Cerebral amyloid angiopathy and Alzheimer's and related pathologies across APOE genotypes

Cerebral amyloid angiopathy (CAA) is associated with age, apolipoprotein E (APOE) genotype, and a pathologic diagnosis of Alzheimer's disease (AD). Yet, the complete spectrum of CAA presence and severity across age, APOE genotype, AD and AD-related disorders (ADRD) is incompletely reported. Additionally, recent experiments suggest associations between CAA and diffuse plaques. Here, we studied CAA in older adults who were followed during longitudinal studies of ageing. Post-mortem brains were evaluated for the presence and severity of CAA and co-pathologies. AD was defined as intermediate-to-high Alzheimer's disease neuropathologic change (ADNC). Regression models were used to analyse the association of AD-related (neuritic and diffuse plaques and neurofibrillary tangles) and ADRD-related counts with CAA, controlling for risk factors including demographics, AD and APOE ɛ4. The 1938 participants with a mean age-at-death of 89.8 years (standard deviation = 6.6) had no (415, 21.4%), mild (795, 41.0%) or moderate-to-severe (728, 37.6%) CAA. The odds of moderate-to-severe CAA were higher in persons who were older [odds ratio (OR) per 10 years older 1.34, 95% CI 1.22-1.63], APOE ɛ4 allele carriers (OR 3.62, 95% CI 2.90-4.52) or comorbid for AD (OR 4.14, 95% CI 3.28-5.23). Despite a strong association between AD and CAA, 117 of 1216 (9.62%) participants with AD had no CAA, while 108 of 581 (18.59%) participants with moderate-to-severe CAA had no AD (i.e. none-to-low ADNC). However, moderate-to-severe CAA was associated with neuritic plaques (OR 1.27, 95% CI 1.09-1.48) and neurofibrillary tangles (OR 1.52, 95% CI 1.32-1.76). Among participants without AD, the odds of severe CAA were ∼28-fold higher in APOE ɛ2 allele carriers when neuritic plaque and neurofibrillary tangle loads were higher. This unexpected association between CAA severity and combined neuritic plaque and neurofibrillary tangle load was not found in APOE ɛ2 allele carriers when there was AD or in APOE ɛ4 allele carriers with or without AD. ADRD was not related to CAA after controlling for AD and APOE ɛ4. Logistic models using moderate-to-severe CAA as the outcome revealed an interaction between neurofibrillary tangles and neuritic plaques in the entire group (P = 0.047) and in APOE ɛ2 allele carriers (P = 0.039). We conclude that CAA is associated with neuritic plaques and neurofibrillary tangles, and this relationship is markedly enhanced in APOE ɛ2 allele carriers (excluding APOE ɛ4) without AD. These findings indicate that further work on the complex relationships between CAA and AD-related lesions must consider AD and APOE status for a more personalized approach to studying CAA.