Comparative analysis of patients with new onset refractory status epilepticus preceded by fever (febrile infection-related epilepsy syndrome) versus without prior fever: An interim analysis

Febrile infection-related epilepsy syndrome (FIRES) is a subset of new onset refractory status epilepticus (NORSE) that involves a febrile infection prior to the onset of the refractory status epilepticus. It is unclear whether FIRES and non-FIRES NORSE are distinct conditions. Here, we compare 34 patients with FIRES to 30 patients with non-FIRES NORSE for demographics, clinical features, neuroimaging, and outcomes. Because patients with FIRES were younger than patients with non-FIRES NORSE (median = 28 vs. 48 years old, p = .048) and more likely cryptogenic (odds ratio = 6.89), we next ran a regression analysis using age or etiology as a covariate. Respiratory and gastrointestinal prodromes occurred more frequently in FIRES patients, but no difference was found for non-infection-related prodromes. Status epilepticus subtype, cerebrospinal fluid (CSF) and magnetic resonance imaging findings, and outcomes were similar. However, FIRES cases were more frequently cryptogenic; had higher CSF interleukin 6, CSF macrophage inflammatory protein-1 alpha (MIP-1a), and serum chemokine ligand 2 (CCL2) levels; and received more antiseizure medications and immunotherapy. After controlling for age or etiology, no differences were observed in presenting symptoms and signs or inflammatory biomarkers, suggesting that FIRES and non-FIRES NORSE are very similar conditions.

Interrater Reliability of Expert Electroencephalographers Identifying Seizures and Rhythmic and Periodic Patterns in EEGs

BACKGROUND AND OBJECTIVES

The validity of brain monitoring using electroencephalography (EEG), particularly to guide care in patients with acute or critical illness, requires that experts can reliably identify seizures and other potentially harmful rhythmic and periodic brain activity, collectively referred to as "ictal-interictal-injury continuum" (IIIC). Previous interrater reliability (IRR) studies are limited by small samples and selection bias. This study was conducted to assess the reliability of experts in identifying IIIC.

METHODS

This prospective analysis included 30 experts with subspecialty clinical neurophysiology training from 18 institutions. Experts independently scored varying numbers of ten-second EEG segments as "seizure (SZ)," "lateralized periodic discharges (LPDs)," "generalized periodic discharges (GPDs)," "lateralized rhythmic delta activity (LRDA)," "generalized rhythmic delta activity (GRDA)," or "other." EEGs were performed for clinical indications at Massachusetts General Hospital between 2006 and 2020. Primary outcome measures were pairwise IRR (average percent agreement [PA] between pairs of experts) and majority IRR (average PA with group consensus) for each class and beyond chance agreement (κ). Secondary outcomes were calibration of expert scoring to group consensus, and latent trait analysis to investigate contributions of bias and noise to scoring variability.

RESULTS

Among 2,711 EEGs, 49% were from women, and the median (IQR) age was 55 (41) years. In total, experts scored 50,697 EEG segments; the median [range] number scored by each expert was 6,287.5 [1,002, 45,267]. Overall pairwise IRR was moderate (PA 52%, κ 42%), and majority IRR was substantial (PA 65%, κ 61%). Noise-bias analysis demonstrated that a single underlying receiver operating curve can account for most variation in experts' false-positive vs true-positive characteristics (median [range] of variance explained ([Formula: see text]): 95 [93, 98]%) and for most variation in experts' precision vs sensitivity characteristics ([Formula: see text]: 75 [59, 89]%). Thus, variation between experts is mostly attributable not to differences in expertise but rather to variation in decision thresholds.

DISCUSSION

Our results provide precise estimates of expert reliability from a large and diverse sample and a parsimonious theory to explain the origin of disagreements between experts. The results also establish a standard for how well an automated IIIC classifier must perform to match experts.

CLASSIFICATION OF EVIDENCE

This study provides Class II evidence that an independent expert review reliably identifies ictal-interictal injury continuum patterns on EEG compared with expert consensus.

Development of Expert-Level Classification of Seizures and Rhythmic and Periodic Patterns During EEG Interpretation

BACKGROUND AND OBJECTIVES

Seizures (SZs) and other SZ-like patterns of brain activity can harm the brain and contribute to in-hospital death, particularly when prolonged. However, experts qualified to interpret EEG data are scarce. Prior attempts to automate this task have been limited by small or inadequately labeled samples and have not convincingly demonstrated generalizable expert-level performance. There exists a critical unmet need for an automated method to classify SZs and other SZ-like events with expert-level reliability. This study was conducted to develop and validate a computer algorithm that matches the reliability and accuracy of experts in identifying SZs and SZ-like events, known as "ictal-interictal-injury continuum" (IIIC) patterns on EEG, including SZs, lateralized and generalized periodic discharges (LPD, GPD), and lateralized and generalized rhythmic delta activity (LRDA, GRDA), and in differentiating these patterns from non-IIIC patterns.

