Accuracy of clinical signs, SEP, and EEG in predicting outcome of hypoxic coma: A meta-analysis

A prediction model of clinical outcomes in prolonged disorders of consciousness: A prospective cohort study

Objective

This study aimed to establish and validate a prediction model for clinical outcomes in patients with prolonged disorders of consciousness (pDOC).

Methods

A total of 170 patients with pDOC enrolled in our rehabilitation unit were included and divided into training (n = 119) and validation sets (n = 51). Independent predictors for improved clinical outcomes were identified by univariate and multivariate logistic regression analyses, and a nomogram model was established. The nomogram performance was quantified using receiver operating curve (ROC) and calibration curves in the training and validated sets. A decision curve analysis (DCA) was performed to evaluate the clinical usefulness of this nomogram model.

Results

Univariate and multivariate logistic regression analyses indicated that age, diagnosis at entry, serum albumin (g/L), and pupillary reflex were the independent prognostic factors that were used to construct the nomogram. The area under the curve in the training and validation sets was 0.845 and 0.801, respectively. This nomogram model showed good calibration with good consistency between the actual and predicted probabilities of improved outcomes. The DCA demonstrated a higher net benefit in clinical decision-making compared to treating all or none.

Conclusion

Several feasible, cost-effective prognostic variables that are widely available in hospitals can provide an efficient and accurate prediction model for improved clinical outcomes and support clinicians to offer suitable clinical care and decision-making to patients with pDOC and their family members.

Food Aspiration Induced Hypoxic Encephalopathy Leading to Status Epilepticus

Cardiopulmonary arrest (CPA) is an emergency that can easily lead to death without appropriate life support. Even if the return of spontaneous circulation (ROSC) is achieved, survivors of sudden cardiac arrest have sustained various degrees of hypoxic encephalopathy. In Japan, rural community hospitals tend to provide care to patients with status epilepticus caused by hypoxic encephalopathy after CPA without sufficient resources. These hospitals neither have enough staff or equipment nor can they perform all the tests required to accurately estimate the prognosis. However, simple methods can be used for better estimation, including reviewing information on arrival, physical examinations, and imaging tests. Herein, we report a case of status epilepticus caused by hypoxic encephalopathy due to food aspiration in a 72-year-old man. For the sake of patients and their families, hospitals without sufficient staff or equipment should try to estimate the prognosis of patients in a manner similar to that described in this report.

The Combination of N60 with Mismatch Negativity Improves the Prediction of Awakening from Coma

BACKGROUND

Our objective was to evaluate the use of event-related potentials and the middle-latency somatosensory evoked potential (MLSEP) for the prediction of awakening in coma, determine the evaluation day that evoked potentials (EPs) best predict an awakening outcome, and determine whether the mismatch negativity (MMN) combined with the MLSEP, when recorded at 7 days after coma, improved the prediction of awakening from coma.

METHODS

Design prospective blinded cohort study. Setting neurointensive care unit of a university hospital. Patients 113 consecutive patients who were severely comatose, whose etiologies of coma included stroke (65 patients), hypoxic-ischemic encephalopathy (28 patients), intracranial infection (6 patients), and other (14 patients). Interventions none. Measurements we gathered Glasgow Coma Scale scores and recorded EPs for all patients who were comatose at 7, 14, and 30 days after coma onset, unless the patients returned to consciousness. The EPs examined included the MLSEP, the middle-latency auditory evoked potential, the N100, and the MMN. With telephone follow-up after 3 months, the patients were classified as awakening or nonawakening according to Glasgow Outcome Scale.

RESULTS

When predicting an awakening outcome, at least the unilateral presence of the N60 had the highest sensitivity (82.7%), whereas the presence of the MMN showed the highest specificity (82.0%). The area under the receiver operating characteristic curve for the EPs were high at 7 days after coma onset. At 7 days after coma onset, the combination of the N60 and MMN offered good predictive performance for awakening (area under the receiver operating characteristic curve = 0.852, 95% confidence interval 0.765-0.940), with increased sensitivity (70.0%) and improved specificity (91.7%).

