Authors :
Presenting Author: Courtney Berezuk, PhD – Harvard Medical School
Kyle Pellerin, BS – Massachusetts General Hospital; Afareen Jaleel, BS – Massachusetts General Hospital; Lauren Bolden, PhD – Brigham and Women's Hospital; Rani Sarkis, MD – Brigham and Women's Hospital; Alice Lam, MD – Massachusetts General Hospital
Rationale:
An estimated 10 to 22% of patients with Alzheimer’s disease (AD) experience unprovoked seizures, which is associated with more rapid clinical progression. Additionally, subclinical epileptiform discharges have been identified on scalp EEG or MEG in 21 to 38% of patients with AD who do not have seizures. Our goal was to assess whether clinical and subclinical epileptiform activity is associated with lower cognitive performance and/or decline in individuals with AD without prior history of seizures.
Methods: We performed 24-hour ambulatory scalp EEGs in 72 participants with early clinical stages of probable AD (mild cognitive impairment or mild dementia). Of those, 56 participants had no prior history of seizures (ADNoEp) and 16 participants had late-onset seizures attributed to AD (ADEp). Within the ADNoEp group, 45 had no epileptiform discharges (ADNoEp-), and 11 had subclinical epileptiform abnormalities (ADNoEp+) on scalp EEG. Cognitive performance was examined (n=61) using available testing from research and clinical settings, and included: MoCA, Logical Memory, Craft Story, Digits Backward, Trail Making Test A and B, FAS and Animals, and Boston Naming. We compared cognitive performance between ADNoEp- vs ADNoEp+ and ADNoEp- vs ADEp. We examined cross-sectional associations between test performance and group membership using linear regression, adjusting for the time between EEG and the nearest test score and duration of subjective cognitive decline at the time of testing. We examined longitudinal association between test performance and group membership using linear mixed-effect models, with a random effect for participant ID and fixed effects for group, time, group*time, age at cognitive decline, and time between EEG and first test score. Longitudinal neuropsychological data was available for n=10-26 (AD-NoEp-), n=4-8 (AD-NoEp+), and n=0-7 (ADEp), depending on the specific test used.
Results: Mean follow-up time was 4.9 ± 4.2 years. Compared to ADNoEp-, the ADNoEp+ group showed significantly worse immediate recall of the Logical Memory stories (B=-2.3, p=.01) cross-sectionally, as well as faster longitudinal decline (B= -0.02, p=.02). The ADNoEp+ group also showed a trend towards worse delayed recall (B=-1.14, p=.06) cross-sectionally. The ADEp group cross-sectionally showed significantly worse Logical Memory immediate recall (B=-3.4, p=.009) and animal fluency (B=-0.8, p=.03) compared to ADNoEp-, as well as a trend towards worse delayed recall (B = -1.96, p=.053). Longitudinally, the ADEp group also showed significantly faster decline on the Trail Making Test A (p< .001) and B (B=-0.05, p< .001). Data was insufficient to examine longitudinal change in memory for the ADEp group.