Loss of cortical grey matter is a diagnostic marker of many neurodegenerative diseases, and is a key mediator of cognitive impairment. We postulated that cerebral amyloid angiopathy (CAA), characterised by cortical vascular amyloid deposits, is associated with cortical tissue loss independent of parenchymal Alzheimer’s disease pathology. We tested this hypothesis in patients with hereditary cerebral haemorrhage with amyloidosis–Dutch type (HCHWA-D), a monogenetic disease with minimal or no concomitant Alzheimer’s disease pathology, as well as in patients with sporadic CAA and healthy and Alzheimer’s disease controls.


In this observational case-control study, we included six groups of participants: patients diagnosed with HCHWA-D using genetic testing; healthy controls age-matched to the HCHWA-D group; patients with probable sporadic CAA without dementia; two independent cohorts of healthy controls age-matched to the CAA group; and patients with Alzheimer’s disease age-matched to the CAA group. De-identified (but unmasked) demographic, clinical, radiological, and genetic data were collected at Massachusetts General Hospital (Boston, MA, USA), at Leiden University (Leiden, Netherlands), and at sites contributing to Alzheimer’s Disease Neuroimaging Initiative (ADNI). The primary outcome measure was cortical thickness. The correlations between cortical thickness and structural lesions, and blood-oxygen-level-dependent time-to-peak (BOLD-TTP; a physiological measure of vascular dysfunction) were analysed to understand the potential mechanistic link between vascular amyloid and cortical thickness. The radiological variables of interest were quantified using previously validated computer-assisted tools, and all results were visually reviewed to ensure their accuracy.


Between March 15, 2006, and Dec 1, 2014, we recruited 369 individuals (26 patients with HCHWA-D and 28 age-matched, healthy controls; 63 patients with sporadic CAA without dementia; two healthy control cohorts with 63 and 126 individuals; and 63 patients with Alzheimer’s disease). The 26 patients with HCHWA-D had thinner cortices (2·31 mm [SD 0·18]) than the 28 healthy controls (mean difference −0·112 mm, 95% CI −0·190 to −0·034, p=0·006). The 63 patients with sporadic CAA without dementia had thinner cortices (2·17 mm [SD 0·11]) than the two healthy control cohorts (n=63, mean difference −0·14 mm, 95% CI −0·17 to −0·10, p<0·0001; and n=126, −0·10, −0·13 to −0·06, p<0·0001). All differences remained independent in multivariable analyses. The 63 patients with Alzheimer’s disease displayed more severe atrophy than the patients with sporadic CAA (2·1 mm [SD 0·14], difference 0·07 mm, 95% CI 0·11 to 0·02, p=0·005). We found strong associations between cortical thickness and vascular dysfunction in the patients with HCHWA-D (ρ=–0·58, p=0·003) or sporadic CAA (r=–0·4, p=0·015), but not in controls. Vascular dysfunction was identified as a mediator of the effect of hereditary CAA on cortical atrophy, accounting for 63% of the total effect.


The appearance of cortical thinning in patients with HCHWA-D indicates that vascular amyloid is an independent contributor to cortical atrophy. These results were reproduced in patients with the more common sporadic CAA. Our findings also suggest that CAA-related cortical atrophy is at least partly mediated by vascular dysfunction. Our results also support the view that small vessel diseases such as CAA can cause cortical atrophy even in the absence of Alzheimer’s disease, a conclusion that can help radiologists, neurologists, and other clinicians who diagnose these common geriatric conditions.