Ann: Update on BLA for Graft Versus Host Disease, page-531

In reply to @reginaldp (HC is doing its normal thing of not allowing me to post a reply):

Many factors influence the timing of an inspection, including progress in the evaluation of the sponsor's data but also external factors such as other inspections in the region (easier to bulk things up and knock over several inspections with one trip to the other side of the world), availability of specialist inspectors, etc. But without question, the scheduling of the inspection is a positive.

As we know, lot variation is a vexed question for biologicals, although pathways are a lot clearer than they were a decade ago. Easy to demonstrate lot-to-lot consistency with a defined chemical entity - basic lab tests will be more than adequate.

I guess theoretically you'd need to conduct a large enough trial that you could detect clinical differences between lots (and with sufficient lots in the trial to be statistically certain you'd accounted for the potential variability). No idea how large such a trial would be, but properly powered it'd probably run into the tens of thousands or more, so not really practical.

Hence MSB looking for markers which correlate with clinical activity - in a strict statistical sense this still requires pretty large studies with a sufficient number of lots to pick up potential lot to lot variation. Not sure we've got enough data on sufficient lots to have achieved this, but equally not sure it is really practical (and the CBER understands these issues and how to deal with them). My guess would be that the current data will be sufficient to get it over the line.

BTW the issue is simpler with the MSB cells - the purification process much more controlled and likely to be much more reproducible than the Osiris cells. With the Osiris cells the variability between donors is compounded by the variability in the purification process. The other source of variation for both is of course the thawing process prior to administration - and this is difficult for MSB to control, and will potentially be even more difficult in routine clinical use.

The inspection itself will likely look at batch manufacturing records (I would be tempted to omit "likely" but I've seen too many exceptions!) - depending on the inspector he/she might ask about variations in parameters and in-process testing, or they might not! For some it's a box-checking exercise, for others not, it's a bit of a crap shoot!

In reality, I don't think lot variation will be a major issue, there's probably enough data there to satisfy FDA. The major issue remains that identified by @Southoz in the post above.

Not sure that helps, it's a complex issue!!