METHODS

We used 6,095 scalp EEGs from 2,711 patients with and without IIIC events to train a deep neural network, SPaRCNet, to perform IIIC event classification. Independent training and test data sets were generated from 50,697 EEG segments, independently annotated by 20 fellowship-trained neurophysiologists. We assessed whether SPaRCNet performs at or above the sensitivity, specificity, precision, and calibration of fellowship-trained neurophysiologists for identifying IIIC events. Statistical performance was assessed by the calibration index and by the percentage of experts whose operating points were below the model's receiver operating characteristic curves (ROCs) and precision recall curves (PRCs) for the 6 pattern classes.

RESULTS

SPaRCNet matches or exceeds most experts in classifying IIIC events based on both calibration and discrimination metrics. For SZ, LPD, GPD, LRDA, GRDA, and "other" classes, SPaRCNet exceeds the following percentages of 20 experts-ROC: 45%, 20%, 50%, 75%, 55%, and 40%; PRC: 50%, 35%, 50%, 90%, 70%, and 45%; and calibration: 95%, 100%, 95%, 100%, 100%, and 80%, respectively.

DISCUSSION

SPaRCNet is the first algorithm to match expert performance in detecting SZs and other SZ-like events in a representative sample of EEGs. With further development, SPaRCNet may thus be a valuable tool for an expedited review of EEGs.

CLASSIFICATION OF EVIDENCE

This study provides Class II evidence that among patients with epilepsy or critical illness undergoing EEG monitoring, SPaRCNet can differentiate (IIIC) patterns from non-IIIC events and expert neurophysiologists.

Putative roles for homeostatic plasticity in epileptogenesis

Homeostatic plasticity allows neural circuits to maintain an average activity level while preserving the ability to learn new associations and efficiently transmit information. This dynamic process usually protects the brain from excessive activity, like seizures. However, in certain contexts, homeostatic plasticity might produce seizures, either in response to an acute provocation or more chronically as a driver of epileptogenesis. Here, we review three seizure conditions in which homeostatic plasticity likely plays an important role: acute drug withdrawal seizures, posttraumatic or disconnection epilepsy, and cyclic seizures. Identifying the homeostatic mechanisms active at different stages of development and in different circuits could allow better targeting of therapies, including determining when neuromodulation might be most effective, proposing ways to prevent epileptogenesis, and determining how to disrupt the cycle of recurring seizure clusters.

Epileptiform Abnormalities in Acute Ischemic Stroke: Impact on Clinical Management and Outcomes

PURPOSE

Studies examining seizures (Szs) and epileptiform abnormalities (EAs) using continuous EEG in acute ischemic stroke (AIS) are limited. Therefore, we aimed to describe the prevalence of Sz and EA in AIS, its impact on anti-Sz drug management, and association with discharge outcomes.

METHODS

The study included 132 patients with AIS who underwent continuous EEG monitoring >6 hours. Continuous EEG was reviewed for background, Sz and EA (lateralized periodic discharges [LPD], generalized periodic discharges, lateralized rhythmic delta activity, and sporadic epileptiform discharges). Relevant clinical, demographic, and imaging factors were abstracted to identify risk factors for Sz and EA. Outcomes included all-cause mortality, functional outcome at discharge (good outcome as modified Rankin scale of 0-2 and poor outcome as modified Rankin scale of 3-6) and changes to anti-Sz drugs (escalation or de-escalation).

RESULTS

The frequency of Sz was 7.6%, and EA was 37.9%. Patients with Sz or EA were more likely to have cortical involvement (84.6% vs. 67.5% P = 0.028). Among the EAs, the presence of LPD was associated with an increased risk of Sz (25.9% in LPD vs. 2.9% without LPD, P = 0.001). Overall, 21.2% patients had anti-Sz drug changes because of continuous EEG findings, 16.7% escalation and 4.5% de-escalation. The presence of EA or Sz was not associated with in-hospital mortality or discharge functional outcomes.

CONCLUSIONS

Despite the high incidence of EA, the rate of Sz in AIS is relatively lower and is associated with the presence of LPDs. These continuous EEG findings resulted in anti-Sz drug changes in one-fifth of the cohort. Epileptiform abnormality and Sz did not affect mortality or discharge functional outcomes.

