Application Of Amplitude-integrated EEG On Evaluating Prognosis And Detecting Epilepsy Of Consciousness Children At PICU

Objectives:To observe the predictive value of the amplitude-integrated EEG in consciousness children of PICU,and compare the performances of aEEG,vEEG,GCS. To study the usefulness of detecting eleptiform discharge by aEEG.To provide a new method of cerebral function monitoring.Methods:According to our inclusion criteria,This study contain 152 critical ill children who had the disturbance of consciousness from January2015 to October 2015.According to our inclusion criteria.All of them accept the monitoring of aEEG and vEEG at least 2 hours. GSC were recorded before the monitoring or at the first 24 hours admitted in PICU.Using the Naqeeb rating system divide aEEG into 3grade:normal,mildly abnormal,severe abnormal.Using the Synek rating system to divide vEEG into 5 grades.The following of outcome reach an end of clinical death or 3 months after leaving hospital.The pediatric cerebral performance category score divide prognosis into 6 grades,they include normal,mild disability, moderate disability,severe disability;coma or vegetative state,brain death.the 1-3 grades represent good outcome(n=85),4-6 grade and clinical death are bad outcome(n=67).χ2 test are used to discuss the relationship between different grads and different outcome.The relationship between aEEG and vEEG,GCS are explored by spearman correction test.Contradistinction of these methods were presented by the area under the receiver operating characteristic(ROC)curve.Concordance of aEEG grade and v EEG, GCS that predicting outcome was studied by kappa coefficient text.There was poor consistency when k is below 0.4.and there was moderate consistency when k is between 0.4 and 0.7.when k is above 0.7,the consistency is good.P<0.05 was considered statistically significant.Using the v EEG as gold standard to analysis the sensitivity and specificity when we use aEEG to detect eleptiform discharge.Results:In 152 cases,71(46.7%) children showed normal aEEG;40( 26.3%) children showed mildly abnormal aEEG,25 of them has the mildly abnormal amplitude,and 15 of them has the normal amplitude but followed by eleptiform discharge;41(27.0%) showed severely abnormal aEEG,30 of them has the severely abnormal amplitude,and 11 of them has the mildly abnormal amplitude but followed by eleptiform discharge.The rate of poor outcome are 4.2%、57.5%、100%.。Different aEEG with difference outcomes of ill children was statistically significant(χ2=100.687P=0.000).Spearman correlation suggested that aEEG has positive relation with vEEG(r=0.618,P=0.000) and negative relation with GCSscore(r=-0.786,P=0.000).The area under the ROC curve of aEEG,vEEG,GCS are 0.939、0.908、0.926、0.932. No significance was found between the area under the ROC curve of these models.Abnormal aEEG has moderate consistency with vEEG(k=0.529)and GCS(k=0.610)when it is used to predict poor outcome.Sensitivity and specificity of aEEG detecting eleptiform discharge were 70.2%,87.8%.Conclusions:1.The results of aEEG can divide into 3 grades:normal,mildly abnormal,severe abnormal aEEG.the rate of poor outcome rise when grade of aEEG is higher.Different aEEG with difference outcomes of ill children was statistically significant.aEEG is useful to evaluate the outcome of consciousness children in PICU.2.Sensitivity and specificity of abnormal aEEG predicting poor outcome were 80.0% 、 95.5%,aEEG has equal efficacy with the other methods.But aEEG is more simple and objective when compared with vEEG and GCS,so aEEG is more suitable to use in PICU.3.The sensitivity and specificity of aEEG detecting eleptiform discharge was 70.2%,87.8%.Primary EEG signal has been compressed,so some eleptiform discharges will be dismissed by aEEG when the abnormal discharge is isolate,low frequency or without apparent change of amplitude.The value of distinguish eleptiform discharge need the help of primary EEG.