Recent advances in clinical electroencephalography.

PURPOSE OF REVIEW: Clinical electroencephalography (EEG) is a conservative medical field. This explains likely the significant gap between clinical practice and new research developments. This narrative review discusses possible causes of this discrepancy and how to circumvent them. More specifically, we summarize recent advances in three applications of clinical EEG: source imaging (ESI), high-frequency oscillations (HFOs) and EEG in critically ill patients. RECENT FINDINGS: Recently published studies on ESI provide further evidence for the accuracy and clinical utility of this method in the multimodal presurgical evaluation of patients with drug-resistant focal epilepsy, and opened new possibilities for further improvement of the accuracy. HFOs have received much attention as a novel biomarker in epilepsy. However, recent studies questioned their clinical utility at the level of individual patients. We discuss the impediments, show up possible solutions and highlight the perspectives of future research in this field. EEG in the ICU has been one of the major driving forces in the development of clinical EEG. We review the achievements and the limitations in this field. SUMMARY: This review will promote clinical implementation of recent advances in EEG, in the fields of ESI, HFOs and EEG in the intensive care.

The predictive value of highly malignant EEG patterns after cardiac arrest: evaluation of the ERC-ESICM recommendations.

PURPOSE: The 2021 guidelines endorsed by the European Resuscitation Council (ERC) and the European Society of Intensive Care Medicine (ESICM) recommend using highly malignant electroencephalogram (EEG) patterns (HMEP; suppression or burst-suppression) at > 24 h after cardiac arrest (CA) in combination with at least one other concordant predictor to prognosticate poor neurological outcome. We evaluated the prognostic accuracy of HMEP in a large multicentre cohort and investigated the added value of absent EEG reactivity. METHODS: This is a pre-planned prognostic substudy of the Targeted Temperature Management trial 2. The presence of HMEP and background reactivity to external stimuli on EEG recorded > 24 h after CA was prospectively reported. Poor outcome was measured at 6 months and defined as a modified Rankin Scale score of 4-6. Prognostication was multimodal, and withdrawal of life-sustaining therapy (WLST) was not allowed before 96 h after CA. RESULTS: 845 patients at 59 sites were included. Of these, 579 (69%) had poor outcome, including 304 (36%) with WLST due to poor neurological prognosis. EEG was recorded at a median of 71 h (interquartile range [IQR] 52-93) after CA. HMEP at > 24 h from CA had 50% [95% confidence interval [CI] 46-54] sensitivity and 93% [90-96] specificity to predict poor outcome. Specificity was similar (93%) in 541 patients without WLST. When HMEP were unreactive, specificity improved to 97% [94-99] (p = 0.008). CONCLUSION: The specificity of the ERC-ESICM-recommended EEG patterns for predicting poor outcome after CA exceeds 90% but is lower than in previous studies, suggesting that large-scale implementation may reduce their accuracy. Combining HMEP with an unreactive EEG background significantly improved specificity. As in other prognostication studies, a self-fulfilling prophecy bias may have contributed to observed results.

Continuous versus Routine Electroencephalography in the Intensive Care Unit: A Review of Current Evidence.

BACKGROUND: Electroencephalography (EEG) has long been used to detect seizures in patients with disorders of consciousness. In recent years, there has been a drastically increased adoption of continuous EEG (cEEG) in the intensive care units (ICUs). Given the resources necessary to record and interpret cEEG, this is still not available in every center and widespread recommendations to use continuous instead of routine EEG (typically lasting 20 min) are still a matter of some debate. Considering recent literature and personal experience, this review offers a rationale and practical advice to address this question. SUMMARY: Despite the development of increasingly performant imaging techniques and several validated biomarkers, EEG remains central to clinicians in the intensive care unit and has been experiencing expanding popularity for at least 2 decades. Not only does EEG allow seizure or status epilepticus detection, which in the ICU often present without clinical movements, but it is also paramount for the prognostic evaluation of comatose patients, especially after cardiac arrest, and for detecting delayed ischemia after subarachnoid hemorrhage. At the end of the last Century, improvements of technical and digital aspects regarding recording and storage of EEG tracings have progressively led to the era of cEEG and automated quantitative analysis. KEY MESSAGES: As compared to repeated rEEG, cEEG in comatose patients does not seem to improve clinical prognosis to a relevant extent, despite allowing a more performant of detection ictal events and consequent therapeutic modifications. The choice between cEEG and rEEG must therefore always be patient-tailored.

