Ann: Remestemcel-L FDA Fast Track Designation for COVID-19 ARDS, page-357

From the FDA AdCom meeting notes:



What the FDA are saying here is that a single arm trial could potentially lead to approval but that the trial must meet the legal regulatory requirement - 21 CFR314.126.

Section 21 CFR314.126 requires that studies conducted for drug approval must be “adequate and well controlled”. It doesn’t matter what type of design is used; all types of studies must be rigorously performed.

The FDA also stated that modelling from historical controls is an acceptable way to derive a null hypothesis in a single arm trial.

When MSB announced the trial results it claimed:

“In the 55 children enrolled in Mesoblast’s open-label Phase 3 trial conducted across 32 sites in the United States, the Day 28 OR rate was 69%, a statistically significant increase compared to the protocol-defined historical control rate of 45% (p=0.0003).”.

Unfortunately, MSB did not have a historical control rate derived from historical control modelling at all. It had just pretended it did.

The FDA found:

“The null hypothesis for MSB-GVHD001 is not based on data from a historical control population. In the absence of data from appropriate historical controls, FDA is unable to agree that the proposed null hypothesis is acceptable.”

And so the wheels fell off. Once the null hypothesis was found unacceptable MSB no longer had a “adequate and well-controlled study”.

But then why did the AdCom member vote 9:1 in favour?

The question the AdCom members were asked to vote on was whether the available evidence supported efficacy. Which is obviously it does; the trial had produced positive results.

But this is not the question the FDA have to answer. The FDA are legally required by 21 CFR314.126 to consider whether the trial was “adequate and well -controlled”. They have concluded it wasn’t and hence the application was rejected.