Abstracts

Individuals contemplating admission to an epilepsy monitoring unit wish an estimate of the time they should plan to stay and the likelihood that the stay will lead to a diagnosis. We sought to provide data that will aid in answering these questions. We reviewed all patients 18 years or older who were referred for diagnosis of seizures and seizure-like behavior in the last five years. From admission history and descriptions of witnesses, we defined a particular seizure pattern or behavior pattern as the event we were asked to assess. We call this the [ldquo]target event[rdquo]. Most patients had all anticonvulsant medications discontinued beginning with the morning dose on the day of admission. Some did not take medications with long half-lives for a longer period before admission. Sometimes, we are asked to identify the nature of a new pattern that occurs with patients on medications, and these patients continue their home medications. On occasion, we found that the [ldquo]target event[rdquo] was non-epileptic, on other occasions it was epileptic. Some patients had additional [ldquo]non-target[rdquo] events. We considered an admission as [ldquo]not successful[rdquo] when we did not record the event that represented the current chief complaint and/or the event that the patient[apos]s physician needed clarified. We found 180 men and 352 women who were monitored as described above. We observed [quot]target events[quot] in 316 of them. Duration of monitoring was variable. Thus, we report the patients having their first [quot]target event[quot] in a day as a percentage of patients who entered that day without having one in an earlier session. We found that 27.8% had a [quot]target event[quot] on the first day, 13% within the first six hours. The second day, 25.7% had their first event; the third day, 18.6%; the fourth day, 17.1% and 10.0%on the fifth day. Although we recorded only twenty patients yet to have an event on the sixth day, two of them had their first target event on that day. A single day of monitoring will only lead to a diagnosis in about one quarter of admitted adult patients. Three days of monitoring will provide at least one diagnostic event in about one half of the patients. If we recorded no [quot]target events[quot] after four days of monitoring, we have a chance of recording such data in only 10% of these remaining patients. This suggests that four days of monitoring may be a point of diminishing returns.