Abstracts

Rationale: Continuous EEG (cEEG) monitoring is utilized as a tool for assessment of cerebral function in critically ill patients and it is well established that these patients are at risk for secondary injuries such as seizures. However, many practical questions remain unanswered regarding the utility of cEEG. Periodic patterns are commonly seen in comatose patients, yet their clinical significance is debated. The purpose of this study is to determine which EEG patterns are associated with the occurrence of seizures and poor outcome, as well as the duration of monitoring required to detect significant EEG patterns. Methods: Retrospective study of critically ill patients undergoing cEEG at Emory University Hospital from 1/2008 to 1/2009. Patients undergoing EEG monitoring for treatment of status epilepticus were excluded. Baseline demographics, admission diagnosis, anti-epileptic drugs, and EEG findings were recorded. Outcome at time of discharge was assessed with the Glasgow Outcome Score (GOS). Clinical outcomes post-discharge were recorded from hospital and clinic notes. EEG patterns were characterized as sharps, Periodic Lateralized Epileptiform Discharges, (PLEDs), Bilateral Periodic Lateralized Epileptiform Discharges, (BiPLEDs), Generalized Periodic Epileptiform Discharges including triphasic waves (GPEDs), Frontal Intermittent Rhythmic Delta Activity, (FIRDA), and Generalized Intermittent Rhythmic Delta Activity, (GIRDA). The association between periodic patterns and seizures was determined using the Fisher exact test. The association between EEG patterns and GOS at discharge was determined using the Wilcoxon rank-sum test. Results: Total number of patients was 406 with a mean age of 55 years. Common diagnosis at time of admission included subarachnoid hemorrhage 114 (28%), unexplained loss of consciousness 105 (26%) and intraparenchymal hemorrhage 64 (16%) followed by brain tumor 35 (9%) and ischemic stroke 29 (7%). A total of 53 (13%) had clinical or subclinical seizures however, prior to cEEG monitoring, 387 (95%) were on an anti-epileptic drug. The mean duration of recording was 3 (+/-) 3.7 days. The median time to first periodic pattern was 2 days (1.0-3.0 days, 95%CI). We were unable to calculate the median time to first seizure due to the small number of incidences. Interictal Epileptiform Discharges (IED) and Periodic Epileptiform Discharges (PED) both demonstrated significant association to seizures; Table 1. PEDs also demonstrated a significant correlation to GOS <3; Table 2. Conclusions: Our study supports the previously documented existence of seizures in critically ill patients. We found a significant association between epileptiform discharges, specifically IEDs and PLEDs, and the occurrence of seizures. There was a lack of clinical data to support the association of seizure with rhythmic activity. In regards to outcome, there was a significant association between GOS <3 and PEDs. These findings will aid in prognostic decisions and issues regarding prophylactic treatment with anti-epileptic medications.