A few things didn't add up in the slide deck.
So the raw mean data (slide data) says:
Placebo 29.18 baseline to 33.26 at 48 weeks = 4.08 not 4.11
A273 27.62 baseline to 30.36 at 48 weeks = 2.74 not 2.26
That would be 33% slowing of cognitive change instead of the 45% they report.
Also that would be a treatment mean difference of -1.34 not -1.85 (if true, this would not be p=0.033 but about p=0.15)
Either Missling was sloppy or wrong. Pick one. He'll probably blame it on the third party again.
The slides and/or PR also do not report the co-primary endpoint (ADCS-ADL) numbers just a +3.5 point responder analysis which was not the endpoint.
The safety data looked ok. Some significant tolerability issues, similar to PDD, were seen. According to slide 26, 161 of 170 ITT (or 168 mITT) patients on placebo patients made it to the maintenance phase after titration --- so 7/168= 4% dropped out during initial titration . However only 301 of 338 ITT (or 335 mITT) made it to the maintenance phase after titration --- so 37/335 = 11% dropped out during titration . Once patients got up on maintenance dose, dropouts were less (1 or 0.6% placebo patients and 17 or 5.7% of A273 treated). So total tolerability dropouts were 8/168 = 5% for placebo and 54/338 = 16%. Likely 50 mg would have more dropouts than 30 mg. Anavex will have to figure out how best to titrate for the next phase 3 study.
The serious TEAEs slide 25 were higher in treated patients --- I'm not sure why calculated based on maintenance only (after the titration) instead of including all in the mITT (i.e slide based on only the 462 patients who remained in the study after being titrated). Fortunately the most severe AE (death) was negligible and unrelated. Falls were lower by % in the treated patients so the dizziness and confusion did not lead to more injuries -- though this would be more convincing if based on the n=503 mITT and not the n= 462 patients remaining after titration.
Market Data 
Markets 
AI
