The inclusion criteria does require imaging to confirm probable Alzheimer's:
"Neuroimaging (computerized tomography (CT) or Magnetic Resonance Imaging (MRI)) within the last 24 months consistent with a diagnosis of probable Alzheimer's disease (AD)"
My educated guess would be that the percentage of misdiagnosed would be lowered greatly by the advanced stage of the disease. Supposedly the mean number of years since first diagnosis in our trial was roughly 5 years. Also one of the reasons a placebo group of similar diagnosis length is vital I suppose.