  The same was true for secondary endpoints—except for one:  Hospitalization due to heart failure.  The study authors noted that the heart failure finding was unexpected and that it should be considered within the context of multiple tests that may have resulted in false positive results.  The authors also cautioned that “a class effect should not be presumed.”  Id. at *1-*2. 

That is the scientific backdrop against which the plaintiffs offered a single general causation expert, whose opinions were ultimately inadmissible for three reasons.  First, the expert relied on a single study (and an inconclusive one at that), to the exclusion of all other studies involving human data.  Four later observational studies found no association, and even the plaintiffs’ expert agreed that the studies were “reasonably designed.”  He nonetheless ignored them, purportedly because of “issues related to confounding,” but he identified no specific issues or confounders in any of the studies.  As for the one study, it was “no smoking gun.”  It observed only a higher incidence of hospitalization in patients treated with saxagliptin, not a causal link.  And the study authors stated that there “are presently no known mechanisms” by which the drug could cause heart failure.  Id. at *4.  The plaintiffs complained that the district court had imposed a “one unreplicated study is not enough” standard, but that was not true.  The district court properly concluded that the expert acted unreliably by relying on the one study while ignoring all other human studies.  Id.

Second, without human data to support his opinions, the expert attempted to rely on animal data instead.  He had, however, no expertise in interpreting animal studies, and he conceded that he was unqualified to conclude that the animals in the studies “actually had heart failure.”  Indeed, multiple animal studies actually dispelled a causal link between saxagliptin and heart failure, which the expert admitted.  Id. at *5.

Third, the expert did not reliably apply the Bradford Hill criteria.  Experts often fall back on the Bradford Hill criteria to bolster their causation opinions.  But here, there were multiple grounds to find that the plaintiffs’ expert got it wrong.  He cherry-picked data.  He drew an “analogy” to a different diabetes medication, but could not explain why, other than to assert that it was “appropriate” because the other medication had “been shown to ‘worsen heart failure.’”  Id. at *5.  A more blatant admission of result-oriented data selection is hard to imagine.

The expert also  applied other factors inconsistently—for example, by claiming in his report that two factors (specificity and biological gradient) were satisfied, but later testifying that they were not.  Id. at *5-*6.  Plaintiffs’ response to this was predictable:  Leave it to the jury to “evaluate and weigh” the testimony.  But that solution applies only to the evaluation of admissible evidence, and “district courts may allow juries to evaluate and weigh only relevant and reliable expert testimony.”  Id.  The decision underscored these points by repeatedly relying on the “amended” language of Rule 702 requiring proof that an expert’s methodology is “reliably applied” to the facts of the case.  Id. at *3, *5, *6 n.7.

Without admissible expert opinions on general causation, the Sixth Circuit affirmed summary judgment, too.  To avoid this result, the plaintiffs first tried to argue that expert causation opinion was not necessary for them to prove their case.  We have not studied the evidence, but we are guessing the plaintiffs were offering to prove their case with internal company documents, regulatory documents, adverse event reports, and the like.  The Sixth Circuit, however, rejected that approach.  The district court reviewed the law of all fifty states and determined that every state requires “the plaintiff in cases involving complex issues of medical causation to present expert testimony.”  Id. at *6.  The plaintiffs could not cite a single case actually holding otherwise.  Id. at *7.  We’ve discussed district court opinions reaching this 50-state conclusion, but we believe Onglyza is the first appellate court to so rule.

The plaintiffs also wanted a second chance to find another expert, but the district court declined, and so did the Sixth Circuit.  Although they characterized their request as an “extension,” they were really asking for a complete do-over of expert discovery, and it was well within the district court’s discretion to say no.  The plaintiffs could not explain why they failed to identify other, reliable experts despite “years of expert discovery.”  Id.  Further, their argument that the “minimal delay” would cause “no prejudice” to defendants was not credible.  The plaintiffs wanted to “essentially restart expert discovery, requiring depositions, briefing, hearings, and motions on plaintiffs’ new expert.”  Id. at *8.  To put a finer point on it, “[t]his would delay the MDL’s resolution by for years—just consider that plaintiffs requested three months to simply identify an expert.”  Id.  The substantial legal expenses and years of delay would most certainly cause prejudice to the defendants. 

We particularly like this last ruling.  Allowing plaintiffs to disclose one expert, then try again if that expert did not work out, would drag out litigation seemingly forever.  As the Sixth Circuit held, “The [district] court properly refused to reward plaintiffs for their failure to identify a reliable general causation expert by imposing significant costs on defendants”  Id. at *8.  We hope the judge in the Acetaminophen MDL is listening.