Prognostic Value Of Abnormal EEG Changes Induced By Zolpidem In Patients With Chronic Disorders Of Consciousness

【Background】With the widespread promotion of intensive care technology and the gradual completion of the social emergency system,the mortality rate of patients with various types of severe brain injury has decreased significantly.However,due to the lack of effective medical methods to restore their brain function,the number of patients with chronic consciousness disorders(CDOC),including vegetative states(VS)or unresponsive arousal syndrome(UWS)and minimally conscious state(MCS),has increased significantly.The daily care and treatment to maintain the survival of patients with CDOC are expensive and their families of patients usually have to bear the economic and spiritual double blow.Meanwhile,the long-term occupation of public medical resources also causes a great medical burden to the society.Therefore,having a good knowledge about the evolution of patients’ clinical symptoms and brain function status,and determining reliable prognostic markers,will enable clinicians and patients’ families to make appropriate decisions in treatment and care.However,to date,the clinical outcome of CDOC patients has not been determined.The relationship between the clinical evolution of VS / UWS or MCS patients and their etiology,age,clinical diagnosis,and other factors is not well established and can not provide reliable prognostic information to guide clinicians and patients’ families in decision-making.In addition,there are currently few evidence-based guidelines for the treatment in patients with DOC.Overall,patients’ responses to various awakening therapies remained unsatisfactory.Zolpidem is a non-benzodiazepine drug that produces pharmacological effects by selectively activating the central nervous system GABAA receptor ω1 subunit with a half-life of 1-4 hours.In recent years,zolpidem has been reported to have a significant behavioral “wake-up” effect in some patients with CDOC,but no significant effect has been found at the population-wide level of study.The behavioral response rate to zolpidem was 5 %,it is suggested that the effect of treatment is related to individual sensitivity.Previous studies have mostly been based on patients with known behavioral responses to zolpidem.The association between the EEG response of zolpidem and prognosis of CDOC is unknown in patients without behavioral response.【Objectives】1.To observe the long-term clinical outcomes of patients with CDOC and determine the prognostic risk factors;2.To explore the prognostic value of zolpidem-induced EEG reactivity in patients with CDOC.【Methods】1.The study was a single-center,retrospective cohort study of VS / UWS and MCS patients admitted to the intensive-care unit of neurologic intensive care unit(NICU)in Xijing Hospital from January 2014 to December 2019.All patients were followed by telephone,the follow-up period was 1-7 years.The primary end point was the functional outcome.The functional outcome was divided into two groups according to the Extended Glasgow Outcome Scale(GOSE).GOSE 1-3 was poor,and 4-8 was good.The secondary end point was survival or death.Demographic information,clinical parameters,electrophysiological markers and serological indexes of patients were collected.Univariate and multivariate analyses were performed to explore the risk factors for long-term functional prognosis,survival curves were plotted and Cox regression analysis was used to explore the risk factors for long-term survival.Finally,the R language was used to establish a nomogram series linear model to predict the long-term clinical outcome of each patient.2.The study was a single-center,prospective,explorative study involving VS /UWS and MCS patients admitted to the intensive-care unit of neurologic intensive care unit(NICU)in Xijing Hospital between December 2017 and June 2019.Volunteers were recruited as health controls.Demographic and clinical data were collected and patients were monitored at least 1 hour resting state EEG,followed by an intervention of 10 mg zolpidem and continued monitoring for 5 hours.Using MATLAB and EEGLAB software,the power spectrum of EEG was analyzed quantitatively in four different frequency bands,including Delta(0.5-4 Hz),theta(4-8 Hz),alpha(8-13 Hz)and beta(13-30 Hz).According to the power spectrum,the EEG patterns of healthy controls and patients before and after taking medicine were summarized,including A,B,C,D four EEG patterns,Mode D was the highest.EEG pattern A: delta frequency power mainly.EEG pattern B: a local peak in theta frequency band.EEG pattern C: a local peak in the theta band and a local peak in the beta band,or only a local peak in the beta band.EEG pattern D: a local peak in the alpha band and a local peak in the beta band,or only a local peak in the alpha band.In this study,the highest pattern of all leads was chosen as the EEG pattern for each patient before and after treatment.First,the changes of power spectrum,electroencephalogram(EEG)patterns and coherence were analyzed before and after treatment in each healthy controls and patients.All patients were then followed up to see if their consciousness improved 6 months after enrollment according to the Coma Recovery Scale-Revised(CRS-R).Among them,consciousness improvement included VS / UWS switching to MCS or EMCS,MCS minus switching to MCS plus or EMCS,MCS plus switching to EMCS.Enventually,the changes of power values and EEG patterns at different frequency bands of different leads were compared before and after medication in healthy control group,consciousness improvement group and consciousness non-improvement group.To explore the effect of zolpidem on resting state EEG in patients with CDOC,we then analyzed the coherence of different frequencies bands of EEG in different regions before and after treatment.