Regarding small vs large trials: Most AD trials (except AVXL's) have indeed been very large -- but that is likely because the expected change in endpoints has been small. AVXL was expecting a larger change -- so, went with a smaller trial to save costs. Ultimately, p-value is what really matters (if safety is not an issue).
In either case, with respect to FDA/US, there may be a tiny change of AA (based on biomarker data) -- but otherwise AVXL needs another trial anyway for FDA approval (this new trial will likely be a large trial -- since the smaller one "failed").