Neurofibrillary tangle distribution in posterior cortical atrophy and typical early-onset Alzheimer's Disease

Background and objectives: Posterior Cortical Atrophy (PCA) is a focal variant of Alzheimer's Disease (AD) characterized by disproportionate and early visuospatial deficits with relative sparing of other cognitive abilities. PCA commonly arises before age 65, consistent with early onset AD (EOAD). Because studies of PCA pathology in patients with detailed cognitive assessment are rare, relationships between severity and distribution of neurofibrillary tangle (NFT) density and visuospatial deficits in PCA are not fully known. Our objectives were to: (1) determine if ratio of NFT density in occipital cortex (OC) versus hippocampus (HC) is higher in PCA than typical EOAD and (2) determine if this ratio correlates with severity of visuospatial deficits in EOAD generally. Methods: We determined density of tau NFT pathology in primary visual cortex (OC), midfrontal cortex (MF), and HC and calculated ratios of these densities for EOAD patients with PCA (n = 12) or non-PCA (n = 33) clinical phenotypes. Groups were compared using linear regression that adjusted for age at death and sex. Correlation between NFT density and visuospatial and other cognitive domain scores were examined using linear regression that adjusted for age at cognitive testing, education, sex and test-death interval. Analyses were repeated with the non-PCA EOAD patients grouped into those with mild (EOAD-Typical, n = 21) or severe (EOAD-Visual, n = 12) visuospatial impairment. Results: PCA patients had lower HC (β = -11.7 ± 5.7, p = 0.045) and a trend toward higher OC (β = 6.9 ± 3.7, p = 0.07) NFT density, and higher OC/HC (β = 0.30 ± 0.14, p = 0.04) and OC/MF (β = 0.56 ± 0.16, p = 0.001) ratios, than typical non-PCA EOAD patients. The EOAD-Visual had higher OC/HC NFT ratios (β = 0.38 ± 0.13, p = 0.008) than the EOAD-Typical non-PCA patients even though they did not meet clinical criteria for PCA. Correlations between OC/HC NFT ratios and visuospatial domain scores were strong (β = -1.29 ± 0.45, p = 0.008) and remained significant when limited to the non-PCA EOAD group (β = -1.22 ± 0.43, p = 0.008). Discussion: PCA is associated with a distribution of NFT pathology (i.e., high OC/HC NFT ratio) that coincides with early predominant visuospatial impairment. Non-PCA EOAD with a memory-predominant presentation and concomitant visuospatial impairment has an NFT distribution profile similar to PCA. NFT pathology in occipital cortex may moderate degree of visuospatial impairment in EOAD regardless of clinical syndrome.