Anavex Life Sciences Corp (AVXL)

[B60611]

Computing absolute numbers for ADAS-Cog and ADCS: I don't think you need to do distribution and regression to find the possible combinations of Blarcamesine/Placebo for slides 19 and 20 (-.5 on ADAS-Cog and +3.5 on ADCS-ADL) . Inspired by a post on ST yesterday, which linked the formula for odds ratio https://www.ncbi.nlm.nih.gov/books/NBK431098/ , and also provided his results for ADAS-Cog, I verified his numbers and also did my own for ADCS-ADL. My method was as follows: Begin at 1 with placebo, and use the linked forumula to find the 2 corresponding Blarcamesine combos that sandwich the odds ratio. Then move up to 2 placebo, etc, and perform the same plug and chug computation (you figure out pretty quickly how to increase efficiency), until there was a direct hit (1.839 for ADAS-Cog and 2.67 for ADCS-ADL).

NOTE, AND THIS IS ESSENTIAL: Slides 19 and 20 state they reflect the ITT population. If slide 16 doesn't reflect the ITT population (Blar n=338, Placebo n=170), then my computations are essentially worthless.

The results are as such: There were no direct hits on ADAS-Cog -.5 until 38 placebo, 117 Blarcamesine. I stopped there, as I was more interested in ADCS-ADL given lack of information about reduction of decline, but the original poster said the next hit was at 77 Placebo, 204 Anavex. I think 38/117 is far more likely to be correct. For ADCS-ADL, the first hit was 12 Placebo, 57 Blarcamesine. The next was 19 Placebo, 85 Blarcamesine. I stopped at that point, as I had calculated/verified the "worst case scenarios" (ie the smallest, thus arguably the less compelling numbers) for ADAS-Cog and ADCS-ADL and was ecstatic with my findings. The eyeball test tells me this will easily achieve ALZ approval at some point, and if another P3 is required, Anavex should attract a lucrative partnership with Big Pharma, who will speed up the process up and add significant pull with the FDA. Ink the partnership before arguing for approval/P4 on the current data, and I think it's extremely likely, especially if the non-pooled data is as good as I and many others suspect it will be.

-.5 for ADAS-Cog and +3.5 for ADCS-ADL are somewhat arbitrarily chosen, and I'm sure they aren't chosen to place the effect of the drug in a bad light. That said, I think inferences can be made, especially by those with a lot of knowledge about the tests. My simplistic take is -.5 ADAS-Cog is more reflective of stabalizing the disease, whereas +3.5 ACDS-ADL is more reflective of "super-responders," which explains why there are likely far more of the former (especially Placebo). If our populations are expected to decline +4 or so on the ADAS-Cog/year, and decline -1.2 on ADCS-ADL (these are just numbers I've seen floated around), that simplistic take would make some sense. But even if -.5 improvement on ADAS-Cog is mostly just stabilizing, the stated Mean ADAS-Cog score improvement of -4.03 points (pg 19) for patients taking Blarcamesine is astounding.

TLDR: I can't wait to see the 50 mg breakdowns, they're going to be incredible.