Abstracts

Rationale: Continuous EEG (cEEG) is increasingly used for inpatients with altered level of consciousness in whom the presence of non-convulsive seizures requires assessment. Here, accurate localisation is less important. Hence, simplified, reliable EEG recordings, aided by digital reformatting and remote data transfer may achieve these goals efficiently, and in settings without specialised epilepsy services. Our aim was to assess the accuracy of a simple, abridged montage for this purpose.Methods: We reviewed 20 randomly selected ICU patients undergoing cEEG to assess brain function or the presence of non-convulsive seizures. We created a simplified montage containing only seven long-interelectrode distances (Fp1, Fp2, T3, T4 , O1, O2 and Cz), yielding eight AP-Biploar channels plus three transverse channels (plus EOG and EKG). Two EEGers independently read all recordings using the abridged and the standard montages, blinded to the results of both montages. Amount and location of seizures, epileptiform discharges, and slowing were the main variables compared. Results: Seven patients were female (35%). Mean age was 42.4 years (SD 21.5). Five were excluded; in two O1 and O2 were artifactual, and in 3 data could not be retrieved. The average recording time was 3 days. The main question was presence and amount of seizures. One patient (6.7%) had normal recordings on both montages (non-epileptic seizures), and 14 (93.3%) had abnormal recordings in both montages. In both montages, five patients (33%) had recorded seizures (seizure group), three had non-convulsive status (20%), one had 3 seizures, and one had status myoclonus. Six patients (40%) had diffuse encephalopathy (encephalopathic group) in both montages. Three (20%) had epileptiform activity in addition to slowing (mixed group). The overall accuracy of the abridged montage for presence and laterality of abnormality was 100% in all groups. However, it failed to localise accurately the onset of seizures in one patient and to detect multifocal discharges in two patients. In the mixed group, it detected the presence of spikes and sharp waves but was inaccurate regarding focus of maximum activity. Muscle artefact was less frequently a problem.Conclusions: An abridged, simplified montage can accurately (100%) detect seizure activity, pseudoseizures, epileptiform discharges and slowing activity in ICU. Advantages of the abridged montage inlcude fewer electrodes to apply and maintain, less time (about one third) to apply while using the familiar 10-20 system, and less muscle artefact. Disadvantages include inability to localise accurately and vulnerability to losing one electrode. Although our data require validation, we propose that an abridged montage is sufficient for the purposes of cEEG in the ICU and in similar settings.