Diabetes is related to cerebral infarction but not to AD pathology in older persons

To examine the potential relation of diabetes to common neuropathologic causes of dementia, cerebral infarction and Alzheimer disease (AD) neuropathology.

Subjects were 233 older Catholic clergy in the Religious Orders Study, who underwent detailed annual evaluations, including neuropsychological testing, and brain autopsy at time of death (mean age 86 years, 45% men). Diabetes was identified by annual direct medication inspection and history. Cognitive function proximate to death was summarized into five cognitive domains, based on 19 neuropsychological tests. Macroscopic cerebral infarctions were recorded from 1 cm coronal slabs. Neuritic plaques, diffuse plaques, and neurofibrillary tangles were counted in Bielschowsky silver-stained sections and summarized to yield composite measures of neuritic plaques, diffuse plaques, tangles, and overall AD pathology. We also used immunohistochemistry with antibodies to amyloid-beta and PHF-tau to obtain quantitative measures of amyloid burden and neurofibrillary tangle density. Multiple logistic and linear regression analyses were used to examine the relation of diabetes to cerebral infarctions and AD pathology, controlling for age, sex, and education.

AD pathology was related to all five cognitive domains (p < 0.01) and infarctions were related to perceptual speed (p < 0.001). Diabetes (present in 15% subjects) was associated with an increased odds of infarction (OR = 2.47, 95% CI: 1.16, 5.24). Diabetes was not related to global AD pathology score, or to specific measures of neuritic plaques, diffuse plaques or tangles, or to amyloid burden or tangle density.

We found a relation between diabetes and cerebral infarction but not between diabetes and Alzheimer disease pathology in older persons.