Cerebral Amyloid Angiopathy in Alzheimer disease: Diagnostic Implications of Boston Criteria 2.0 and Associations with CSF Biomarkers

Background. Cerabral amyloid angiopathy (CAA) frequently co-occurs with Alzheimer’s disease (AD) and is a major risk factor for amyloid-related imaging abnormalities. However, its prevalence according to Boston criteria 2.0 (v2.0) and associations with CSF biomarkers remain incompletely explored. Methods. We cross-sectionally evaluated 189 cognitively impaired amyloid-positive AD patients with available MRI and CSF data. Hemorrhagic markers (HMs) and non-hemorrhagic markers (nHMs) of CAA were assessed. Patients were classified as absent, possible, or probable CAA according to Boston criteria 1.0, 1.5, and 2.0. Bayesian regression evaluated predicted probabilities of CAA and associations between CSF Aβ40, Aβ42, total tau and p-tau181 levels, CAA status and MRI markers. Results. Median age was 74 years and 57.7% were female. According to Boston criteria v2.0, possible and probable CAA were identified in 59.3% and 28.6% of patients. Predicted probabilities of probable and possible CAA increased by 8.7% and 45.3%, respectively, when transitioning from v1.5 to v2.0. Under v2.0, probable CAA, but not probable plus possible CAA, was associated with lower CSF Aβ42 (−13.9%; 95%CrI, −23.8 to −2.6) and Aβ40 (−11.6%; 95%CrI, −21.9 to 0.9). Consistently, HMs were associated with lower Aβ40 (−11.9%; 95%CrI, −22.4 to −0.2) and Aβ42 (−14.1%; 95%CrI, −24.1 to −3.1), whereas nHMs were not. Conclusions. Boston criteria v2.0 identified probable CAA in approximately 30% of AD patients and possible CAA in nearly 60%, with potential implications for anti-amyloid therapy eligibility. This increased detection was driven by the inclusion of nHMs. Only probable CAA was associated with reduced CSF Aβ levels, which were specifically linked to HMs.