Exactly Laurie. So far what the drug has never "failed" to do is improve test scores.
Company A runs a very small, non controlled trial without mandatory PET scans and finds several patients who improve test scores whereas the entire group is viewed, when divided into subgroups and compared within themselves (based on several variables), as one of the sub groups slowing the rate of decline more than the other as proof the drug does "something".
Company B runs a double blind, placebo controlled exploratory trial and find that the small sub group of 16 memantine patients showed a large statistically significant test score improvement of roughly 6 pts. Viewed a failure by some it is used to design a second FDA controlled confirmatory trial in patients not on memantine. 2nd trial showed nearly identical improvement to the first trial comparing patient's own baseline scores (4.8 pts or roughly 6 pts if placebo had acted similarly to first trial) but in the more moderate, larger subgroup. In other words severe patients not getting the benefit with once test cycle. Once again viewed as "failed" although NIH chooses to back an additional trial, likely to be kicked off mid year based on past timelines.
My guess is if company A finds a subgroup that is more prone to Sig-1 stimulation when another subgroup shows no such activation it will create a "failed" trial in 2.5 years and will have to be re- run as a confirmatory trial but IMO regardless of how the market reacts (probably not positively), as someone who wants to find a beneficial treatment for AD I would view it as positive as additional info could provide for a future treatment for a percentage of sufferers. JMHO
News 
Market Data 
Discover 
