Evaluating and Predicting Diagnostic Yield of Ambulatory EEG for Interictal Epileptiform Discharges and Seizures
Abstract number :
1.237
Submission category :
3. Neurophysiology / 3C. Other Clinical EEG
Year :
2025
Submission ID :
67
Source :
www.aesnet.org
Presentation date :
12/6/2025 12:00:00 AM
Published date :
Authors :
Presenting Author: Nathan Kindja, MD – University of Pittsburgh Department of Neurology
Lavanya Biju, BS – University of Pittsburgh Neurology
James Castellano, MD – University of Pittsburgh Department of Neurology
Alexandra Urban, MD – University of Pittsburgh Medical Center
Anto Bagic, MD, PhD – University of Pittsburgh Department of Neurology
Joanna Fong-Isariyawongse, MD – University of Pittsburgh Department of Neurology
Wesley Kerr, MD, PhD – University of Pittsburgh Department of Neurology
Vijayalakshmi Rajasekaran, MD – University of Pittsburgh Department of Neurology
Rationale: Diagnosing and characterizing epilepsy can be challenging due to the episodic nature of seizures, which also may mimic other conditions. In-hospital electroencephalography (EEG) can address these challenges but is resource-intensive and has limited availability. Ambulatory EEG (aEEG) has emerged as a cost-effective, accessible alternative, enabling prolonged monitoring in outpatient settings, but may have limited technical quality and does not allow for real-time patient testing. To improve the selection of those patients for whom aEEG will improve diagnosis and inform the selection of the duration of aEEG, we evaluated aEEG and clinical characteristics that may influence the likelihood of capturing epileptiform discharges or seizures.
Methods: We evaluated all aEEGs conducted in adults ( >=18 years old) between April 2022 and October 2022 at a comprehensive epilepsy center. We evaluated if the pre-specified length of aEEG and clinical characteristics were associated with identification of epileptiform discharges or seizure capture with logistic regression. We used recursive feature elimination (RFE) to identify the key characteristics associated with capturing these findings.
Results: Of 179 patients who underwent aEEG (average age 44, 58% female), 35% (62/179) of patients had interictal epileptiform discharges, and 6% (11/179) had a seizure during aEEG. Epileptiform discharges were more likely observed when a prior aEEG was abnormal (Odds Ratio [OR] 3.37, p=0.039) or more concurrent antiseizure medications were in use (OR 1.45/medication, p=0.007; Figure 1A). Epileptiform discharges were less likely observed when MRI brain was either normal (OR 0.28, p=0.012) or had nonspecific abnormalities (OR 0.19, p=0.003). A seizure was more likely observed when seizure frequency was higher (OR 1.71 per log seizure/month, p=0.034; Figure 1B), the aEEG was scheduled for 72 hours (OR 26, p=0.003), or epileptiform discharges were captured in a prior epilepsy monitoring unit admission (OR 22, p=0.004). Seizures were less likely to occur with a comorbidity of chronic pain (OR 0.0049, p=0.026) or a prior abnormal aEEG (OR 0.03, p=0.042).
Conclusions: This study demonstrated the diagnostic yield of aEEG for capturing interictal epileptiform discharges and observing seizures. Additionally, we identified clinical characteristics associated with either higher or lower diagnostic yield, which can guide choices of the length of aEEG. If the goal of aEEG was seizure capture, a longer 72-hour study narrowly had the highest yield, although shorter studies may be appropriate especially in people with high seizure frequency or prior EMU admission (Figure 2). On the whole, 48 hours emerged as a reasonable recording duration for seizure capture, with diminishing returns for longer durations. If the goal of aEEG was to identify interictal epileptiform findings, the yield of aEEG was higher in patients with a high suspicion of epilepsy (prior interictals and more antiseizure medications) and, conversely, was lower if neuroimaging did not indicate epilepsy.
Funding: n/a
Neurophysiology