Deep active learning for Interictal Ictal Injury Continuum EEG patterns

OBJECTIVES

Seizures and seizure-like electroencephalography (EEG) patterns, collectively referred to as "ictal interictal injury continuum" (IIIC) patterns, are commonly encountered in critically ill patients. Automated detection is important for patient care and to enable research. However, training accurate detectors requires a large labeled dataset. Active Learning (AL) may help select informative examples to label, but the optimal AL approach remains unclear.

METHODS

We assembled >200,000 h of EEG from 1,454 hospitalized patients. From these, we collected 9,808 labeled and 120,000 unlabeled 10-second EEG segments. Labels included 6 IIIC patterns. In each AL iteration, a Dense-Net Convolutional Neural Network (CNN) learned vector representations for EEG segments using available labels, which were used to create a 2D embedding map. Nearest-neighbor label spreading within the embedding map was used to create additional pseudo-labeled data. A second Dense-Net was trained using real- and pseudo-labels. We evaluated several strategies for selecting candidate points for experts to label next. Finally, we compared two methods for class balancing within queries: standard balanced-based querying (SBBQ), and high confidence spread-based balanced querying (HCSBBQ).

RESULTS

Our results show: 1) Label spreading increased convergence speed for AL. 2) All query criteria produced similar results to random sampling. 3) HCSBBQ query balancing performed best. Using label spreading and HCSBBQ query balancing, we were able to train models approaching expert-level performance across all pattern categories after obtaining ∼7000 expert labels.

CONCLUSION

Our results provide guidance regarding the use of AL to efficiently label large EEG datasets in critically ill patients.

Association of Epileptiform Abnormality on Electroencephalography with Development of Epilepsy After Acute Brain Injury

BACKGROUND/OBJECTIVES

Epileptiform abnormalities (EA) on continuous electroencephalography (cEEG) are associated with increased risk of acute seizures; however, data on their association with development of long-term epilepsy are limited. We aimed to investigate the association of EA in patients with acute brain injury (ABI): ischemic or hemorrhagic stroke, traumatic brain injury, encephalitis, or posterior reversible encephalopathy syndrome, and subsequent development of epilepsy.

METHODS

This was a retrospective, single-center study of patients with ABI who had at least 6 hours of cEEG during the index admission between 1/1/2017 and 12/31/2018 and at least 12 months of follow-up. We compared patients with EAs; defined as lateralized periodic discharges (LPDs), lateralized rhythmic delta activity (LRDA), generalized periodic discharges (GPDs), and sporadic interictal epileptiform discharges (sIEDs) to patients without EAs on cEEG. The primary outcome was the new development of epilepsy, defined as the occurrence of spontaneous clinical seizures following hospital discharge. Secondary outcomes included time to development of epilepsy and use of anti-seizure medications (ASMs) at the time of last follow-up visit.

RESULTS

One hundred and one patients with ABI met study inclusion criteria. Thirty-one patients (30.7%) had EAs on cEEG. The median (IQR) time to cEEG was 2 (1-5) days. During a median (IQR) follow-up period of 19.1 (16.2-24.3) months, 25.7% of patients developed epilepsy; the percentage of patients who developed epilepsy was higher in those with EAs compared to those without EAs (41.9% vs. 18.6%, p = 0.025). Patients with EAs were more likely to be continued on ASMs during follow-up compared to patients without EAs (67.7% vs. 38.6%, p = 0.009). Using multivariable Cox regression analysis, after adjusting for age, mental status, electrographic seizures on cEEG, sex, ABI etiology, and ASM treatment on discharge, patients with EAs had a significantly increased risk of developing epilepsy compared to patients without EA (hazard ratio 3.39; 95% CI 1.39-8.26; p = 0.007).

CONCLUSIONS

EAs on cEEG in patients with ABI are associated with a greater than three-fold increased risk of new-onset epilepsy. cEEG findings in ABI may therefore be a useful risk stratification tool for assessing long-term risk of seizures and serve as a biomarker for new-onset epilepsy.

Assessment of the Validity of the 2HELPS2B Score for Inpatient Seizure Risk Prediction

Importance

Seizure risk stratification is needed to boost inpatient seizure detection and to improve continuous electroencephalogram (cEEG) cost-effectiveness. 2HELPS2B can address this need but requires validation.

Objective

To use an independent cohort to validate the 2HELPS2B score and develop a practical guide for its use.