CONCLUSIONS

The N60 and MMN were the strongest prognostic factors for an awakening outcome. Furthermore, at 7 days after coma onset, the combination of the N60 and MMN improved the prediction of an awakening outcome in patients who were comatose.

Prognostic roles of sleep electroencephalography pattern and circadian rhythm biomarkers in the recovery of consciousness in patients with coma: a prospective cohort study

OBJECTIVE

To investigate the potential prognostic value of sleep electroencephalography (EEG) pattern and serum circadian rhythm biomarkers in the recovery of consciousness in patients at the acute stage of coma.

METHODS

A prospective observational study which included 75 patients with coma was conducted. Twenty-four-hour continuous polysomnography (PSG) was performed to determine the sleep EEG pattern according to the modified Valente's Grade (mVG) that we proposed. Serum levels of melatonin and orexin-A at four consecutive time points during the PSG were examined. Patients were then followed for one month to determine their level of consciousness. Multivariate logistic regression analysis was performed to examine associations between demographics, aetiology, baseline clinical features (pupillary and corneal reflex, and neuron-specific enolase [NSE]), clinical scores (Glasgow Coma Scale-Motor Response [GCS-M], Full Outline of Unresponsiveness [FOUR] scale, Acute Physiology and Chronic Health Evaluation II [APACHE II] scale), mVG, serum circadian biomarkers, and recovery of consciousness within one month.

RESULTS

Within one month of enrolment, 34 patients regained consciousness and 36 patients remained non-conscious. Spearman rank correlation revealed a significant association between mVG and state of consciousness after one month. Significant variation in serum melatonin or orexin-A was not detected in either the conscious or non-conscious groups. Hypoxic aetiology, APACHE II, and mVG were independently associated with recovery of consciousness within one month.

CONCLUSION

Sleep EEG structure, hypoxic aetiology, and APACHE II can independently predict recovery of consciousness in patients with acute coma. Taken together, we encourage neurologists to use sleep elements to assess patients with acute coma.

Estimating the False Positive Rate of Absent Somatosensory Evoked Potentials in Cardiac Arrest Prognostication

OBJECTIVES

Absence of somatosensory evoked potentials is considered a nearly perfect predictor of poor outcome after cardiac arrest. However, reports of good outcomes despite absent somatosensory evoked potentials and high rates of withdrawal of life-sustaining therapies have raised concerns that estimates of the prognostic value of absent somatosensory evoked potentials may be biased by self-fulfilling prophecies. We aimed to develop an unbiased estimate of the false positive rate of absent somatosensory evoked potentials as a predictor of poor outcome after cardiac arrest.

DATA SOURCES

PubMed.

STUDY SELECTION

We selected 35 studies in cardiac arrest prognostication that reported somatosensory evoked potentials.

DATA EXTRACTION

In each study, we identified rates of withdrawal of life-sustaining therapies and good outcomes despite absent somatosensory evoked potentials. We appraised studies for potential biases using the Quality in Prognosis Studies tool. Using these data, we developed a statistical model to estimate the false positive rate of absent somatosensory evoked potentials adjusted for withdrawal of life-sustaining therapies rate.

DATA SYNTHESIS

Two-thousand one-hundred thirty-three subjects underwent somatosensory evoked potential testing. Five-hundred ninety-four had absent somatosensory evoked potentials; of these, 14 had good functional outcomes. The rate of withdrawal of life-sustaining therapies for subjects with absent somatosensory evoked potential could be estimated in 14 of the 35 studies (mean 80%, median 100%). The false positive rate for absent somatosensory evoked potential in predicting poor neurologic outcome, adjusted for a withdrawal of life-sustaining therapies rate of 80%, is 7.7% (95% CI, 4-13%).

CONCLUSIONS

Absent cortical somatosensory evoked potentials do not infallibly predict poor outcome in patients with coma following cardiac arrest. The chances of survival in subjects with absent somatosensory evoked potentials, though low, may be substantially higher than generally believed.