Staged treatment response in status epilepticus: Lessons from the SENSE registry.

OBJECTIVES: Although in epilepsy patients the likelihood of becoming seizure-free decreases substantially with each unsuccessful treatment, to our knowledge this has been poorly investigated in status epilepticus (SE). We aimed to evaluate the proportion of SE cessation and functional outcome after successive treatment steps. METHODS: We conducted a post hoc analysis of a prospective, observational, multicenter cohort (Sustained Effort Network for treatment of Status Epilepticus [SENSE]), in which 1049 incident adult SE episodes were prospectively recorded at nine European centers. We analyzed 996 SE episodes without coma induction before the third treatment step. Rates of SE cessation, mortality (in ongoing SE or after SE control), and favorable functional outcome (assessed with modified Rankin scale) were evaluated after each step. RESULTS: SE was treated successfully in 838 patients (84.1%), 147 (14.8%) had a fatal outcome (36% of them died while still in SE), and 11 patients were transferred to palliative care while still in SE. Patients were treated with a median of three treatment steps (range 1-13), with 540 (54.2%) receiving more than two steps (refractory SE [RSE]) and 95 (9.5%) more than five steps. SE was controlled after the first two steps in 45%, with an additional 21% treated after the third, and 14% after the fourth step. Likelihood of SE cessation (p < 0.001), survival (p = 0.003), and reaching good functional outcome (p < 0.001) decreased significantly between the first two treatment lines and the third, especially in patients not experiencing generalized convulsive SE, but remained relatively stable afterwards. SIGNIFICANCE: The significant worsening of SE prognosis after the second step clinically supports the concept of RSE. However, and differing from findings in human epilepsy, RSE remains treatable in about one third of patients, even after several failed treatment steps. Clinical judgment remains essential to determine the aggressiveness and duration of SE treatment, and to avoid premature treatment cessation in patients with SE.

Status epilepticus in the ICU.

Status epilepticus (SE) is a common medical emergency associated with significant morbidity and mortality. Management that follows published guidelines is best suited to improve outcomes, with the most severe cases frequently being managed in the intensive care unit (ICU). Diagnosis of convulsive SE can be made without electroencephalography (EEG), but EEG is required to reliably diagnose nonconvulsive SE. Rapidly narrowing down underlying causes for SE is crucial, as this may guide additional management steps. Causes may range from underlying epilepsy to acute brain injuries such as trauma, cardiac arrest, stroke, and infections. Initial management consists of rapid administration of benzodiazepines and one of the following non-sedating intravenous antiseizure medications (ASM): (fos-)phenytoin, levetiracetam, or valproate; other ASM are increasingly used, such as lacosamide or brivaracetam. SE that continues despite these medications is called refractory, and most commonly treated with continuous infusions of midazolam or propofol. Alternatives include further non-sedating ASM and non-pharmacologic approaches. SE that reemerges after weaning or continues despite management with propofol or midazolam is labeled super-refractory SE. At this step, management may include non-sedating or sedating compounds including ketamine and barbiturates. Continuous video EEG is necessary for the management of refractory and super-refractory SE, as these are almost always nonconvulsive. If possible, management of the underlying cause of seizures is crucial particularly for patients with autoimmune encephalitis. Short-term mortality ranges from 10 to 15% after SE and is primarily related to increasing age, underlying etiology, and medical comorbidities. Refractoriness of treatment is clearly related to outcome with mortality rising from 10% in responsive cases, to 25% in refractory, and nearly 40% in super-refractory SE.

Outcome of comatose patients following cardiac arrest: When mRS completes CPC.

