• Mengdi Jiang, Yingying Su, Gang Liu, Huijin Huang, and Fei Tian.
• Department of Neurology, Xuanwu Hospital, Capital Medical University, Beijing, China.
• Resuscitation. 2020 Jun 1; 151: 33-38.

ObjectiveTo explore the value of electroencephalogram (EEG) pattern in predicting awakening of comatose patients after cardiopulmonary resuscitation (CPR).MethodsA retrospective cohort study was conducted on comatose patients after CPR in the neuro-critical care unit of Xuanwu Hospital, Capital Medical University, from 2002 to 2018. The included patients received clinical evaluation, and the Glasgow coma scale (GCS) score was recorded. Bedside EEG monitoring was performed for visual grading and reactivity detection. The 3-month prognostic assessment was performed using the Glasgow outcome scale (GOS). The patients were dichotomized into the awakening group (GOS 3-5) and the unawakening group (GOS 1-2).ResultsA total of 160 patients were included. There was no significant difference in the baseline data between the two groups except that the GCS score of the awakening group was higher (P = 0.000). Different EEG patterns were used to predict awakening from coma. As a result, the slow wave pattern showed the highest accuracy (73.1%, 95% CI: 0.66-0.79), and the sensitivity and specificity reached 61.3% (95% CI: 0.48-0.73) and 80.6% (95% CI: 0.71-0.88), respectively. Compared with EEG reactivity, the slow wave pattern was more sensitive in predicting awakening. It was also more specific than GCS in predicting awakening. The slow wave pattern within different time frame after coma was used to predict the prognosis of awakening, suggesting that the accuracy (100%, 95% CI: 0.75-1.00), sensitivity (100%, 95% CI: 0.46-1.00), and specificity (100%, 95% CI: 0.63-1.00) of predicting awakening within 8-14 days were the highest.ConclusionsThe slow wave pattern of EEG had a good predictive value for awakening in comatose patients after CPR, and the highest accuracy occurred within 8-14 days from coma.Copyright © 2019. Published by Elsevier B.V.

12 keywords...

hide…