Odds Ratios presented as a substitute for the trial population mean performance is exactly the issue. As secondary analysis/information Odds Ratios with n would tell a meaningful story. Anavex did not have the time to perform and present the full and hopeful secondary analysis at CTAD 2022.
There are clinical trials comparing one treatment to another where Odds Ratios are used to assess the relative risk of eg. some side effect and determined which treatment is the safest. There are no clinical trials comparing a treatment to placebo, that I know of, having been approved with Odds Ratios as the outcome measure. That is, if in an overall trial population N you get a subgroup n of patients with y threshold for this or that scoring system, it is a pass.
It is the other way round. That is, if in an overall trial population N the mean score from baseline to EOT versus placebo is clinically meaningful and statistically significant, it is a pass. Then in secondary analysis you might, and I think really should, ask the question - ok great so how many out the treated patients actually saw clinical benefit and was that subgroup stat sig vs those on placebo meeting criteria? Case in point the recent amyloid plaque removal drugs that don't really do much for not so many and comes with risks. Here there is a general issue of sponsors running very large trials to achieved traditional approval, but really not helping that many patients.
For the Anavex P2b/3 AD trial with A2-73, it could be that once the standard baseline to EOT mean score vs. placebo outcome measures all become known, it isn't an outright traditional pass. However, in secondary analysis of subgroups by genetics, other biomarkers, dose/concentration and this and that, it becomes clear beyond reasonable doubt that some n of the dosed trial population actually improved and others stabilised given some very specific and potentially complex characteristics, it could be an Accelerated Approval pass.
Anavex have done themselves, at least a short term, a disfavour by retracting the original 2 of 3 presented mean scores and throwing in some naked Odds Ratios instead and then having declared all endpoints met - on the face of it that simply does not compute. I would grant that in the circumstances of having the data very late before CTAD unable to complete that full subgroup secondary analysis Anavex chose perhaps in some desperation a simple/lacking responder analysis presented as trial population Odds Ratios with no n. They obviously were not prepared to present a failed P3 trial and wanted the time and grace to do the full subgroup analysis batting for AA from a P2b trial based on clear responder results that beyond reasonable doubt actually works well for a subset of AD patients.
If AA is granted then the met all endpoints statement will be forgiven, if not then the class action lawyers will be all over the case.
The longer we wait, the sooner we will get rich!