Initial absence of N20 waveforms from median nerve somatosensory evoked potentials in a patient with cardiac arrest and good outcomes

A 34-year-old male was brought to the hospital with a chest gunshot wound. Pulseless upon arrival, blood pressure was absent for 10 minutes. A thoracotomy resulted in return of spontaneous circulation. On hospital day 5, with brainstem reflexes present, he was unresponsive to call or pain, exhibited generalized hyperreflexia and bilateral Babinskys. Median nerve somatosensory evoked potentials (mSSEPs) and brainstem auditory evoked potentials were obtained. International Federation of Clinical Neurophysiology recommendations for mSSEPs and brainstem auditory evoked potentials were followed. Despite absence of the N20 responses from cortical mSSEPs no withdrawal from care was agreed upon. After awaking on day 7, mSSEPs were repeated and present. The patient survived and was discharged with minor deficits. Bilateral absence of N20 responses from mSSEPs performed beyond 48 hours after resuscitation from cardiac arrest is highly associated with bad neurological outcomes. However, variation due to hypothermia, noisy signals, medications, and brain hypo-perfusion must be taken into account.

Clinical neurophysiology of neurologic rehabilitation

Clinical neurophysiologic testing provides valuable support in predicting outcome in the setting of disorders of consciousness (DOC), including coma and traumatic brain injury (TBI). Electroencephalography (EEG) and evoked potentials (EP) are simple to apply, inexpensive, safe, and available in most rehabilitation facilities. This chapter reviews the use of EEG and EP in postanoxic coma and TBI. Bilateral absence of cortical somatosensory evoked potentials (SSEP) may be regarded as a predictor of poor outcome in hypoxic brain damage. Flash VEP may be useful to differentiate between good and poor outcome. In addition, low EEG frequencies, burst suppression, and isoelectric EEG patterns prognosticate poor outcomes in hypoxic brain damage. While a loss of cortical SSEP is generally regarded as a negative prognostic sign in the acute phase of hypoxic brain damage, absence of cortical SSEP responses is not necessarily associated with poor outcome in TBI. Event-related potentials (ERPs) can provide support in outcome prediction. In particular, the N100, mismatch negativity, P300, and N400 may improve accuracy of outcome prediction DOC of different etiologies. Some evidence suggests that ERPs may be superior to SSEP in predicting functional and DOC outcomes (Lew et al., 2003). ERPs are measured brain responses resulting from specific cognitive tasks, sensory stimulation, or planned motor activity.

A Machine Learning Framework for Automatic and Continuous MMN Detection With Preliminary Results for Coma Outcome Prediction

Mismatch negativity (MMN) is a component of the event-related potential (ERP) that is elicited through an odd-ball paradigm. The existence of the MMN in a coma patient has a good correlation with coma emergence; however, this component can be difficult to detect. Previously, MMN detection was based on visual inspection of the averaged ERPs by a skilled clinician, a process that is expensive and not always feasible in practice. In this paper, we propose a practical machine learning (ML) based approach for detection of MMN component, thus, improving the accuracy of prediction of emergence from coma. Furthermore, the method can operate on an automatic and continuous basis thus alleviating the need for clinician involvement. The proposed method is capable of the MMN detection over intervals as short as two minutes. This finer time resolution enables identification of waxing and waning cycles of a conscious state. An auditory odd-ball paradigm was applied to 22 healthy subjects and 2 coma patients. A coma patient is tested by measuring the similarity of the patient's ERP responses with the aggregate healthy responses. Because the training process for measuring similarity requires only healthy subjects, the complexity and practicality of training procedure of the proposed method are greatly improved relative to training on coma patients directly. Since there are only two coma patients involved with this study, the results are reported on a very preliminary basis. Preliminary results indicate we can detect the MMN component with an accuracy of 92.7% on healthy subjects. The method successfully predicted emergence in both coma patients when conventional methods failed. The proposed method for collecting training data using exclusively healthy subjects is a novel approach that may prove useful in future, unrelated studies where ML methods are used.

Exploratory Use of Decision Tree Analysis in Classification of Outcome in Hypoxic-Ischemic Brain Injury

Background

Prognostication following hypoxic ischemic encephalopathy (brain injury) is important for clinical management. The aim of this exploratory study is to use a decision tree model to find clinical and MRI associates of severe disability and death in this condition. We evaluate clinical model and then the added value of MRI data.