Design, Setting, and Participants

This multicenter retrospective medical record review analyzed clinical and EEG data from patients 18 years or older with a clinical indication for cEEG and an EEG duration of 12 hours or longer who were receiving consecutive cEEG at 6 centers from January 2012 to January 2019. 2HELPS2B was evaluated with the validation cohort using the mean calibration error (CAL), a measure of the difference between prediction and actual results. A Kaplan-Meier survival analysis was used to determine the duration of EEG monitoring to achieve a seizure risk of less than 5% based on the 2HELPS2B score calculated on first- hour (screening) EEG. Participants undergoing elective epilepsy monitoring and those who had experienced cardiac arrest were excluded. No participants who met the inclusion criteria were excluded.

Main Outcomes and Measures

The main outcome was a CAL error of less than 5% in the validation cohort.

Results

The study included 2111 participants (median age, 51 years; 1113 men [52.7%]; median EEG duration, 48 hours) and the primary outcome was met with a validation cohort CAL error of 4.0% compared with a CAL of 2.7% in the foundational cohort (P = .13). For the 2HELPS2B score calculated on only the first hour of EEG in those without seizures during that hour, the CAL error remained at less than 5.0% at 4.2% and allowed for stratifying patients into low- (2HELPS2B = 0; <5% risk of seizures), medium- (2HELPS2B = 1; 12% risk of seizures), and high-risk (2HELPS2B, ≥2; risk of seizures, >25%) groups. Each of the categories had an associated minimum recommended duration of EEG monitoring to achieve at least a less than 5% risk of seizures, a 2HELPS2B score of 0 at 1-hour screening EEG, a 2HELPS2B score of 1 at 12 hours, and a 2HELPS2B score of 2 or greater at 24 hours.

Conclusions and Relevance

In this study, 2HELPS2B was validated as a clinical tool to aid in seizure detection, clinical communication, and cEEG use in hospitalized patients. In patients without prior clinical seizures, a screening 1-hour EEG that showed no epileptiform findings was an adequate screen. In patients with any highly epileptiform EEG patterns during the first hour of EEG (ie, a 2HELPS2B score of ≥2), at least 24 hours of recording is recommended.

A standardized nomenclature for spectrogram EEG patterns: Inter-rater agreement and correspondence with common intensive care unit EEG patterns

OBJECTIVE

To determine the inter-rater agreement (IRA) of a standardized nomenclature for EEG spectrogram patterns, and to estimate the probability distribution of ictal-interictal continuum (IIC) patterns vs. other EEG patterns within each category in this nomenclature.

METHODS

We defined seven spectrogram categories: "Solid Flames", "Irregular Flames", "Broadband-monotonous", "Narrowband-monotonous", "Stripes", "Low power", and "Artifact". Ten electroencephalographers scored 115 spectrograms and the corresponding raw EEG samples. Gwet's agreement coefficient was used to calculate IRA.

RESULTS

Solid Flames represented seizures or IIC patterns 69.4% of the time. Irregular Flames represented seizures or IIC patterns 38.7% of the time. Broadband-monotonous primarily corresponded with seizures or IIC (54.3%) and Narrowband-monotonous with focal or generalized slowing (43.8%). Stripes were associated with burst-suppression (37.2%) and generalized suppression (34.4%). Low Power category was associated with generalized suppression (94%). There was "near perfect" agreement for Solid Flames (κ = 94.36), Low power (κ = 92.61), and Artifact (κ = 93.72). There was "substantial agreement" for all other categories (κ = 74.65-79.49).

CONCLUSIONS

This EEG spectrogram nomenclature has high IRA among electroencephalographers.

SIGNIFICANCE

The nomenclature can be a useful tool for EEG screening. Future studies are needed to determine if using this nomenclature shortens time to IIC identification, and how best to use it in practice to reduce time to intervention.

Validation of the 2HELPS2B Seizure Risk Score in Acute Brain Injury Patients

BACKGROUND AND OBJECTIVE

Seizures are common after traumatic brain injury (TBI), aneurysmal subarachnoid hemorrhage (aSAH), subdural hematoma (SDH), and non-traumatic intraparenchymal hemorrhage (IPH)-collectively defined herein as acute brain injury (ABI). Most seizures in ABI are subclinical, meaning that they are only detectable with EEG. A method is required to identify patients at greatest risk of seizures and thereby in need of prolonged continuous EEG monitoring. 2HELPS2B is a simple point system developed to address this need. 2HELPS2B estimates seizure risk for hospitalized patients using five EEG findings and one clinical finding (pre-EEG seizure). The initial 2HELPS2B study did not specifically assess the ABI subpopulation. In this study, we aim to validate the 2HELPS2B score in ABI and determine its relative predictive accuracy compared to a broader set of clinical and electrographic factors.

