This is indeed a confusing result, as previous large-scale studies have shown an increased risk of type 2 diabetes mellitus patients for developing AD, for example, the recent study by Xu et al., 2009
One important confounding variable is, Are the T2DM patients on drugs to treat diabetes? The study by Sanz et al. does not mention this at all. I can’t access the full paper, but the abstract states:
"Sixty-three participants (10.4 percent) had DM at baseline. In a mixed model adjusted for sex, age, educational level, dementia severity, cholinesterase inhibitor use, and vascular factors (hypertension, atrial fibrillation, coronary heart disease, and hypercholesterolemia), there were no differences between the groups in MMSE baseline scores (–0.75, p = 0.20), but cognitive decline was slower in the group with DM (0.38, p = 0.01)."
It does not appear that there was a control for diabetes treatments. If diabetes has been successfully treated in those patients, they should not have a higher risk of developing AD at all. Some people...