Method

The inclusion criteria were as follows: age ≥17 years, cardio-respiratory arrest, and coma on admission (2003-2011). Decision tree analysis was used to find clinical [Glasgow Coma Score (GCS), features about cardiac arrest, therapeutic hypothermia, age, and sex] and MRI (infarct volume) associates of severe disability and death. We used the area under the ROC (auROC) to determine accuracy of model. There were 41 (63.7% males) patients having MRI imaging with the average age 51.5 ± 18.9 years old. The decision trees showed that infarct volume and age were important factors for discrimination between mild to moderate disability and severe disability and death at day 0 and day 2. The auROC for this model was 0.94 (95% CI 0.82-1.00). At day 7, GCS value was the only predictor; the auROC was 0.96 (95% CI 0.86-1.00).

Conclusion

Our findings provide proof of concept for further exploration of the role of MR imaging and decision tree analysis in the early prognostication of hypoxic ischemic brain injury.

Chronic disorders of consciousness

Over the last 20 years, studies have provided greater insight into disorders of consciousness (DOC), also known as altered state of consciousness. Increased brain residual functions have been identified in patients with DOC due to the successful application of novel next-generation imaging technologies. Many unconscious patients have now been confirmed to retain considerable cognitive functions. It is hoped that greater insight regarding the psychological state of patients may be achieved through the use of functional magnetic resonance imaging and brain-computer interfaces. However, issues surrounding the research and treatment of DOC remain problematic. These include differing opinions on the definition of consciousness, difficulties in diagnosis, assessment, prognosis and/or treatment, and newly emerging ethical, legal and social issues. To overcome these, appropriate care must be offered to patients with DOC by clinicians and families, as DOC patients may now be considered to live in more than just a vegetative state. The present article reviews the controversy surrounding the definition of consciousness and the reliability of novel technologies, prognostic prediction, communication with DOC patients and treatment methods. The ethical and social issues surrounding the treatment of DOC and future perspectives are also considered.

Whole-scalp EEG mapping of somatosensory evoked potentials in macaque monkeys

High-density scalp EEG recordings are widely used to study whole-brain neuronal networks in humans non-invasively. Here, we validate EEG mapping of somatosensory evoked potentials (SSEPs) in macaque monkeys (Macaca fascicularis) for the long-term investigation of large-scale neuronal networks and their reorganisation after lesions requiring a craniotomy. SSEPs were acquired from 33 scalp electrodes in five adult anaesthetized animals after electrical median or tibial nerve stimulation. SSEP scalp potential maps were identified by cluster analysis and identified in individual recordings. A distributed, linear inverse solution was used to estimate the intracortical sources of the scalp potentials. SSEPs were characterised by a sequence of components with unique scalp topographies. Source analysis confirmed that median nerve SSEP component maps were in accordance with the somatotopic organisation of the sensorimotor cortex. Most importantly, SSEP recordings were stable both intra- and interindividually. We aim to apply this method to the study of recovery and reorganisation of large-scale neuronal networks following a focal cortical lesion requiring a craniotomy. As a prerequisite, the present study demonstrated that a 300-mm² unilateral craniotomy over the sensorimotor cortex necessary to induce a cortical lesion, followed by bone flap repositioning, suture and gap plugging with calcium phosphate cement, did not induce major distortions of the SSEPs. In conclusion, SSEPs can be successfully and reproducibly recorded from high-density EEG caps in macaque monkeys before and after a craniotomy, opening new possibilities for the long-term follow-up of the cortical reorganisation of large-scale networks in macaque monkeys after a cortical lesion.

Development of a simple score to predict outcome for unresponsive wakefulness syndrome

Introduction

Accurate assessment of prognosis for patients with unresponsive wakefulness syndrome (UWS; formerly vegetative state) may help clinicians and families guide the type and intensity of therapy; however, there is no suitable and accurate means to predict the outcome so far. We aimed to develop a simple bedside scoring system to predict the likelihood of awareness recovery in patients with UWS.

Methods

We prospectively enrolled 56 patients (age range 10 to 73 years) with UWS 3 to 12 weeks post-onset. We collected demographic data and performed neurological, serological and neurophysiological tests at study entry. Each patient received a one year follow-up, during which awareness recovery was assessed by experienced physicians on the basis of clinical criteria. Univariate and multivariable analyses were employed to assess the relationships between predictors and awareness recovery.