AIM: Good outcome in patients following cardiac arrest (CA) is usually defined as Cerebral Performance Category (CPC) 1-2, while CPC 3 is debated, and CPC 4-5 represent poor outcome. We aimed to assess when the modified Rankin Scale (mRS) can improve CPC outcome description, especially in CPC 3. We further aimed to correlate neuron specific enolase (NSE) with both functional measures to explore their relationship with neuronal damage. METHODS: Peak NSE within the first 48 hours, and CPC and mRS at 3 months were prospectively collected for 665 consecutive comatose adults following CA treated between April 2016 and April 2023. For each CPC category, mRS was described. We considered good outcome as mRS 1-3, in line with existing recommendations. CPC and mRS were correlated to peak serum NSE using non-parametric assessments. RESULTS: CPC 1, 2, 4 and 5 correlated almost perfectly with mRS in terms of good and poor outcomes. However, CPC 3 was heterogeneously associated to the dichotomized mRS (53.1% had good outcome (mRS 0-3), 46.9% poor outcome (mRS 4-6)). NSE was strongly correlated with CPC (Spearman's rho 0.616, P < 0.001) and mRS (Spearman's rho 0.613, P < 0.001). CONCLUSION: CPC and mRS correlate similarly with neuronal damage. Whilst CPC 1-2 and CPC 4-5 are strongly associated with mRS 0-3 and, respectively, with mRS 5-6, CPC 3 is heterogenous: both good and poor mRS scores are found within this category. Therefore, we suggest that the mRS should be routinely assessed in patients with CPC 3 to refine outcome description.

The effect of sedation and time after cardiac arrest on coma outcome prognostication based on EEG power spectra.

Early prognostication of long-term outcome of comatose patients after cardiac arrest remains challenging. Electroencephalography-based power spectra after cardiac arrest have been shown to help with the identification of patients with favourable outcome during the first day of coma. Here, we aim at comparing the power spectra prognostic value during the first and second day after coma onset following cardiac arrest and to investigate the impact of sedation on prognostication. In this cohort observational study, we included comatose patients (N = 91) after cardiac arrest for whom resting-state electroencephalography was collected on the first and second day after cardiac arrest in four Swiss hospitals. We evaluated whether the average power spectra values at 4.6-15.2 Hz were predictive of patients' outcome based on the best cerebral performance category score at 3 months, with scores ranging from 1 to 5 and dichotomized as favourable (1-2) and unfavourable (3-5). We assessed the effect of sedation and its interaction with the electroencephalography-based power spectra on patient outcome prediction through a generalized linear mixed model. Power spectra values provided 100% positive predictive value (95% confidence intervals: 0.81-1.00) on the first day of coma, with correctly predicted 18 out of 45 favourable outcome patients. On the second day, power spectra values were not predictive of patients' outcome (positive predictive value: 0.46, 95% confidence intervals: 0.19-0.75). On the first day, we did not find evidence of any significant contribution of sedative infusion rates to the patient outcome prediction (P > 0.05). Comatose patients' outcome prediction based on electroencephalographic power spectra is higher on the first compared with the second day after cardiac arrest. Sedation does not appear to impact patient outcome prediction.

Acute Valproate-Induced Encephalopathy in Status Epilepticus: A Registry-Based Assessment.

BACKGROUND: Valproate-induced encephalopathy (VIE) affects between 0.1% and 2.5% of patients under long-term epilepsy treatment. Its frequency and characteristics in adults with status epilepticus (SE) is, however, unknown. OBJECTIVE: The aim of this study was to characterize the frequency and the clinico-biological characteristics of VIE in adult SE patients. METHODS: We reviewed all patients included in our institutional SE registry who were treated for an SE episode between November 2021 and February 2023 and identified 39 patients who received valproate for their SE treatment. Acute VIE was defined by worsening of consciousness having led to the discontinuation of valproate, and improvement of consciousness within 96 hours after discontinuation of valproate during acute hospital treatment. RESULTS: Patients had a mean valproate intravenous loading dose of 34.5 mg/kg and a mean maintenance dose of 15.3 mg/kg/d (1078 mg/d). Four out of 29 patients with measured ammonium had hyperammonemia. We identified four (10%) patients fulfilling acute VIE criteria. Median time from administration of valproate to the occurrence of VIE, and to resolution of VIE after cessation of valproate treatment, was 2 days for each. Three of the four VIE patients had no associated hyperammonemia. Patients who developed VIE more frequently had a history of liver disease (p = 0.023), and tended to be younger, but other clinical variables did not differ significantly from patients without VIE, including valproate loading or maintenance doses, SE cause, duration or severity, other concomitant antiseizure medications (none received topiramate, phenobarbital, or primidone). CONCLUSION: Pending larger studies, VIE in SE seems relatively frequent and difficult to foresee; clinical alertness to symptoms is mandatory, even without hyperammonemia, and valproate withdrawal should be considered in suspected cases.

