Abstracts

Rationale: Epileptologists  consider an epilepsy monitoring unit (EMU) admission a failure when there is no event or diagnostic pattern recorded. Administrators view EMU admissions requiring longer lengths of stay (LOS) to capture the typical event or diagnostic pattern as inefficient, costly, and negatively impacting hospital scorecard benchmarks. Because of the nature of different  types of EMU admissions, presurgical Phase I and II studies may be held separate from general service LOS hospital statistics, but routine diagnostic EMU admissions are considered with all services. It would be useful to optimize diagnostic potential in multiday EMU routine admissions.  We sought to address the problem of trying to identify common factors associated with EMU admissions that were more frequently nondiagnostic.   Methods: We reviewed 197 individual adult patients admitted for 'nondiagnostic' EMU testing of at least five days duration to the Wake Forest Comprehensive Epilepsy Center EMU between January 2012 and  December 2018. Patients underwent standard paractices of the EMU, with baseline recording including photic stimulation and hyperventilation. Routine daily recording included provocative techniques, including repeated photic stimulation and hyperventilation, sleep deprivation, antiepileptic medication reduction or elimination, exercise, and patient specific triggers. 21 channel video EEG recordings with electrodes placed in accordance with the international 10-20 system and supplemental physiologic electrodes for EKG were reviewed in raw format using standard montages of the American Clinical Neurophysiology Society, with patient alerts also reviewed in detail with continuously recorded video.  Data was also recorded for medications removed, baseline spell frequency at time of admission, and duration spells had been present at time of admission. When prior records were avilable, data was recorded from previous EEGs and EMU studies and whether there was diagnostic information obtained from these studies. Results: We identified 197 patients, age 18-62, who had 5+ days with nondiagnostic EMU admissions during the time period, with 27 of these patients admitted for 7 days.  All had sleep deprivation at least once. 135 were taking antiepileptic drugs on admission; all but 25 had complete removal of their medications during admission.  116 patients were taking a benzodiazepine on admission, only 8 had complete removal of benzodiazepine. Baseline spell frequency was daily in 20, 2-4x per week in 29, weekly in 28, monthly in 53, and less than monthly in 67. Duration of spells was less than 6 months in 36 patients, 6-12 months in 61 patients, 1-5 years in 72, and unknown duration in 28 patients. Of the patients who had records available, there were 177 with normal EEGs and 55 with 2 normal EEGs. There were also 34 who had prior nondiagnostic video EEGs. Conclusions: With current concerns for reductions of LOS, it has become increasingly important for epileptologists to focus on optimizing duration among other aspects of EMU admissions. The goal of obtaining diagnostic information in minimal inpatient time and then affecting a treatment plan safely prior to discharge is a prodigious challenge. Our results indicate that there may be some basic variables which are predictive of failed nonsurgical EMU referrals, however, and careful planning may lead to higher diagnostic yield and avoidance of unneccessary EMU admissions. We found that this group of nondiagnostic patients tended to be treated with benzodiazepines, which were not weaned off during admission. Also low frequency of spells was noted in this series, as 60% of patients had spells no more frequently than once a month. Less frequent spells may be more appropriately managed in the outpatient setting.  Further study is needed to determine if these variables are unique chracteristics of nondiagnostic compared to diagnostic admissions. Funding: No funding