Results

A total of 56 participants were included in the study; of these, 24 patients recovered awareness, 3 with moderate disabilities, 8 with severe disabilities, 12 were in a minimally conscious state, and 1 died after recovery. During the study, 23 patients remained in UWS and 9 died in UWS. Motor response, type of brain injury, electroencephalogram reactivity, sleep spindles and N20 were shown to be independent predictors for awareness recovery. Based on their coefficients in the model, we assigned these predictors with 1 point each and created a 5-point score for prediction of awareness recovery. The resulting score showed good predictive accuracy in the derivation cohort. The area under the receiver operating characteristic curve for the score was 0.918 with 87.50% sensitivity.

Conclusion

This simple bedside prognostic score can be used to predict the probability of awareness recovery in UWS, thus provide families and clinicians with useful outcome information.

MMN and novelty P3 in coma and other altered states of consciousness: a review

In recent decades, there has been a growing interest in the assessment of patients in altered states of consciousness. There is a need for accurate and early prediction of awakening and recovery from coma. Neurophysiological assessment of coma was once restricted to brainstem auditory and primary cortex somatosensory evoked potentials elicited in the 30 ms range, which have both shown good predictive value for poor coma outcome only. In this paper, we review how passive auditory oddball paradigms including deviant and novel sounds have proved their efficiency in assessing brain function at a higher level, without requiring the patient's active involvement, thus providing an enhanced tool for the prediction of coma outcome. The presence of an MMN in response to deviant stimuli highlights preserved automatic sensory memory processes. Recorded during coma, MMN has shown high specificity as a predictor of recovery of consciousness. The presence of a novelty P3 in response to the subject's own first name presented as a novel (rare) stimulus has shown a good correlation with coma awakening. There is now a growing interest in the search for markers of consciousness, if there are any, in unresponsive patients (chronic vegetative or minimally conscious states). We discuss the different ERP patterns observed in these patients. The presence of novelty P3, including parietal components and possibly followed by a late parietal positivity, raises the possibility that some awareness processes are at work in these unresponsive patients.

Prognosis in severe brain injury

BACKGROUND

The prediction of neurologic outcome is a fundamental concern in the resuscitation of patients with severe brain injury.

OBJECTIVE

To provide an evidence-based update on neurologic prognosis following traumatic brain injury and hypoxic-ischemic encephalopathy after cardiac arrest.

DATA SOURCE

Search of the PubMed database and manual review of bibliographies from selected articles to identify original data relating to prognostic methods and outcome prediction models in patients with neurologic trauma or hypoxic-ischemic encephalopathy.

DATA SYNTHESIS AND CONCLUSION

Articles were scrutinized regarding study design, population evaluated, interventions, outcomes, and limitations. Outcome prediction in severe brain injury is reliant on features of the neurologic examination, anatomical and physiological changes identified with CT and MRI, abnormalities detected with electroencephalography and evoked potentials, and physiological and biochemical derangements at both the brain and systemic levels. Use of such information in univariable association studies generally lacks specificity in classifying neurologic outcome. Furthermore, the accuracy of established prognostic classifiers may be affected by the introduction of outcome-modifying interventions, such as therapeutic hypothermia following cardiac arrest. Although greater specificity may be achieved with scoring systems derived from multivariable models, they generally fail to predict outcome with sufficient accuracy to be meaningful at the single patient level. Discriminative models which integrate knowledge of genetic determinants and biologic processes governing both injury and repair and account for the effects of resuscitative and rehabilitative care are needed.

Continuous electroencephalography monitoring for early prediction of neurological outcome in postanoxic patients after cardiac arrest: a prospective cohort study

OBJECTIVE

To evaluate the value of continuous electroencephalography in early prognostication in patients treated with hypothermia after cardiac arrest.

DESIGN

Prospective cohort study.

SETTING

Medical intensive care unit.

PATIENTS

Sixty patients admitted to the intensive care unit for therapeutic hypothermia after cardiac arrest.

INTERVENTION

None.

