# Pathological insights into cerebral amyloid angiopathy underlying intracerebral haemorrhage: population-based autopsy study

## TL;DR

This study examines how cerebral amyloid angiopathy contributes to brain hemorrhages and finds that amyloid deposits are widespread, not limited to bleeding areas.

## Contribution

The study provides new insights into the diffuse distribution of CAA pathology and evaluates diagnostic accuracy of CAA grading in post-mortem brain samples.

## Key findings

- CAA pathology is diffusely distributed across all cerebral lobes regardless of ICH location.
- APOE ε2/ε4 alleles and higher Thal/Braak stages correlate with increased CAA measures.
- Vonsattel grade ≥1 on cortical biopsy has 100% sensitivity for detecting CAA.

## Abstract

Lobar intracerebral haemorrhage (ICH) is associated with cerebral amyloid angiopathy (CAA) pathology. Uncertainty remains about the mechanisms leading from CAA to ICH. We investigated the distribution and characteristics of CAA, and its clinical and neuropathological associations. Participants underwent research autopsy in the Lothian IntraCerebral Haemorrhage, Pathology, Imaging and Neurological Outcome (LINCHPIN) study. Neuropathologists rated tissue for CAA using standardised consensus criteria, as well as non-amyloid small vessel disease, Thal phase, and Braak stage. We compared the presence, distribution, and severity of CAA among different brain regions, and in the lobe or hemisphere affected by lobar ICH to corresponding contralateral regions. We evaluated the diagnostic accuracy of Vonsattel CAA grade on a post-mortem cortical specimen (simulating surgical biopsy) versus the reference standard of moderate-to-severe parenchymal CAA at autopsy. Among 162 participants, parenchymal CAA, meningeal CAA, and CAA-associated vasculopathy were diffusely distributed among all cerebral lobes irrespective of the ICH location, but capillary CAA showed an occipital predominance. In lobar ICH, all CAA measures did not differ between the ICH lobe or hemisphere and the contralateral unaffected region. CAA measures did not increase with age, but they were higher in carriers of APOE ε2 or ε4 alleles and in individuals with higher Thal phase or Braak stage. Using a rule-out category of Vonsattel grade ≥ 1 to diagnose CAA on a simulated cortical biopsy achieved 100% sensitivity (95%CI 93.4–100), and a rule-in category of Vonsattel grade ≥ 2 had 79.5% specificity (95%CI 63.5–90.7) versus the reference standard. The distribution and severity of parenchymal CAA, meningeal CAA, and CAA-associated vasculopathy are diffuse regardless of ICH location, indicating the need to better understand the factors underlying bleeding in CAA-affected vessels.

The online version contains supplementary material available at 10.1007/s00401-026-02980-0.