Abstracts

Rationale: Continuous EEG monitoring is frequently used to diagnose and quantify the severity of cerebral dysfunction but its clinical significance is poorly understood. Standardized critical care EEG terminology exists but the prognostic significance of individual electrographic patterns in non-cardiac arrest patients is unknown. We aim to evaluate electrographic features (based on standardized critical care EEG terminology) that predict poor clinical outcomes in patients diagnosed with cerebral dysfunction unrelated to cardiac arrest.  Methods: Consecutive patients diagnosed with cerebral dysfunction on continuous EEG monitoring at our institution were prospectively studied during a two-month period. Patients with cardiac arrest were excluded. Electrographic predictors included burst suppression, continuity, anterior-posterior gradient (APG), posterior dominant rhythm (PDR), variability, focal slowing, voltage asymmetry, frequency asymmetry, stage II sleep, seizures, GRDA, GPDs, LRDA and LPDS. Statistical analysis was performed using logistic regression. The statistical model was adjusted for clinical features including demographics, etiology, imaging, antiepileptic drugs and sedation. Clinical outcome measures included Glasgow Coma Scale (GCS) at time of EEG and in-hospital mortality. Results: A total of 100 consecutive patients with cerebral dysfunction on continuous EEG monitoring were identified. The mean age was 62.5 years and 55 were women. Median length of stay was 12.5 days and there were 18 in-hospital deaths. Burst suppression (OR 5.6) was associated with poor GCS scores (<8) at time of EEG , independent of etiology and sedation. Absence of APG (HR=6) and frequency asymmetry on EEG (HR =9) were associated with increased in hospital mortality, independent of etiology in this study. Statistical significance was determined at P<0.05.  Conclusions: We prospectively evaluated EEG predictors of poor clinical outcomes in patients with cerebral dysfunction unrelated to cardiac arrest. We found burst suppression was independently associated with poor mental status at time of EEG study. Absent APG and frequency asymmetry were independently associated with increased in-hospital mortality.  Importantly, other EEG features showed no significant associations with poor clinical outcomes independent of etiology and sedation use in this prospective study. EEG may have a limited prognostic role in patients with cerebral dysfunction unrelated to cardiac arrest. Certain electrographic features on EEG may independently predict poor outcomes. Further studies are needed to validate if grouping certain predictors in the form of a standardized classification system may be beneficial in terms of predicting the severity of cerebral dysfunction unrelated to cardiac arrest. Funding: None