| BackgroundAltered mental state is the emergency in neurocritical care unit (NICU). Early and accurate evaluation and prediction of prognosis is the key point. Various imaging modalities, such as computed tomography (CT), magnetic resonance imaging (MRI), digital subtraction angiography (DSA) are often applied to diagnose patients with altered mental state. However, sometimes patients can not thansfer to imaging for disease severity. Also, imaging can’t be used as a bedside complement continuously. Electroencephalography (EEG) is one of the most useful methods to evaluate a patient’s state of consciousness. EEG allows an immediate, inexpensive, and non-invasive bedside assessment of the brain function condition. EEG has been widely used in many NICU and emergency rooms.EEG grading scales can classify different EEG types. By this way, physicians can know more direct and convenient about medical information of patients, then carry out the treatment plan as soon as possible. Synek and Young classification are the most practical and typical and have been widely used by physicians. However, each classification has some defects. Synek classification had designed for patients who resulted in diffuse cerebral anoxia from suffered cardiac arrest and excluded cases who presented focal lesions such as cerebral infarction or cerebral hemorrhage. Because Young classification defines8/8>50%as grade I, many EEG types have categorized into this grade, then Young classification has lacked increasing abnormality process. Also, triphasic wave has rare occurred in patients from NICU. Meanwhile, two classifications have not included RAWOD or RAWCS pattern designed for acute massive cerebral infarction. Wang xiao-mei et al built a revised EEG classification for patients with severe cerebrovascular diseases. In this classification, slow wave activity was divided into focal and diffuse, and included RAWOD or RAWCS pattern. However, EEG reactivity was not fully used and spindle wave coma was rare occurred in NICU.Electroencephalogram reactivity is a strong positive prognostic factor for consciousness recovery. A study reported that96%patients with preserved EEG reactivity had a globally good outcome,93%of the patients in whom EEG reactivity was absent had a bad outcome. EEG reactivity is an excellent outcome predictor with a high specificity of94%and sensitivity of90%, and was superior to the Glasgow Coma Scale and central conduction time with evoked responses.The purposes of our study were:1. To compare the predicting efficacy of different EEG classification (Synek, Young and Wang xiao-mei) in patients with altered mental state.2. To evaluate the predicting efficacy of revised EEG classification in the prospective study.Subjects and methodsThere were three parts in this study.First, To compare the predicting efficacy of different EEG classification (Synek, Young and Wang xiao-mei) in patients with altered mental state.A study was carried out at the NICU of Nanfang hospital, Southern Medical University of China. March2009and March2011. Inclusion criteria included:1) Adult patients, age range from18-90years.2) patients who had EEG record in the first three day after admitting NICU.3) altered consciousness during the EEG test.4) CT scan or MRI was performed to confirm the lesions, brain stem diseases or cerebellum diseases or diffuse diseases.5) All patients received standard medical therapy. Exclusion criteria included:1) age<18or>90.2) patients receiving antiepileptic or sedative drugs within the24h before EEG examination.3) during therapeutic hypothermia or temperature<32℃.4) patients died from factors of other systems but neurology system, hearing impairment, and seriously failure of liver, renal, respiratory and other systems when admission.5) lost to follow-up6) apparent interference artifact. The following information was collected for all the patients within24h after admission:basic demographic data (age, gender), diagnosis, GCS score, treatment, history and so on.All EEG examinations were performed using the same NicoletOne EEG monitor. EEG test was performed when patients were quiet. Each EEG record was for>20minutes. Silver-chloride EEG electrodes were applied to the scalp according to the international10-20system of electrode placement. EEG reactivity was tested using external auditory stimulation as a sudden loud noise (loud click) and painful stimulation (nail bed). Reactive electroencephalographic pattern was defined as a change in the frequency or amplitude of the background activity. Using three EEG classification (Synek, Young and Wang xiao-mei) to judge the first EEG result of patients.Patients were followed up at3months after onset. The outcome for survivors was assessed using Modified Rankin Scores.SPSS13.0software was used for analysis. Means and standard deviations were calculated for measurement data. Differences between groups were assessed using the Wilcoxon test. Spearman correlation analysis was performed on the relationship between EEG classification and mRS. Statistical significance was taken as P<0.05.Second, to build the revised EEG classification.Using the same data from the first study to evalute the relationship between different variables of EEG