Hi Stanjupier – very quickly – in response to a couple of...

Hi Stanjupier – very quickly – in response to a couple of questions you have asked me.

At the AdCom meeting no-one (not the experts, FDA or MSB) expressed any difficulty with designing a RCT for adults with GvHD.

According to MSB they were well advanced in the design of the adult study and had been in discussion with centres to take part in the study.

If you are interested in what the design of the study that MSB are proposing the best bet would be to ask MSB. Not me or Davisite.

In terms of MSB leaving no stone unturned to avoid another trial? This is a little bit of a no-brainer. In 17 years and a couple of dozen RCTs MSB have never met a per-specified endpoint – for anything. When you don’t meet per-specified endpoints in trials this is very problematic. In the context here … it would be fatal for approval.

“Southoz has said the FDA wanted matching at the patient level. I take it that means ensuring the characteristics are balanced between groups in a RCT?”

No – you are misunderstanding this point.

The FDA rejected MSBs approach to deriving the control group estimate in the single arm GvHD trial. The FDA asked MSB to provide further justification for the control group estimate and so MSB conducted the MAGIC study.

The MAGIC study essentially went okay lets match at a group level patients from the MAGIC database with patients who were in the single arm trial. So that at the end of this process on key prognostic factors you would have a rough equivalence – eg both groups would have a similar average age.

The FDA rejected this approach. What it wanted was MSB to match at the individual level – a case control study. That is if the single arm study had a child 10 years old with prognostic factors XYZ find a similar child in the MAGIC database to match to. And to do this within a confirmatory framework – that is pre-specify the “control” endpoint estimate.

According to the FDA the reason this misunderstanding of what they wanted and what MSB did occurred because MSB failed to submit its statistical analysis plan to them. So the MAGIC study turned out to be a waste of money – it didn’t convince the FDA of anything.

This mistake in MAGIC, together with the error that MSB had made in its original approach to the control group estimate and the totality of the negative evidence from 17 years worth of methodologically stronger (RCTs) trials - meant game over for MSB - despite the 9:1 vote.