Predictors of cognitive impairment in drug-resistant epilepsy: the role of interictal EEG abnormalities

TL;DR

This study explores whether EEG abnormalities can predict cognitive impairment in patients with drug-resistant epilepsy, finding that seizure frequency is a stronger predictor.

Contribution

The study evaluates the incremental predictive value of interictal EEG abnormalities beyond established clinical predictors of cognitive impairment in drug-resistant epilepsy.

Findings

Interictal EEG abnormalities were more prevalent in patients with cognitive impairment.

Seizure frequency was the strongest independent predictor of cognitive impairment.

EEG abnormalities did not provide additional predictive value after adjusting for seizure burden.

Abstract

Cognitive impairment (CI) affects approximately one-third of patients with drug-resistant epilepsy (DRE), underscoring the need for accessible predictors. Interictal electroencephalographic (EEG) abnormalities have been proposed as potential indicators of cognitive dysfunction; however, their independent diagnostic utility is unclear. This study aimed to investigate the association between interictal EEG patterns and CI in adults with DRE, with a specific focus on evaluating their incremental predictive value beyond established clinical predictors.

In this cross-sectional study of 90 adults with DRE were recruited over a six-month period. Participants were stratified into two groups based on their Montreal Cognitive Assessment (MoCA): those with cognitive impairment (Cases; n = 45; MoCA < 26) and those with preserved cognition (Controls; n = 45; MoCA ≥ 26). All participants underwent routine interictal scalp EEG, An EEG recording was classified as abnormal if epileptiform discharges or significant background slowing was identified. The relationships between cognitive status and various clinical variables-including age, monthly seizure frequency and epilepsy type were analyzed using multivariable logistic regression, with expressed as odds ratios alongside their 95% confidence intervals.

The frequency of monthly seizures was significantly higher in the CI group compared to the control group (9.6 ± 2.8 vs. 5.4 ± 2.1 seizures/month, P < 0.001). Interictal EEG abnormalities were also more prevalant in CI group (77.8% vs. 57.8%; OR = 2.56, 95% CI: 1.02–6.41, P = 0.041). However, in the adjusted multivariable model, only seizure frequency reained a signifcant independent association with CI (adjusted OR = 0.46, 95% CI: 0.32–0.65, P < 0.001), indicating that EEG abnormalities did not confer significant additional predictive power after accounting for seizure burden.

Seizure burden emerged as the predominant predictor of CI with DRE, while interictal EEG abnormalities demonstrated a univariate correlation with cogntive status, this association was not independent in the adjusted analysis. EEG findings may still provide contextual or supportive clinical context, emphasize that a comprehensive approach integrating seizure management with cognitive assessments is warranted, rather than relying primarily on interictal EEG for cognitive risk stratification.

Figures