Survival analysis of electroencephalogram monitoring patients in intensive care unit

Objective To observe and evaluate the predictive value of electroencephalogram (EEG) abnormalities of the EEG monitoring of patients with brain dysfunction in the intensive care unit(ICU). Methods Total of 58 cases with brain dysfunction under EEG were collected from the ICU of the XiangYa Hospital, Central South University from January 2014 to December 2015. EEG was performed to monitor those patients and data was collected, analyzed and classified according to both Synek and Young EEG scales to evaluate its predictive value. The statistical analysis was performed with SPSS 23.0 software(MAC, USA)and statistical significant was considered as P<0.05. Numerical values were given as means ± SD and t-test was performed to compare data of different groups. Kaplan-Meier survival estimator was used to draw the survival curve, and the survival analysis was postulated by COX regression analysis. Results Data from 58 patients were collected and classified according to both Synek and Young EEG scales, positive waveforms as periodic discharge or delta-predominant background were found among 50 patients, electrographic seizures were found in 7 patients, patients with EEG abnormality possessed a high level of S100β and showed statistical differences. The 28-day mortality was independently associated with Acute Physiology and Chronic Health Evaluation (APACHE Ⅱ score)(OR: 1.08; 95% CI [1.03 to 1.14])、Synek Grade>2(OR: 0.17; 95% CI [0.03 to 0.80])、electrographic seizures (OR: 23.70; 95% CI [2.02 to 277.73]) and slow rhythm (OR: 8.54; 95% CI [1.72 to 42.32]). Conclusions The 28-day mortality of patients under EEG with brain dysfunction was independently associated with Synek Grade>2, electrographic seizures and slow EEG rhythm. Key words: Electroencephalogram monitoring; Brain dysfunction; Synek grade; Young grade; EEG abnormalities; Survival analysis; Intensive care unit; 28-day mortality