In response to Neuroinvest I summarized my recent thoughts on the Dex Ph III trial on the Yahoo thread. Just to keep track of the post I'll duplicate it here as well.
Just some overall summary on my 'empirical analysis'. I've now raised my confidence in the ph III to near 90%. It really isn't possible to go higher - there are always unknown unknowns. Just ask the Vioxx people.
My rationale for yet another raising since my last one:
1) People say that neuroprotectants are where drugs go to die. So I researched the heck out of past failures. None(!) that I looked at had Ph II results anywhere near as promising as the Dex trials (albeit all the ones I researched were stroke). I might be unhappy that Pharmos culled the bad arm, or, ... . But in comparison to previous Ph II's their results are awesomely good and the paper is well thought through.
2) People say that there are lots of surprises:
a) Surprise safety problems - and indeed in the trials I looked at there were a fair number of 'surprise' mortality problems. But all but one had at least moderately significant Ph II tox issues. It is possible something will crop up in Dex Ph III, but I'd guess efficacy will more than offset it.
b) Surprise placebo improvement. A key example given to me was the Citiclone trial. But the Citiclone trial shouldn't have been a surprise. The Ph II didn't allow thrombolytic agents, but the ph III did. That inevitably increased the control arm and hit a ceiling effect. (Note that the other reason that the ph III failed, as they acknowledge in the Ph III paper, is that they got too tricky with their post hoc analysis of the Ph II results and picked a poor endpoint.)
3) The Citiclone trials, taken in combination, are fairly convincing that there is efficacy in Citiclone. That provides proof of concept to Dex that neuroprotectants don't just work in rats.
4) The biggest potential weakness in the Dex Ph III is that they overshot or undershot when they changed the entry criteria to avoid the measurement ceiling effect obvious in the Ph II. Pharmos has much more data than is publicly available to know how likely/unlikely this is but still it is a point of concern. But I think that their agreement with the FDA on calibrating the individual subgroups separately probably largely mitigates this worry. The groups that hit the ceiling won't swamp out the groups that didn't.
In total I'd say that there is still risk just because of the unknown unknowns, but I've put a bunch of those to bed and feeling ever more comfortable with the probability of ph III success.
BTW - On this board, where I will not be eviscerated just for whispering about a possible failure mode, I thought I'd add some thoughts/examples on the unknown unknowns that could bite us. Maybe in Israel, the location of the successful ph ii, there is a particular treatment standard for TBI which acts in synergy with Dex and makes it vastly more effective even though Dex or the synergistic agent alone does nothing much. But this synergistic agent is not used outside of Israel or stopped being used after 2001 (the Ph ii's were before 2001). This is not likely, but is really just meant to be an example of an unknown unknown that is essentially unknowable. I've been bitten by them in real life.
Clark |