and prognosis.SPSS13.0software was used for analysis. Differences between variables of EEG and prognosis were assessed using the Wilcoxon test. Spearman correlation analysis was performed on the relationship between revised EEG classification and mRS. Statistical significance was taken as P<0.05.Third, to evaluate the predicting efficacy of the revised EEG classification in the prospective study.A study was carried out at the NICU of Nanfang hospital, Southern Medical University of China. Inclusion criteria included:1) Adult patients, age range from18-90years.2) CT scan or MRI was performed to confirm the lesions, brain stem diseases or cerebellum diseases or diffuse diseases.3) altered consciousness during the EEG test.4) patients died from factors of other systems but neurology system, hearing impairment, and seriously failure of liver, renal, respiratory and other systems when admission.65patients were included in the Janurary2012and March2012. The following information was collected for all the patients after admission:basic demographic data (age, gender), diagnosis, GCS score, treatment, history and EEG and so on. Patients were followed up at3months after onset. The outcome for survivors was assessed using Modified Rankin Scores.60patients were included into the final analysis.SPSS13.0software was used for analysis. Means and standard deviations were calculated for measurement data. The performances of the revised EEG classification were presented as calibration and discrimination, which were measured by the Hosmer-Lemeshow goodness-of-fit test and the area under the receiver operating characteristic (ROC) curve, respectively.ResultsFirst, to compare the predicting efficacy of different EEG classification (Synek, Young and Wang xiao-mei).82patients were included into the final analysis.14patients with drowsiness;27patients with stupor;41patients in coma.51.2%patients were cerebral infarction and35.4%were cerebral hemorrhage.44(53.7%) patients were survival and38(46.3%) patients were died.There were42(51.2%) patients could not use Young classification and20(24.4%) patients could not use Synek classification. All EEG can used Wang xiao-mei classification. There was a significant difference between the survival and death groups tested by Synek classification (P<0.05), so did Young and Wang xiao-mei classification (P<0.05). Both Synek and Wang xiao-mei classification were well-correlated with MRS (P<0.05), and the Spearman correlation coefficient were0.371and0.419, relatively. However, the Young classification was not well-correlated with MRS (P=0.792). Second, to build the revised EEG classification.29.2%patients with preserved EEG reactivity died,82.4%of the patients in whom EEG reactivity was absent died. There was a significant difference between the survival and death groups tested by EEG reactivity (P<0.05), so did amplitude of EEG. However, there was no significant difference between the survival and death groups tested by background rhythm of EEG. Revised EEG classification is based on three groups, EEG reactivity, EEG non-reactivity and RAWOD or RAWCS. Grade1theta/delta activity (<50%) or amplitude of alpha activity reduced in the involved side. Grade2focal or unilateral theta/delta activity (>=50%). Grade3diffuse theta or delta activity (>=50%). Grade4RAWOD or RAWCS. Grade5diffuse alpha or theta or delta activity with no reactivity; burst suppression pattern; epileptiform discharges in the form of polyspikes or sharp waves; low output EEG (<20uV); Grade6lsoelectric EEG (<10uV). In our study, all patients can use the revised EEG classification.There was a significant difference between the survival and death groups tested by revised EEG classification (P<0.05). Revised classification was well-correlated with MRS (P<0.05), and the Spearman correlation coefficient was0.480. Third, to evaluate the predicting efficacy of the revised EEG classification in the prospective study.60patients were included into the final analysis in the prospective study.5patients with drowsiness;22patients with stupor;33patients in coma.40.0%patients were cerebral infarction and45.0%were cerebral hemorrhage.33(55.0%) patients were survival and27(45.0%) patients were died.The Hosmer-Lemeshow chi-square statistics were0.608(P=0.962, degree of freedom=4) for revised EEG classification. The AUROCC of revised EEG classification was0.769, the cutoff value was3.5, the3-month mortality≤Grade Ⅲ was25.6%, the3-month mortality> Grade Ⅲ was80.9%.Conclusion1Synek and Wang xiao-mei classification were well-correlated with MRS. However, the Young classification was not well-correlated with MRS. 2There were51.2%patients could not use Young classification and24.4%patients could not use Synek classification.3EEG reactivity plays an important role in evaluating the prognosis of patients with supratentorial focal lesion.29.2%patients with preserved EEG reactivity died,82.4%of the patients in whom EEG reactivity was absent died.4Revised EEG classification is based on three groups. Group1EEG reactivity; Group2RAWOD or RAWCS and Group3EEG non-reactivity.5Revised EEG classification was fit to approximately predict3-month mortality in patients with altered mental state. |
PDF Full Text Request