Methods: Subjects enrolled in two vascular-focused studies of cognition underwent PIB-PET and MRI (n=23, mean age 79.8±7.8 years, 91% with vascular risk factors). PIB DVR images (Logan, cerebellar reference) were classified as positive (PIB+) or negative (PIB-) based on visual reads and MRI scans were rated for the presence of infarcts, all blinded to clinical data. Subjects were classified as having cerebrovascular disease (CVD+) if there was a clinical history of stroke/TIA or an infarct on MRI. Patients with a CDR≥0.5 were considered cognitively impaired (n=14/23, median CDR=0.5).

Results: 10/23 subjects had PIB+ scans (mean PIB Index 1.31±0.30, versus 0.95±0.05 in PIB- scans, p<0.001). Highest regional uptake in PIB+ scans was in the precuneus (mean DVR 1.40±0.35). 16/23 subjects were CVD+ (50% with lacunar infarcts, 25% with cortical infarcts and 25% with both). Cognitive impairment was found in 1/6 CVD-/PIB-, 5/7 CVD+/PIB-, 0/1 CVD-/PIB+ and 8/9 CVD+/PIB+ subjects (χ2=9.77, p=0.02). In a logistic regression that included cognitive impairment as the outcome and PIB, CVD, age and education as predictors, cognitive impairment was associated with CVD+ (odds ratio=32, 95% confidence interval 1.5-694, p=0.03) but not with PIB+ status (p=0.78). A trend for an association between CVD+ and cognitive impairment remained after excluding subjects with cortical infarcts (p=0.08).

Conclusions: In this small, mildly impaired, elderly cohort enriched for vascular disease, CVD was more strongly linked to cognition than was amyloid. CVD is often viewed as a secondary factor in cognitive impairment, but this data suggests it is not infrequently the primary cause of mild impairment. Further studies are needed to determine the relative contributions of amyloid and CVD to changes in cognition and brain structure, and function.

Potential conflicts of interest: Nothing to report.

Funding: Supported by the National Institute on Aging P01-AG12435, R01-AG031563, R01-AG10129, K23-AG031861; Alzheimer’s Association ZEN-08-87090.