Clonazepam Loading Dose in Status Epilepticus: Is More Always Better?

BACKGROUND AND OBJECTIVES: Benzodiazepines are the first treatment line in status epilepticus (SE). Despite their well-established benefit, benzodiazepines are frequently underdosed with potential detrimental consequences. In some European countries, clonazepam (CLZ) is commonly used as the first line treatment. The aim of this study was to explore the correlation between CLZ loading doses and SE outcome. METHODS: This study included a retrospective analysis of a prospective registry in Lausanne, Switzerland (CHUV Lausanne University Hospital), including all SE episodes treated between February 2016 and February 2021. Only adults (> 16 years old) were included with CLZ used as the first treatment line. Post-anoxic SE were excluded because of significant differences in physiopathology and prognosis. Patient characteristics, SE features, the validated SE severity score (STESS), and treatment characteristics were prospectively recorded. We considered loading doses of 0.015 mg/kg or higher (following commonly recommended loading doses) as high doses. We analyzed outcome in terms of number of treatment lines after the CLZ, proportion of refractory episodes, intubation for airways protection, intubation for SE treatment, and mortality. We performed univariable analyses to investigate the association between loading doses and clinical response. A multivariable stepwise backward binary logistic regression was applied for adjusting for potential confounders. Multivariable linear regression was similarly used to analyze CLZ dose as a continuous variable. RESULTS: We collected 251 SE episodes in 225 adult patients. Median CLZ loading dose was 0.010 mg/kg. CLZ high doses were used in 21.9% of SE episodes (in 43.8% for > 80% of the high dose). Thirteen percent of patients with SE were intubated for airways control, while intubation was needed in 12.7% for SE treatment. High CLZ loading doses were independently associated with younger age (median 62 versus 68 years old, p = 0.002), lesser weight (65 kg versus 75 kg, p = 0.001) and more frequent intubation for airways protection (23% vs 11%, p = 0.013), but differing CLZ dose was not associated with any outcome parameter. CONCLUSION: CLZ high doses were more frequently used for SE treatment in younger patients with healthy weight and were more often associated with intubation for airways protection, probably as an adverse event. Varying CLZ dose did not alter outcome in SE, raising the possibility that commonly recommended doses are above what is needed, at least in some patients. Our results suggest that CLZ doses in SE may be individualized depending on the clinical setting.

Relationship between serum neuron-specific enolase and EEG after cardiac arrest: A reappraisal.

OBJECTIVE: Electroencephalogram (EEG) and serum neuron specific enolase (NSE) are frequently used prognosticators after cardiac arrest (CA). This study explored the association between NSE and EEG, considering the role of EEG timing, its background continuity, reactivity, occurrence of epileptiform discharges, and pre-defined malignancy degree. METHODS: Retrospective analysis including 445 consecutive adults from a prospective registry, surviving the first 24 hours after CA and undergoing multimodal evaluation. EEG were interpreted blinded to NSE results. RESULTS: Higher NSE was associated with poor EEG prognosticators, such as increasing malignancy, repetitive epileptiform discharges and lack of background reactivity, independently of EEG timing (including sedation and temperature). When stratified for background continuity, NSE was higher with repetitive epileptiform discharges, except in the case of suppressed EEGs. This relationship showed some variation according to the recording time. CONCLUSIONS: Neuronal injury after CA, reflected by NSE, correlates with several EEG features: increasing EEG malignancy, lack of background reactivity, and presence of repetitive epileptiform discharges. The correlation between epileptiform discharges and NSE is influenced by underlying EEG background and timing. SIGNIFICANCE: This study, describing the complex interplay between serum NSE and epileptiform features, suggests that epileptiform discharges reflect neuronal injury particularly in non-suppressed EEG.

Status epilepticus prognosis following levetiracetam administration: Analysis of loading doses.

