Abstracts

RATIONALE: Video-EEG monitoring is increasingly utilized in the diagnosis of epilepsy. Seizure classification can usually be accomplished by recording a typical seizure or specific interictal epileptiform discharges (IEDs). However there is no detailed data available to guide decisions regarding how long patients should be kept in hospital to obtain diagnostically useful EEG information. METHODS: The time to first IED and the first seizure were analyzed prospectively in 59 patients with epilepsy admitted for seizure classification. EEG recordings were obtained using routine 10-20 scalp electrode placement with sphenoidal electrodes in selected patients. IEDs and seizures were detected by physician review of the EEG and an automated spike and seizure detection program utilizing the Gotman method. RESULTS: 46 patients (78%) had their typical seizures recorded during admission. IEDs were identified in 32 of these patients (70%), the majority of which (84%) occurred within 24 hours. Overall, the mean time to first seizure was 48 hours. However, there was a trend towards a more prolonged time to first seizure in the group without IEDs (62 vs. 42 hours, p=0.14). Only one of the 46 patients (2%) had a seizure recorded beyond 5 days. Thirteen patients had no seizures recorded. Nine of these patients had IEDs, 78% of which occurred within 24 hours. Four patients (7%) had no seizures or IEDs detected despite a convincing history of epilepsy and a suggestive outpatient EEG. Medication was completely withdrawn on admission in the same proportion of patients with and without recorded seizures. In addition, the number of patients with sub-therapeutic, therapeutic and toxic drug levels on admission was similar in each group. CONCLUSIONS: Nearly all diagnostically useful EEG information was obtained within 5 days of admission. If patients are going to have IEDs it is highly probable these will be evident within 24 hours. Patients without IEDs are just as liable to have seizures but these are likely to occur later during the admission. These findings have implications for planning duration of stay in the epilepsy-monitoring unit.