Replies to post #465196 on Anavex Life Sciences Corp (AVXL)

[Doc328]

I just read the PR a minute ago. I still think the likelihood that AVXL files this year based on the single 2b/3 with either or both FDA and MAA is still 1-2 %.

From today's PR

For the primary endpoint ADAS-Cog13, blarcamesine is significantly better than placebo for both 50 mg (-2.149; P = 0.021) at 48 weeks and for 30 mg blarcamesine dosage groups (-1.934; P = 0.026) at 48 weeks. The key secondary endpoint CDR-SB was significantly improved vs. placebo in both 50 mg (-0.465; P = 0.045) and 30 mg (-0.502; P = 0.020) assigned dose groups.

From September 14, 2023

The trial was successful, since the differences in the least-squares mean (LSM) change from baseline to 48 weeks between the blarcamesine and placebo groups were -1.783 [95% CI, -3.314 to -0.251]; (P = 0.0226) for ADAS-Cog13, and -0.456 [95% CI, -0.831 to -0.080]; (P = 0.0175) for CDR-SB in patients with early Alzheimer’s disease

Note the very strange result saying 50 mg progressed 2.1 points less on ADAS-Cog than placebo AND 30 mg progressed 1.9 points less with p values around 0.021-0.26. Yet in September 2023, for 30 mg plus 50 mg using more accepted MMRM/LSM, combined dose was 1.78 fewer points progression compared to placebo and p=0.026.

There is no positive value solution for (n1+n2)*1.78 = (n1)*2.1 + (n2)*1.9 So only answer is that different cohorts were used. My guess is the new data is only using completers.(PP cohort) which is not an acceptable method for regulatory agencies. That also explains the high dropout rates (for the MMRM/LSM the mITT cohort was probably the standard of all patients who received at least one dose and had at least one post baseline visit (i.e. 12-13 weeks).

Another clue that there were different cohorts is that the p values for 1/2 sized 30 vs 50 mg groups are the same as for the larger combined cohort, compared to placebo.

The PR was written for investors not scientists or clinicians

Anavex is still playing tricks with the stats.

[frrol]

Some notes:
- Yes, we've now got an EMA submission target, Q4. Good to see.
- The trial's demographic profile not a concern.
- The dropouts are high and disappointing, and will be considered by the EMA, not just the FDA. In perspective though:
1) This should not be a big surprise to anyone paying attention. The 3 of 4 credible investors here pointed dizziness out as an issue of our drug since the phase 2a.
2) This is a tolerability issue, not a (medical) safety issue, unlike ARIAs.
- These are not the "full" results, but are "fuller". In endpoints, we got dose breakouts, and the pTau and Nf-L data.
- pTau and Nf-L reduction correlation was not stat-sig. Not unexpected, since company withheld it until now. Notably, company provides no explanation (unlike for ADL, and dropouts), which is unfortunate. Means there is none plausible. Was somewhat hoping they potentially could be chalked up to the early disease stage, or recognized high test variances, etc.
- S1 status correlation may be same story, not stat sig. We'll see in Oct.
- As Doc points out, there are results inconsistencies that raise questions about the actual statistical method used in this latest analysis, and the considerable dropout rates raise questions about what populations were used (mITT). The paper will clarify this stuff for us, but this was a medical presentation and should have been made clear in the slides.
Like Doc and Inv2014, I didn't see anything here to change my view on chances of success with EMA for better, or worse. And I'm hoping our peer-review is thorough.