I'm starting to cool off and forcing myself to think more critically here.
I think I'm going to ignore the biased sample population argument for now because clearly that wasn't the case in the supplementary Phase 1b (111) study as those patients had been made known about the same WR score criteria ahead of time. That trial went on to see stat sig improvements that were not observed in the Phase 2 trial. So I think that's a red herring and we should focus our attention elsewhere, although I don't deny that trial design is critical moving forward.
I posted this some time ago, but I do remember reading somewhere that the cochlea, being a very delicate membrane, has physical limitations to what can be injected into it before it breaks. Chris Loose also said today that they hadn't trialed multiple doses in animals before because the cochlea would be overwhelmed/break. I'm wondering then whether that analogy -
that when you re-seed the lawn you must stay off the grass - does carry some weight. Because regardless of whether someone received 1 dose or 4 doses, EVERYONE received 4 doses of the carrier gel, regardless of whether it was carrying the small molecules or not. I also recall reading in Will McLean's paper that a colony took 12 days to form, which begs the question: why the hell did they only give it one week in between each dose? Even two weeks would be pushing it.
Is it possible, for example, that 4 doses of FX-322 was actually equivalent to 1 dose, because each previous dose washed away the other? On this point, do we have any historical data to glean from regarding multiple dosing for any ear-related condition? For example, is there a study out there that looked at multiple weekly injections of intratympanic steroids against placebo? I'm pretty sure this is a discussion some of us had over on the Otonomy thread about Meniere's and I don't remember finding anything other than maybe one study. In any case, if the lawn analogy is true, this could explain why they didn't see differences between 1x, x2 and x4 FX-322 groups. Unfortunately, this still doesn't explain the lack of difference between placebo within each category. The only explanation we have for that is the failed trial design.
So we have some plausible explanations given the above, but how does one still reconcile this with the failed EHF audiogram in the Phase 1 study which showed no placebo effect? The only explanation we seem to have, as
@FGG and
@Diesel have pointed out, is that there is preferential IHC regeneration over OHCs in vivo - the idea being that IHC regeneration will not appear on an audiogram. But again, if IHCs restore clarity, why have we not found statistically significant results in the WIN scores? The only answer again is design error and a biased sample of patients.
To put things simply, for this Phase 2 outcome to have happened, it would have taken a catalogue of errors and unique set of circumstances. So I'm going from "this drug doesn't work" to more of a "perfect storm" take on this. Is it possible this was the perfect storm? We just don't know yet, but obviously things don't look too good right now.
This may seem like a tangent, but I have an interest in air crash investigations. What you find in almost all cases of air crashes is that there is almost never just one cause. You find that there are a catalogue of errors, one after the other, usually combined with a unique set of circumstances, that all together lead to the crash. Perhaps this is one of those occasions.