Abstracts

Rationale: In a prior study in the Veteran population, only 13% of routine EEG studies detected any abnormal finding and epileptiform discharges were detected in only 3% (Bozorg, Lacayo et al. 2010). Recent reports support using extended outpatient EEG recordings to increase the yield of detecting abnormalities. In one study reviewing EEG in patients with positive findings, about 64% of epileptiform discharges were detected within the first 30 minutes, which increased to 100% after 3-hours of recording (Lee et al. 2013). In patients with suspected epilepsy, detection of habitual events increased from 8% to 15% when the video-EEG recording was continued for an additional 4 hours following the initial routine 20-minute recording (Modur and Rigdon 2008). We sought to determine whether extended 3-4 hour outpatient video-EEG after routine video-EEG increased the diagnostic yield in our Veteran population.Methods: Retrospective review was conducted on a set of 196 outpatient extended video-EEG (eVEEG) studies in Veterans who had previously undergone routine video-EEG (rVEEG) during a prior clinical encounter. Patients were classified by diagnosis into either clinical epilepsy (CE, n=73) or rule out seizures (ROS, n=123) groups based on information on record leading to EEG evaluation. Video-EEG evaluations were then classified as diagnostic if interpretations were suggestive of focal onset epilepsy (FOE), generalized epilepsy (GE), or non-epileptic events (NEE). The diagnostic yields for rVEEG alone, and for eVEEG after non-diagnostic rVEEG, were calculated and compared among the CE and ROS diagnosis groups. We also compared the eVEEG diagnostic rate (yield of eVEEG after a non-diagnostic rVEEG/yield of both video-EEG evaluations) among these two diagnosis groups. P-values were corrected for false discovery rate given the multiple comparisons performed.Results: In the ROS group, eVEEG was only superior to rVEEG in terms of capturing stage II sleep (p<0.001). However, in the CE group, eVEEG was superior to rVEEG for capturing of stage II sleep (p<0.001), and also for detecting overall abnormalities (p=0.017), non-epileptiform abnormalities (p=0.014) and epileptiform abnormalities (p<0.001). The rVEEG was diagnostic in 19/73 (26%) and 17/123 (14%) of the CE and ROS groups, respectively. After a non-diagnostic rVEEG, eVEEG was diagnostic in 29/54 (54%) and 20/106 (19%) of the CE and ROS groups, respectively. The combined diagnostic yield of both rVEEG and eVEEG was 48/73 (66%) and 37/123 (30%) in the CE and ROS groups, respectively. The eVEEG diagnostic rate with 95% confidence interval (CI) was 0.60 [CI: 0.45-0.74] and 0.54 [CI: 0.37-0.71] in the CE and ROS groups, respectively. The diagnostic yield of eVEEG was also more pronounced in the CE group, demonstrating a significantly higher yield than for the ROS group (p<0.001).Conclusions: We conclude that performance of eVEEG after non-diagnostic rVEEG is worthwhile for both patients with clinically diagnosed epilepsy and in patients undergoing workup for ruling out seizures.