心肺脑复苏后患者振幅整合脑电图对脑功能预后的早期评价

Early evaluation of patients with amplitude-integrated electroencephalogram on brain function prognosis after cardiopulmonary cerebral resuscitation

目的 探讨振幅整合脑电图（aEEG）在早期评价心肺脑复苏（CPCR）患者脑功能预后中的作用.方法 回顾性分析2016年3月至2017年3月河南省人民医院重症医学科（ICU）收治的成人CPCR患者的临床资料,记录其CPCR后住院时间、复苏时长、急性生理学与慢性健康状况评分系统Ⅱ（APACHEⅡ）评分、72 h内aEEG数据及格拉斯哥昏迷评分（GCS）.主要临床结局为发病后3个月的脑功能预后〔格拉斯哥-匹兹堡脑功能评分（CPC）〕.用Spearman秩相关分析aEEG与GCS之间及其与脑功能预后的相关性;用Logistic回归分析aEEG和GCS对脑功能预后的影响;用受试者工作特征曲线（ROC）评价aEEG和GCS对脑功能预后的预测能力.结果 共纳入31例CPCR患者,其中男性18例,女性13例;平均年龄（41.84±16.96）岁,复苏时长（19.42±10.79）min,住院时间（14.84±10.86）d,APACHEⅡ评分（19.29±6.42）分;aEEGⅠ级（正常振幅）7例,Ⅱ级（轻中度异常振幅）13例,Ⅲ级（重度异常振幅）11例;GCSⅠ级（9～14分）7例,Ⅱ级（4～8分）14例,Ⅲ级（3分）10例;存活19例,死亡12例;脑功能预后良好（CPC 1～2分）8例,脑功能预后不良（CPC 3～5分）23例.不同脑功能预后两组间年龄、性别、复苏时长、住院时间及APACHEⅡ评分差异无统计学意义,而aEEG分级和GCS分级差异有统计学意义.Cochran-Armitage趋势检验分析显示,aEEG和GCS分级越高,CPCR患者脑功能预后越差（均P-trend〈0.01）;Spearman秩相关分析显示,aEEG与GCS存在中度相关（r=0.6206,P=0.0003）;aEEG、GCS均与脑功能预后呈正相关（r1=0.7796、P1〈0.0001,r2=0.7021、P2〈0.0001）.Logistic回归分析显示,单因素分析中,aEEG和GCS对早期脑功能预后均有显著影响〔aEEG的优势比（OR）=37.234,95％可信区间（95％CI）=3.168～437.652,P=0.004,GCS的OR=12.333,95％CI=1.992～76.352,P=0.007〕;多因素分析中,仅aEEG对早期脑功能预后有显著影响（OR=26.932,95％CI=1.729～419.471,P=0.019）.ROC曲线分析显示,aEEG预测CPCR患者脑功能预后的ROC曲线下面积（AUC）=0.913,最佳截断值为1.5时,敏感度为95.7％,特异度为75.0％;GCS的AUC=0.851,最佳截断值为1.5时,敏感度为91.3％,特异度为62.5％.结论 aEEG可作为ICU内CPCR患者脑功能预后的早期评价指标,其预测能力优于GCS评分.

Objective To explore the characteristic of early evaluation of patients with amplitude-integrated electroencephalogram （aEEG） on brain function prognosis after cardiopulmonary cerebral resuscitation （CPCR）. Methods A retrospective analysis of the clinical data of patients with adult CPCR in intensive care unit （ICU） of Henan Provincial People＇s Hospital from March 2016 to March 2017 was performed. The length of stay, recovery time, acute physiology and chronic health evaluation Ⅱ （APACHE Ⅱ） score, aEEG and Glasgow coma scale （GCS） within 72 hours were recorded. The main clinical outcome was the prognosis of brain function （Glasgow-Pittsburgh cerebral performance category, CPC） in patients with CPCR after 3 months. Relationship between aEEG and GCS and their correlation with brain function prognosis was analyzed by Spearman rank correlation analysis. The effects of aEEG and GCS on prognosis of brain function were evaluated by Logistic regression analysis. The predictive ability of aEEG and GCS for brain function prognosis was evaluated by receiver operating characteristic （ROC） curve.Results A total of 31 patients with CPCR were enrolled, with 18 males and 13 females; mean age was （41.84±16.96） years old; recovery time average was （19.42±10.79） minutes; the length of stay was （14.84±10.86） days; APACHE Ⅱ score 19.29±6.42; aEEG grade Ⅰ（normal amplitude） in 7 cases, grade Ⅱ （mild to moderate abnormal amplitude） in 13 cases, grade Ⅲ （severe abnormal amplitude） in 11 cases; GCS grade Ⅰ （9-14 scores） in 7 cases, grade Ⅱ （4-8 scores） in 14 cases, grade Ⅲ （3 scores） in 10 cases; 19 survivals, 12 deaths; the prognosis of brain function was good （CPC 1-2） in 8 cases, and the prognosis of brain function was poor （CPC 3-5） in 23 cases. There was no significant difference in age, gender, recovery time, length of stay and APACHE Ⅱ score between two groups with different brain function prognosis, while aEEG grade and GCS grade were significantly different. Cochran-Armitage trend test showed that the higher the grade of aEEG and GCS, the worse the prognosis of CPCR patients （bothP-trend 〈 0.01）. With the increase in GCS classification, the classification of aEEG was also increasing （r = 0.6206,P = 0.0003）. Both aEEG and GCS were positively correlated with the prognosis of brain function （r1 = 0.7796,P1 〈 0.0001;r2 = 0.7021,P2 〈 0.0001）. Univariate Logistic regression analysis showed that aEEG and GCS had significant effect on early brain function prognosis [aEEG： odds ratio （OR） = 37.234, 95%confidence interval （95%CI） = 3.168-437.652,P = 0.004, GCS：OR = 12.333, 95%CI = 1.992-76.352,P = 0.007]; after adjusting for aEEG and GCS, only aEEG had significant effect on the early prognosis of brain function （OR = 26.932, 95%CI = 1.729-419.471,P = 0.019）. The ROC curve analysis showed that in the evaluation of the prognosis of CPCR patients with brain function, the area under ROC curve （AUC） of aEEG was 0.913, when the cut-off value of aEEG was 1.5, the sensitivity was 95.7% and the specificity was 75.0%. The AUC of GCS was 0.851, the best cut-off value was 1.5, the sensitivity was 91.3% and the specificity was 62.5%.Conclusion aEEG and GCS scores have a good correlation in the evaluation of brain function prognosis in patients with CPCR, the accuracy of aEEG in the early evaluation of the prognosis of patients with CPCR is higher than the GCS score.