BACKGROUND AND PURPOSE: Recommended loading doses of levetiracetam (LEV) for status epilepticus (SE) treatment have increased over time. However, this was not evidence-based, and the benefit of the increase remains unclear. The effect of different LEV loading doses on SE prognosis was explored. METHODS: This is a retrospective analysis of an SE adult registry (January 2016-December 2021), including patients receiving LEV as a second-line SE treatment. Patients were stratified according to LEV loading doses (threshold 35 mg/kg). Main outcomes were global mortality, LEV use as last SE treatment, and return to baseline conditions at discharge, exploring LEV as a dichotomized or continuous dose. RESULTS: Among 202 patients, 44 received LEV at ≥35 mg/kg and 158 below it. Global mortality, adjusted for SE severity and potentially fatal aetiology, was more frequent in the high LEV dose group (27.2% vs. 17.1%, odds ratio 3.14, 95% confidence interval 1.23-8.06; p = 0.017), whilst LEV prescription as last treatment and return to baseline conditions were comparable. Considering continuous LEV dosages or mortality in ongoing SE, however, no outcome reached statistical significance. CONCLUSIONS: Lower LEV loading doses do not seem to correlate with worse clinical outcome, challenging current guidelines. Further studies, ideally prospective, are needed on this topic.

Multimodal Prediction of Favorable Outcome After Cardiac Arrest: A Cohort Study.

OBJECTIVES: Prognostic guidelines after cardiac arrest (CA) focus on unfavorable outcome prediction; favorable outcome prognostication received less attention. Our aim was to identify favorable outcome predictors and combine them into a multimodal model. DESIGN: Retrospective analysis of prospectively collected data (January 2016 to June 2021). SETTING: Two academic hospitals (Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland; Brigham and Women's Hospital, Boston, MA). PATIENTS: Four hundred ninety-nine consecutive comatose adults admitted after CA. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: CA variables (initial rhythm, time to return of spontaneous circulation), clinical examination (Full Outline of UnResponsiveness [FOUR] score at 72 hr, early myoclonus), electroencephalography (EEG) (reactivity, continuity, epileptiform features, and prespecified highly malignant patterns), somatosensory-evoked potentials, quantified pupillometry, and serum neuron-specific enolase (NSE) were retrieved. Neurologic outcome was assessed at 3 months using Cerebral Performance Category (CPC); 1 and 2 were considered as favorable outcome. Predictive performance of each variable toward favorable outcomes were calculated, and most discriminant items were combined to obtain a multimodal prognostic score, using multivariable ordinal logistic regression, receiving operator characteristic curves, and cross-validation. Our analysis identified a prognostic score including six modalities (1 point each): 1) early (12-36 hr) EEG not highly malignant, 2) early EEG background reactivity, 3) late (36-72 hr) EEG background reactivity and 4) continuity, 5) peak serum NSE within 48 hours less than or equal to 41 µg/L, and 6) FOUR score greater than or equal to 5 at 72 hours. At greater than or equal to 4 out of 6 points, sensitivity for CPC 1-2 was 97.5% (95% CI, 92.9-99.5%) and accuracy was 77.5% (95% CI, 72.7-81.8%); area under the curve was 0.88 (95% CI, 0.85-0.91). The score showed similar performances in the validation cohort. CONCLUSIONS: This study describes and externally validates a multimodal score, including clinical, EEG and biological items available within 72 hours, showing a high performance in identifying early comatose CA survivors who will reach functional independence at 3 months.

Auditory stimulation and deep learning predict awakening from coma after cardiac arrest.

Assessing the integrity of neural functions in coma after cardiac arrest remains an open challenge. Prognostication of coma outcome relies mainly on visual expert scoring of physiological signals, which is prone to subjectivity and leaves a considerable number of patients in a 'grey zone', with uncertain prognosis. Quantitative analysis of EEG responses to auditory stimuli can provide a window into neural functions in coma and information about patients' chances of awakening. However, responses to standardized auditory stimulation are far from being used in a clinical routine due to heterogeneous and cumbersome protocols. Here, we hypothesize that convolutional neural networks can assist in extracting interpretable patterns of EEG responses to auditory stimuli during the first day of coma that are predictive of patients' chances of awakening and survival at 3 months. We used convolutional neural networks (CNNs) to model single-trial EEG responses to auditory stimuli in the first day of coma, under standardized sedation and targeted temperature management, in a multicentre and multiprotocol patient cohort and predict outcome at 3 months. The use of CNNs resulted in a positive predictive power for predicting awakening of 0.83 ± 0.04 and 0.81 ± 0.06 and an area under the curve in predicting outcome of 0.69 ± 0.05 and 0.70 ± 0.05, for patients undergoing therapeutic hypothermia and normothermia, respectively. These results also persisted in a subset of patients that were in a clinical 'grey zone'. The network's confidence in predicting outcome was based on interpretable features: it strongly correlated to the neural synchrony and complexity of EEG responses and was modulated by independent clinical evaluations, such as the EEG reactivity, background burst-suppression or motor responses. Our results highlight the strong potential of interpretable deep learning algorithms in combination with auditory stimulation to improve prognostication of coma outcome.