MEASUREMENTS AND MAIN RESULTS

In all patients, continuous electroencephalogram and daily somatosensory evoked potentials were recorded during the first 5 days of admission or until intensive care unit discharge. Neurological outcomes were based on each patient's best achieved Cerebral Performance Category score within 6 months. Twenty-seven of 56 patients (48%) achieved good neurological outcome (Cerebral Performance Category score 1-2).At 12 hrs after resuscitation, 43% of the patients with good neurological outcome showed continuous, diffuse slow electroencephalogram rhythms, whereas this was never observed in patients with poor outcome.The sensitivity for predicting poor neurological outcome of low-voltage and isoelectric electroencephalogram patterns 24 hrs after resuscitation was 40% (95% confidence interval 19%-64%) with a 100% specificity (confidence interval 86%-100%), whereas the sensitivity and specificity of absent somatosensory evoked potential responses during the first 24 hrs were 24% (confidence interval 10%-44%) and 100% (confidence interval: 87%-100%), respectively. The negative predictive value for poor outcome of low-voltage and isoelectric electroencephalogram patterns was 68% (confidence interval 50%-81%) compared to 55% (confidence interval 40%-60%) for bilateral somatosensory evoked potential absence, both with a positive predictive value of 100% (confidence interval 63%-100% and 59%-100% respectively). Burst-suppression patterns after 24 hrs were also associated with poor neurological outcome, but not inevitably so.

CONCLUSIONS

In patients treated with hypothermia, electroencephalogram monitoring during the first 24 hrs after resuscitation can contribute to the prediction of both good and poor neurological outcome. Continuous patterns within 12 hrs predicted good outcome. Isoelectric or low-voltage electroencephalograms after 24 hrs predicted poor outcome with a sensitivity almost two times larger than bilateral absent somatosensory evoked potential responses.

Predicting neurological outcome after cardiac arrest

PURPOSE OF REVIEW

Therapeutic hypothermia and aggressive management of postresuscitation disease considerably improved outcome after adult cardiac arrest over the past decade. However, therapeutic hypothermia alters prognostic accuracy. Parameters for outcome prediction, validated by the American Academy of Neurology before the introduction of therapeutic hypothermia, need further update.

RECENT FINDINGS

Therapeutic hypothermia delays the recovery of motor responses and may render clinical evaluation unreliable. Additional modalities are required to predict prognosis after cardiac arrest and therapeutic hypothermia. Electroencephalography (EEG) can be performed during therapeutic hypothermia or shortly thereafter; continuous/reactive EEG background strongly predicts good recovery from cardiac arrest. On the contrary, unreactive/spontaneous burst-suppression EEG pattern, together with absent N20 on somatosensory evoked potentials (SSEP), is almost 100% predictive of irreversible coma. Therapeutic hypothermia alters the predictive value of serum markers of brain injury [neuron-specific enolase (NSE), S-100B]. Good recovery can occur despite NSE levels >33 μg/l, thus this cut-off value should not be used to guide therapy. Diffusion MRI may help predicting long-term neurological sequelae of hypoxic-ischemic encephalopathy.

SUMMARY

Awakening from postanoxic coma is increasingly observed, despite early absence of motor signs and frank elevation of serum markers of brain injury. A new multimodal approach to prognostication is therefore required, which may particularly improve early prediction of favorable clinical evolution after cardiac arrest.

Prognosis after cardiac arrest and hypothermia: a new paradigm

Before the use of hypothermia as a treatment for comatose post-cardiac arrest patients, several prognostic variables were widely accepted as reliable and valid for the prediction of poor outcome. Recent studies using hypothermia have reported on patients with recovery of consciousness in spite of absent or extensor motor responses after 3 days, absent bilateral cortical N20 responses after 24 h, serum neuron-specific enolase levels greater than 33 μg/L, and early myoclonus status epilepticus. Hypothermia and its associated use of sedative and paralytic agents may delay neurologic recovery and affect the optimal timing of prognostic variables. Recent developments in brain imaging may provide additional objective prognostic information and deserve further study. Pending the results of future validation studies in patients treated with hypothermia, we recommend that irreversible management decisions not be made based on a single prognostic parameter, and, if there is uncertainty, these decisions be delayed for several days to allow for repeated testing.