FDA Panel Nearly Unanimous on Liver Transplant Perfusion Device

TransMedics’ Organ Care System liver perfusion device got near unanimous support for an indication in liver transplantation by an FDA advisory committee.

The agency’s Gastroenterology and Urology Devices Panel voted 14-0 in favor of both efficacy and safety, supporting an indication for the resuscitation, preservation, and assessment of liver allografts from donors after brain death or selected donors after circulatory death (DCD, age ≤55 in a near-physiologic, normothermic and functioning state).

The only dissension came on whether the benefits outweigh the risks for the portable extracorporeal liver perfusion and monitoring system. The panel had one “No” vote and one abstention, both indicating the limitations of the data that came up over and over at the meeting Wednesday.

“I don’t think that the benefits have really been proven,” said Colleen Gallagher, PhD, a critical care physician at the University of Texas MD Anderson Cancer Center in Houston, in explaining her “No” vote.

Even so, she fully anticipated approval. “The post [marketing] study will give us good data that may lead to proving the benefits,” she said, which might include increased utilization of donor livers and sharing organs over greater distances. “Those will probably be proven, but they’re not there yet.”

Nearly all the panelists likewise expressed discomfort with the evidence, most commonly due to the surrogate endpoint used in the pivotal PROTECT trial.

That 300-patient trial met both noninferiority and superiority for the warm perfusion device versus standard of care cold storage on early allograft dysfunction (EAD) within 7 days in both the per protocol (18.0% vs 31.2%) and modified intent-to-treat analyses (17.9% vs 32.4%). The majority of those cases were driven by aspartate aminotransferase (AST) elevations alone.

Whether EAD is a good surrogate for clinical outcomes was the core question, said panel chair Steven Schwaitzberg, MD, of the University of Buffalo, New York.

The key evidence supporting it was a 2010 study that defined EAD and linked it to a more than 10-fold elevated risk of mortality within 6 months and seven-fold risk of graft loss at that same point.

In PROTECT, survival as a secondary endpoint in the trial was an identical 99.3% in both arms at 30 days and north of 98% at discharge. Kaplan-Meier curves also showed similarity for both graft and recipient survival out to 24 months.

A much larger trial on the order of 2,000 patients would have been required to be powered for survival, TransMedics presenters noted.

The fact that there was a decrease in EAD that wasn’t reflected in survival muddied the waters, but wasn’t a sticking point for the FDA advisors. They called it acceptable, albeit not ideal.

“While I would have preferred a stronger clinical outcome, I believe this will suffice,” said Jason Dominitz, MD, MHS, of the VA Puget Sound Health Care System in Seattle, suggesting graft failure or survival would have been preferable.

Julie Heimbach, MD, of the Mayo Clinic in Rochester, Minnesota, was willing to rely on the 2010 data. “Clearly, it gets to the biology and the impact of this device,” she said. “I have the same mixed opinion … that we can’t be 100 percent sure about efficacy with this data of EAD, but the data definitely look safe and it’s encouraging.”

Secondary endpoint findings that the warm perfusion device lowered ischemic biliary complications through 12 months (2.6% vs 9.9% among controls, P=0.02) were more impressive to Lynt Johnson, MD, MBA, of George Washington University’s Liver and Pancreas Institute for Quality in Washington, D.C.

“I would ask the question is whether survival is an appropriate endpoint, because really we’re looking at something that has immediate impact on the function of the graft. The biliary tree is more susceptible to ischemia than any other part of the liver,” he noted. “To me [it] validates the value of this technology, particularly in the DCD population.”

The FDA staff review of the data was critical of study design issues, including early randomization of participants prior to donor liver retrieval such that there could have been bias introduced in the treating surgeon accepting grafts based on known allocation to the device arm as well as re-randomization of patients for whom an organ was not accepted for transplant.

More than a third of donor livers considered for implantation were screen failures and thus excluded from the study, largely due to accessory veins, which are a contraindication for the device. And 30% of consented organ recipients were excluded from PROTECT with little data collected on them. One in 10 of those patients were randomized, but transplanted outside of the trial protocol and not followed for outcomes.

Panelists said they understood the complex, practical reasons for these design features, but wrestled with the influence it might have had on the outcomes. In the end, it wasn’t a “fatal flaw,” but did impact generalizability, said W. Ray Kim, MD, of Stanford University School of Medicine in Redwood, California.

But because of such issues, the advisory committee disagreed with TransMedics’ proposal to simply follow the PROTECT participants and those in the continued access program for 2 years to gain additional insights after approval.

The panel instead aligned with the FDA’s assessment of what would be needed for a post-marketing study, leveraging the TOP registry already established for TransMedics’ device for lung transplant, but adding that they would like to see more data on rejected organs and on DCD and older donor livers.

The FDA does not have to follow the recommendations of its advisory committees on approval but usually does.