Thank you for the opportunity to speak today. My name is Dr. Stephanie Fox-Rawlings. I am a Senior Fellow at the National Center for Health Research. Our research center analyzes scientific and medical data to provide objective health information to patients, providers and policy makers. We do not accept funding from drug or device companies so I have no conflicts of interest.
There is a need for better outcomes for patients undergoing lung transplants. New devices and/or new methods should be tested. It is difficult to design clinical trials to test ex vivo pumps, however it is necessary to demonstrate effectiveness and safety for the intended use of these pumps before approval. The INSPIRE trial was intended to demonstrate the equivalence of TransMedics Organ Care System (OCS) to the current standard of cold storage. However, the trial had significant design and implementation problems. As a result of those flaws, it could not show that OCS was as effective as the current standard of care.
The INSPIRE study had numerous opportunities for unintended bias to be introduced to make OCS appear more comparable to control.
1) Some study sites disproportionately conducted surgeries for one arm, so differences in site conditions that were unrelated to OCS vs. cold storage could alter outcomes.
2) The higher number of ‘off-study’ transplantations in the OCS arm is concerning. Since the study couldn’t be double blind, the difference in off-study transplantations could be caused by doctors moving lower quality lungs off-study instead of using OCS.
3) The criteria used to define screening failures and other protocol violations were not consistent between arms or sites. For example, FDA scientists identified examples of OCS lungs that could not be used for the study due “active” donor lung disease, but similar cases were allowed for the control arm. 74% of screening failures occurred in the OCS arm, which often led to “off-study” transplants. That could obviously bias the results. It would not have to be intentional; doctors in the OCS arm could unconsciously tend to avoid situations that would result in poorer outcomes.
Most important, the OCS did not meet the non-inferiority effectiveness endpoints designated before the trial started. To meet the primary effectiveness endpoint, both the analyzed population and one of the co-primary endpoints had to be modified. This occurred after an unplanned analysis that the FDA advised against. This type of post hoc manipulation of data does not meet acceptable scientific standards needed for clinical trials.
It would not be ethical to rely on post-approval studies to demonstrate efficacy and safety. Post-approval studies for devices tend to be small and are delayed for years. The proposed new-enrollment study is based on a registry, so it will be difficult to compare outcomes to cold storage.
There is tremendous pressure to have more treatment options, but if the FDA approves OCS despite the lack of unbiased evidence of non-inferiority, many lives will be put at risk. Improving the studies – including following the clear advice of the FDA – will answer all the essential questions about whether OCS is non-inferior to cold storage, and for whom.