Benign EEG for prognostication of favorable outcome after cardiac arrest: A reappraisal.

AIM: The current EEG role for prognostication after cardiac arrest (CA) essentially aims at reliably identifying patients with poor prognosis ("highly malignant" patterns, defined by Westhall et al. in 2014). Conversely, "benign EEGs", defined by the absence of elements of "highly malignant" and "malignant" categories, has limited sensitivity in detecting good prognosis. We postulate that a less stringent "benign EEG" definition would improve sensitivity to detect patients with favorable outcomes. METHODS: Retrospectively assessing our registry of unconscious adults after CA (1.2018-8.2021), we scored EEGs within 72 h after CA using a modified "benign EEG" classification (allowing discontinuity, low-voltage, or reversed anterio-posterior amplitude development), versus Westhall's "benign EEG" classification (not allowing the former items). We compared predictive performances towards good outcome (Cerebral Performance Category 1-2 at 3 months), using 2x2 tables (and binomial 95% confidence intervals) and proportions comparisons. RESULTS: Among 381 patients (mean age 61.9 ± 15.4 years, 104 (27.2%) females, 240 (62.9%) having cardiac origin), the modified "benign EEG" definition identified a higher number of patients with potential good outcome (252, 66%, vs 163, 43%). Sensitivity of the modified EEG definition was 0.97 (95% CI: 0.92-0.97) vs 0.71 (95% CI: 0.62-0.78) (p < 0.001). Positive predictive values (PPV) were 0.53 (95% CI: 0.46-0.59) versus 0.59 (95% CI: 0.51-0.67; p = 0.17). Similar statistics were observed at definite recording times, and for survivors. DISCUSSION: The modified "benign EEG" classification demonstrated a markedly higher sensitivity towards favorable outcome, with minor impact on PPV. Adaptation of "benign EEG" criteria may improve efficient identification of patients who may reach a good outcome.

Treatment-emergent adverse events and antiseizure medication actual drug load.

OBJECTIVE: The correlation between treatment-emergent adverse events (TEAE) and antiseizure medication (ASM) drug load is a controversial topic. Previous studies used daily defined dosage (DDD) to measure drug load. We aim to assess if ASM adjusted to body weight and plasma levels were associated with TEAE. METHODS: We analyzed clinical visits of a trial on therapeutic drug monitoring in outpatients with epilepsy. TEAE, treatment, and its changes, as well as ASM plasma levels, were recorded at each visit. Each medication level was stratified according to its position in relation to its proposed reference range (below, in the lower half, upper half, or above). RESULTS: We analyzed 424 visits (151 participants). Treatment-emergent adverse events were reported in 84 (20%) visits. There was no significant difference when comparing visits with TEAE with those without TEAE in terms of ASM drug load (calculated with DDD), corrected for body weight, their changes since the last visit, as well as summed plasma levels compared to reference ranges. SIGNIFICANCE: Actual drug load seems not to represent a major determinant of TEAE recorded during routine visits, even when accounting thoroughly for the patient's exposure to the treatment. The use of structured questionnaires and neuropsychometric tests may assess more accurately the potential consequences of drug loads.

Delirium in Adults With COVID-19-Related Acute Respiratory Distress Syndrome: Comparison With Other Etiologies.

