I had blogged before about a minority view that Alzheimer's disease begins with circulatory problems, and the neuronal damage comes later. This hypothesis is emphatically a minority view, and has a lot of genetic data (my favorite) to explain away. However-- and this is important-- it is testable: there are very sensitive evaluations of blood flow in the brain (not least fMRI!), and prospective studies are presumably underway.
What is abundantly clear, though, is once the disease starts, that both neuronal and vascular function are compromised. Today's EurekaNet lists a story to appear in Radiology comparing both vascular flow and indicators of brain structural damage between patients with dementia (either Parkinson's or Alzheimer's) and age-matched controls with normal cognitive function. They found that total brain volume was comparable between patients and normal people, but vascular flow was strongly reduced. Thus the flow density is reduced in the persons with disease. Measures of brain structural damage as expected also showed the correlates of the diseases.
The conclusion is that vascular and neuronal measures both show damage in people with sporadic Alzheimer's even at an early stage of diagnosis. What remains critical-and with enough time, can be done-- is to ask which system is damaged first.