Abstracts

To establish the number of monitoring days needed for event characterization in patients admitted for video EEG monitoring (VEM), and to determine the proportions who were ultimately diagnosed with psychogenic non-epileptic events (PNE) and true epileptic seizures (ES).
A retrospective chart review was performed on 199 consecutive admissions for behavioral event characterization to the Epilepsy Monitoring Unit at Duke University Medical Center (DUMC). The diagnosis of epileptic seizure, psychogenic non-epileptic event, neither or both was determined from the discharge summary. Kaplan-Meier survival curves were performed to determine the number monitoring days before the first behavioral event occurred.
Of the 199 patients admitted for VEM, 83.9T (n=167) had a clinical event during admission and a definitive diagnosis was made in 75.9% (n=151). Of patients who had clinical events, 58.9% (n=96) had their first event on admission day 1 and 28.8 (n=47) had their first event on day 2. Overall, 88% of patients who had an event did so during the first two days of admission. There was no significant diffference in time to first event between patients with ES and those with PNE (Spearman r = -0.06, 95% CI -0.21 tp 0.10). Factors associated with ES included an abnormal baseline EEG (p[lt]0.001), an abnormal brain MRI (p=0.007), and a patient history of events lasting less than 1 minute (p=0.11). Pre-monitoring factors that were not associated with ES included the presence of an abnormality on neurologic examination, a history of head trauma, and a family history of seizure disorder.
VEM differentiated between epileptic seizures and psychogenic non-epileptic events in the majority of patients selected for the procedure. There was no difference between time to first event for those demonstrating ES vs. those demonstrating PNE. The majority of behavioral events were characterized within 2 days of admission and prolonged monitoring beyond 2 days was associated with a low subsequent diagnostic yield.