BACKGROUND AND OBJECTIVES: Neurologic complications have been associated with COVID-19, including delirium. Such complications have been reported to be frequent among intensive care unit (ICU)-admitted patients. We hypothesized that the rate of neurologic complications would be higher in COVID-19 associated acute respiratory distress syndrome (ARDS) than those who develop ARDS from a different cause. METHODS: We conducted a retrospective cohort study in the adult ICU of Lausanne University Hospital, including all consecutive patients fulfilling the Berlin criteria for ARDS hospitalized between December 2017 and June 2021, stratifying exposure between COVID-19 or not. The primary outcome was delirium onset during ICU stay, defined by the confusion assessment method (CAM-ICU). Exploratory outcomes included development of neurologic complications of the central nervous system (stroke, hemorrhage, and vasculitis), critical illness weakness, and 30- and 180-day all-cause mortality. RESULTS: Three hundred eleven patients were included in the study (253 with COVID-19 and 58 with other causes) and CAM-ICU could be assessed in 231 (74.3% in COVID-19 vs 74.1% in non-COVID-19). The proportion of patients developing delirium was similar in patients with COVID-19 and controls in univariate comparison (69.1% vs 60.5%, p = 0.246). Yet, patients with COVID-19 had a higher body mass index, lower ICU severity, longer mechanical ventilation, and higher sedation doses (propofol and dexmedetomidine). After adjusting for these factors in a multivariable analysis, the risk of delirium remained comparable across groups (adjusted OR [95% CI]: 0.86 [0.35-2.1]). Similarly, COVID-19-related ARDS had no effect on all-cause mortality at 30 days (adjusted OR: 0.87 [0.39-1.92]) and 180 days (adjusted OR: 0.67 [0.33-1.35]). Finally, neurologic complications affecting the CNS (adjusted OR: 1.15 [0.25-5.29]) and critical illness weakness (adjusted OR: 2.99 [0.97-9.1]) were not higher in the COVID-19 group. DISCUSSION: Compared with other etiologies, patients with COVID-19 did not have higher incidence of delirium and other neurologic complications, after accounting for underlying disease severity in patients with ARDS. Management of COVID-19-associated ARDS needed longer invasive ventilation and higher sedation, which could explain higher rates of delirium in uncontrolled studies.

Diagnostic and prognostic EEG analysis of critically ill patients: A deep learning study.

Visual interpretation of electroencephalography (EEG) is time consuming, may lack objectivity, and is restricted to features detectable by a human. Computer-based approaches, especially deep learning, could potentially overcome these limitations. However, most deep learning studies focus on a specific question or a single pathology. Here we explore the potential of deep learning for EEG-based diagnostic and prognostic assessment of patients with acute consciousness impairment (ACI) of various etiologies. EEGs from 358 adults from a randomized controlled trial (CERTA, NCT03129438) were retrospectively analyzed. A convolutional neural network was used to predict the clinical outcome (based either on survival or on best cerebral performance category) and to determine the etiology (four diagnostic categories). The largest probability output served as marker for the confidence of the network in its prediction ("certainty factor"); we also systematically compared the predictions with raw EEG data, and used a visualization algorithm (Grad-CAM) to highlight discriminative patterns. When all patients were considered, the area under the receiver operating characteristic curve (AUC) was 0.721 for predicting survival and 0.703 for predicting the outcome based on best CPC; for patients with certainty factor ≥ 60 % the AUCs increased to 0.776 and 0.755 respectively; and for certainty factor ≥ 75 % to 0.852 and 0.879. The accuracy for predicting the etiology was 54.5 %; the accuracy increased to 67.7 %, 70.3 % and 84.1 % for patients with certainty factor of 50 %, 60 % and 75 % respectively. Visual analysis showed that the network learnt EEG patterns typically recognized by human experts, and suggested new criteria. This work demonstrates for the first time the potential of deep learning-based EEG analysis in critically ill patients with various etiologies of ACI. Certainty factor and post-hoc correlation of input data with prediction help to better characterize the method and pave the route for future implementations in clinical routine.

Refractory Status Epilepticus: Risk Factors and Analysis of Intubation in the Multicenter SENSE Registry.

