Self-reported seizure frequency and time to first event in the seizure monitoring unit

Purpose: To compare seizure frequency reported in the clinic with time to first diagnostic event during video-EEG monitoring. The effect of the artificial environment of the monitoring unit on self-reported seizure frequency was explored. Methods: The 155 consecutive patients were seen in the Washington University Epilepsy Center and subsequently underwent video-EEG monitoring during 2001. Of these, 112 had a diagnostic event during monitoring; 31 left without having a definite event; and 12 could not provide an estimate of seizure frequency in the clinic. The time to first event was compared with self-reported seizure frequency. The patients were then divided into three equal groups (tertiles) based on mean seizure frequency, and time to first seizure was compared between groups. Then the numbers of patients staying > 7 days without ever having an event were compared between the low and high seizure-frequency groups. Finally, Kaplan-Meier survival curves were calculated. Results: No correlation was found between self-reported seizure rate and time to diagnostic event (r= 0.18; p = 0.06). Time to first event was 2.8 days in the low seizure-frequency group (mean, 2.2/month), 2.1 days in the medium (mean, 8.8/month), and 2.1 days in the high (mean, 24.1/month) groups, which were not significantly different (p = 0.19). Of patients in the low-frequency group, 79% had an event within 7 days. Conclusions: In the artificial environment of the monitoring unit, self-reported outpatient seizure frequency is not an accurate predictor of duration of video-EEG monitoring required to make a definitive classification of clinical events and should not contribute to the decision as to whether to refer a patient for monitoring.