Abstract
Little is known about whether clinical, radiological or neuropathological features are associated with cognitive impairment before intracerebral haemorrhage. We conducted a community-based cohort study of 125 adults with intracerebral haemorrhage (lobar n = 71, non-lobar n = 54) with consent to brain autopsy. We compared small vessel disease biomarkers on diagnostic CT head and neuropathological findings including neurofibrillary tangles and amyloid plaques in adults without cognitive impairment versus cognitive impairment without dementia versus dementia before intracerebral haemorrhage, stratified by lobar and non-lobar intracerebral haemorrhage. In non-lobar intracerebral haemorrhage, severe cortical atrophy was less common in those without cognitive impairment (8/36, 22%) and cognitive impairment without dementia (0/9, 0%) versus dementia (5/9, 56%); P = 0.008. Irrespective of intracerebral haemorrhage location, adults without cognitive impairment had milder neurofibrillary tangle pathology measured by median Braak stage (lobar intracerebral haemorrhage: no cognitive impairment 2 [interquartile range, 2-3] versus cognitive impairment without dementia 4 [2-6] versus dementia 5.5 [4-6]; P = 0.004; non-lobar intracerebral haemorrhage: no cognitive impairment 2 [1-2] versus cognitive impairment without dementia 2 [1-2] versus dementia 5 [3-6]; P < 0.001). Irrespective of intracerebral haemorrhage location, adults without cognitive impairment had milder amyloid plaque pathology measured by median Thal stage (lobar intracerebral haemorrhage: no cognitive impairment 2 [1-2] versus cognitive impairment without dementia 2 [2-3] versus dementia 2.5 [2-3.5]; P = 0.033; non-lobar intracerebral haemorrhage: no cognitive impairment 1 [0-1] versus cognitive impairment without dementia 0 [0-2] versus dementia 3 [2-3]; P = 0.002). Our findings suggest that irrespective of intracerebral haemorrhage location, adults with cognitive impairment before an intracerebral haemorrhage have more Alzheimer's disease neuropathologic change.