【Results】1.A total of 151 eligible VS / UWS and MCS patients were enrolled in this study.At the follow-up date,11 patients were lost to follow up and the remaining 140 patients were enrolled in the study.Finally,94 patients(67.1%)got unfavorable outcome(GOSE score 1-3)and 39 patients(32.9%)died.Patients with traumatic brain injury had the best long-term functional prognosis(N = 9,50.0 %)and the best survival rate(n = 15,83.3 %).While patients with hypoxic-ischemic encephalopathy had the best functional prognosis(N = 38,76.0 %)and the lowest survival rate(n = 33,66.6 %),but there was no significant statistical difference.Nearly half of the patients(17,50.0%)had a good long-term functional outcome,while only 29(27.4%)of VS/UWS patients had a good long-term functional outcome.There was no significant difference between VS/UWS patients(71.7%)and MCS patients(73.5%)in long-term survival.Logistic regression analysis showed that there were significant differences between the two groups in CRS-R score(p = 0.007),EEG reactivity(p = 0.027),EEG sleep spindles(p = 0.001)and EEG Synek scale(p = 0.005).Mode III had better discriminative power(AUC = 0.827,95% CI:0.755-0.900)than mode II(AUC = 0.779,95% CI: 0.696-0.863)and mode I(AUC =0.790,95% CI: 0.710-0.870).Cox regression analysis showed that there were significant differences in Synek scale(p = 0.020),APACHE II score(p = 0.043)and educational level(p = 0.040)between survival and death groups.Cox regression analysis showed that the Synek scael(p = 0.020),APACHE II(p = 0.043)and educational level(p = 0.040)had significant influence on survival outcome.The nomogram series linear model can predict the long-term functional outcome and survival outcome of each patient.2.In this study,36 patients and 7 healthy controls met the criteria,2 patients and 1healthy control were excluded due to the excessive movement artifacts and occular contamination in EEG.The results showed that the EEG patterns of 6 healthy controls before and after taking zolpidem were pattern D,Pattern A was the main baseline EEG pattern,the rest were pattern B,pattern C,and pattern D,respectively.Pattern A was still the main EEG pattern after zopidem administration,the rest were pattern D and pattern B,and there was no pattern C.The power spectrum and coherence of different frequency bands showed one changing trend in 6 healthy controls before and after zolpidem intake,while the patients showed four changing trends of power spectrum and three changing trends of coherence.Three patients were lost to followed-up at endpoint,the resting 31 patients and 6 healthy controls were included in futher analysis.Finally,18 patients’ consciousness improved and 13 patients did not improve,and one of them died.Univariate analysis showed that the improvement of EEG pattern in conscious group was significantly better than that in unconscious group(p = 0.032).In the healthy control group,the power of delta and theta frequencies were significantly increased,while the power of alpha and beta frequencies were significantly decreased after zopidem intake.The patients in the favorable outcome group showed the opposite pattern of EEG activation,that is,the delta and theta power were significantly decreased,while the alpha and beta power were significantly increased after drug administration.The power of all frequency bands were significantly decreased in the unfavorable outcome group.Compared with other frequency bands,beta frequency bands at 1 hour post drug administration had the best predictive power(AUC = 0.829,95 % CI:0.810-0.847),sensitivity(94.38 %),negative predictive value(90.1 %)and accuracy(81.48 %)for predicting the outcome of patients at endpoint.The coherence of different brain regions in different frequency bands of EEG was consistent with the corresponding power spectrum,especially in frontal lobe and occipital lobe.【Conclusions】1.The rate of poor long-term functional prognosis in CDOC patients was 67.1%and the mortality was 32.9 %.Baseline CRS-R score,EEG reactivity,EEG sleep spindles,EEG Synek scale combined with demographic characteristics of patients,including age,gender and etiology were more valuable in predicting the long-term functional outcome of patients with DOC.Synek scacle,AACHE II score and educational level were the three main factors influencing the long-term survival rate of DOC patients.Patients with traumatic brain injury had the best long-term functional prognosis and the highest survival rate,while patients with hypoxic-ischemic encephalopathy had the lowest functional prognosis and lowest survival rate.2.The abnormal pattern of EEG activation evoked by zolpidem in patients with improved consciousness,as opposed to the healthy controls,and the overall suppression pattern evoked by zolpidem in patients with unimproved consciousness,were helpful in assessing the prognosis of CDOC patients,it can effectively screen out the patients who have good response to zolpidem,and its EEG response can further guide clinicians and patients’ families to make decision of the follow-up treatment program.