The association of seizure control with neuropathology in dementia

Seizures in people with dementia (PWD) are associated with faster cognitive decline and worse clinical outcomes. However, the relationship between ongoing seizure activity and post-mortem neuropathology in PWD remains unexplored. We compared post-mortem findings in PWD with active, remote and no seizures using multicentre data from 39 Alzheimer's Disease Research Centres from 2005 to 2021. PWD were grouped by seizure status into active (seizures over the preceding 1 year), remote (prior seizures but none in the preceding 1 year) and no seizures (controls). Baseline demographics, cognition, mortality and post-mortem findings of primary and contributing (co-pathologies) Alzheimer's disease (AD), frontotemporal lobar degeneration, Lewy body, vascular pathologies and neurodegeneration were compared among the groups using Pearson's χ2 test, Fisher's exact test, t-test and ANOVA. Of 10 474 deceased PWD, active seizure participants suffered the highest mortality among the groups (proportion deceased among the groups: active = 56%, remote = 35%, controls = 34%, P < 0.001). Among 6085 (58.1% of deceased) who underwent autopsy, 294 had active, 151 had remote and 5640 had no seizures. PWD and active seizures died at a younger age (active = 75.8, remote = 77.9, controls = 80.8 years, P < 0.001) and had more severe dementia (Clinical Dementia Rating score: active = 2.36, remote = 1.90, controls = 1.69, P < 0.001). In post hoc analyses, those with primary post-mortem diagnosis of AD with active seizures had more severe and later stages of AD pathology and ATN (amyloid, tau and neurodegeneration), as evidenced by Braak stage for neurofibrillary (tau) degeneration and Consortium to Establish a Registry for Alzheimer's Disease (CERAD) score density of neuritic (amyloid) plaques, than remote seizure participants and controls. Active seizure participants had more neurodegeneration, evidenced by cerebral atrophy, hippocampal atrophy and locus coeruleus hypopigmentation, than controls. Among participants with primary post-mortem diagnosis of non-AD, in post hoc analyses, active seizure participants had worse AD co-pathology, evidenced by higher Braak stages than remote seizures and controls and a higher Thal phase of beta-amyloid plaques than remote seizure participants. Neurodegeneration (cerebral/hippocampal atrophy) and locus coeruleus hypopigmentation were comparable among the groups. In both primary post-mortem AD and non-AD diagnoses, frontotemporal lobar degeneration (co)pathology was less prevalent among active seizure participants than controls, while vascular pathology, Circle of Willis atherosclerosis, Lewy body pathology, lobar atrophy and substantia nigra hypopigmentation were comparable among the three groups. This study shows that active seizures, compared with remote seizures, are associated with earlier death and post-mortem evidence of more severe ATN pathology. Active seizures are associated with more advanced AD pathology in AD and worse AD co-pathology in non-AD dementias. Therefore, clinicians should be vigilant in detecting ongoing seizures, because this could reflect a worse prognosis in PWD.