BACKGROUND AND OBJECTIVES: Refractory status epilepticus (RSE) bears significant morbidity and mortality. Therapy escalation and in some cases intubation are recommended. Most existing studies are retrospective and focus on intensive care units. We aimed to describe routine-care management and analyze determinants of RSE development and outcomes in a large multicenter cohort. METHODS: We performed post hoc analysis of an observational, cohort study, which prospectively registered all consecutive adults with SE at 9 centers from 3 central European countries. Only incident episodes were included. Ongoing SE despite 2 antiseizure medications was defined as RSE. Patients intubated during first-line or second-line treatments were excluded. Variables investigated included demographics, severity (Status Epilepticus Severity Score), etiology, and guideline-compliant treatment (defined as fixed minimum doses). Outcome parameters included survival and mRS at baseline, and discharge (good: 0-2, or absence of worsening compared with prehospitalization). RESULTS: Among 1,179 SE episodes from 1,049 adults, 996 patients were eligible (median age: 70 years, 52% female), of which 545 (54.7%) developed RSE. RSE was associated with higher baseline mRS (p < 0.001) and treatment deviation from guidelines (p < 0.001, OR 0.09; 95% CI 0.06-0.1). Good outcomes were observed in 52.7% of refractory patients, correlating with lower status epilepticus severity (p < 0.001), absence of acute etiology (p < 0.001, OR 0.5; 95% CI 0.3-0.7), adequate first-line benzodiazepine dose (p < 0.001, OR 2.5; 95% CI 1.6-4.0), shorter durations of SE and hospitalization (both p < 0.001), and lack of intubation (p < 0.001, OR 0.4; 95% CI 0.3-0.6). Most (71.7%) refractory patients were not intubated. Intubation was associated with younger age (p = 0.006), more severe consciousness disturbances (p < 0.001, OR 3.2; 95% CI 2.1-4.8), more severe SE types (p < 0.001), higher severity score (p = 0.009), acute etiologies (p = 0.01, OR 1.6; 95% CI 1.1-2.4), more antiseizure medications (p < 0.001), initial treatment after shorter latency (p = 0.01), worse outcomes (p < 0.001, OR 0.4; 95% CI 0.4-0.6), and longer hospitalizations (p < 0.001). DISCUSSION: Variables associated with RSE were identified, some potentially preventable. More than 70% of RSE were treated without intubation, suggesting that focal RSE without deep impairment of consciousness, in older patients, may be successfully treated outside ICUs. TRIAL REGISTRATION INFORMATION: Original cohort study registered at the German Clinical Trials Register (DRKS00000725).

Continuous versus routine EEG in patients after cardiac arrest: Analysis of a randomized controlled trial (CERTA).

BACKGROUND: Electroencephalography (EEG) is essential to assess prognosis in patients after cardiac arrest (CA). Use of continuous EEG (cEEG) is increasing in critically-ill patients, but it is more resource-consuming than routine EEG (rEEG). Observational studies did not show a major impact of cEEG versus rEEG on outcome, but randomized studies are lacking. METHODS: We analyzed data of the CERTA trial (NCT03129438), including comatose adults after CA undergoing cEEG (30-48 hours) or two rEEG (20-30 minutes each). We explored correlations between recording EEG type and mortality (primary outcome), or Cerebral Performance Categories (CPC, secondary outcome), assessed blindly at 6 months, using uni- and multivariable analyses (adjusting for other prognostic variables showing some imbalance across groups). RESULTS: We analyzed 112 adults (52 underwent rEEG, 60 cEEG,); 31 (27.7%) were women; 68 (60.7%) patients died. In univariate analysis, mortality (rEEG 59%, cEEG 65%, p = 0.318) and good outcome (CPC 1-2; rEEG 33%, cEEG 27%, p = 0.247) were comparable across EEG groups. This did not change after multiple logistic regressions, adjusting for shockable rhythm, time to return of spontaneous circulation, serum neuron-specific enolase, EEG background reactivity, regarding mortality (cEEG vs rEEG: OR 1.60, 95% CI 0.43-5.83, p = 0.477), and good outcome (OR 0.51, 95% CI 0.14-1.90, p = 0.318). CONCLUSION: This analysis suggests that cEEG or repeated rEEG are related to comparable outcomes of comatose patients after CA. Pending a prospective, large randomized trial, this finding does not support the routine use of cEEG for prognostication in this setting. TRIAL REGISTRATION: Continuous EEG Randomized Trial in Adults (CERTA); NCT03129438; July 25, 2019.