METHODS

We queried the Critical Care EEG Monitoring Research Consortium database for ABI patients age ≥ 18 with > 6 h of continuous EEG monitoring; data were collected between February 2013 and November 2018. The primary outcome was electrographic seizure. Clinical factors considered were age, coma, encephalopathy, ABI subtype, and acute suspected or confirmed pre-EEG clinical seizure. Electrographic factors included 18 EEG findings. Predictive accuracy was assessed using a machine-learning paradigm with area under the receiver operator characteristic (ROC) curve as the primary outcome metric. Three models (clinical factors alone, EEG factors alone, EEG and clinical factors combined) were generated using elastic-net logistic regression. Models were compared to each other and to the 2HELPS2B model. All models were evaluated by calculating the area under the curve (AUC) of a ROC analysis and then compared using permutation testing of AUC with bootstrapping to generate confidence intervals.

RESULTS

A total of 1528 ABI patients were included. Total seizure incidence was 13.9%. Seizure incidence among ABI subtype varied: IPH 17.2%, SDH 19.1%, aSAH 7.6%, TBI 9.2%. Age ≥ 65 (p = 0.015) and pre-cEEG acute clinical seizure (p < 0.001) positively affected seizure incidence. Clinical factors AUC = 0.65 [95% CI 0.60-0.71], EEG factors AUC = 0.82 [95% CI 0.77-0.87], and EEG and clinical factors combined AUC = 0.84 [95% CI 0.80-0.88]. 2HELPS2B AUC = 0.81 [95% CI 0.76-0.85]. The 2HELPS2B AUC did not differ from EEG factors (p = 0.51), or EEG and clinical factors combined (p = 0.23), but was superior to clinical factors alone (p < 0.001).

CONCLUSIONS

Accurate seizure risk forecasting in ABI requires the assessment of EEG markers of pathologic electro-cerebral activity (e.g., sporadic epileptiform discharges and lateralized periodic discharges). The 2HELPS2B score is a reliable and simple method to quantify these EEG findings and their associated risk of seizure.

Comparison of machine learning models for seizure prediction in hospitalized patients

Objective

To compare machine learning methods for predicting inpatient seizures risk and determine the feasibility of 1‐h screening EEG to identify low‐risk patients (<5% seizures risk in 48 h).

Methods

The Critical Care EEG Monitoring Research Consortium (CCEMRC) multicenter database contains 7716 continuous EEGs (cEEG). Neural networks (NN), elastic net logistic regression (EN), and sparse linear integer model (RiskSLIM) were trained to predict seizures. RiskSLIM was used previously to generate 2HELPS2B model of seizure predictions. Data were divided into training (60% for model fitting) and evaluation (40% for model evaluation) cohorts. Performance was measured using area under the receiver operating curve (AUC), mean risk calibration (CAL), and negative predictive value (NPV). A secondary analysis was performed using Monte Carlo simulation (MCS) to normalize all EEG recordings to 48 h and use only the first hour of EEG as a “screening EEG” to generate predictions.

Results

RiskSLIM recreated the 2HELPS2B model. All models had comparable AUC: evaluation cohort (NN: 0.85, EN: 0.84, 2HELPS2B: 0.83) and MCS (NN: 0.82, EN; 0.82, 2HELPS2B: 0.81) and NPV (absence of seizures in the group that the models predicted to be low risk): evaluation cohort (NN: 97%, EN: 97%, 2HELPS2B: 97%) and MCS (NN: 97%, EN: 99%, 2HELPS2B: 97%). 2HELPS2B model was able to identify the largest proportion of low‐risk patients.

Interpretation

For seizure risk stratification of hospitalized patients, the RiskSLIM generated 2HELPS2B model compares favorably to the complex NN and EN generated models. 2HELPS2B is able to accurately and quickly identify low‐risk patients with only a 1‐h screening EEG.

Sensitivity of quantitative EEG for seizure identification in the intensive care unit

OBJECTIVE

To evaluate the sensitivity of quantitative EEG (QEEG) for electrographic seizure identification in the intensive care unit (ICU).

METHODS

Six-hour EEG epochs chosen from 15 patients underwent transformation into QEEG displays. Each epoch was reviewed in 3 formats: raw EEG, QEEG + raw, and QEEG-only. Epochs were also analyzed by a proprietary seizure detection algorithm. Nine neurophysiologists reviewed raw EEGs to identify seizures to serve as the gold standard. Nine other neurophysiologists with experience in QEEG evaluated the epochs in QEEG formats, with and without concomitant raw EEG. Sensitivity and false-positive rates (FPRs) for seizure identification were calculated and median review time assessed.

RESULTS

Mean sensitivity for seizure identification ranged from 51% to 67% for QEEG-only and 63%-68% for QEEG + raw. FPRs averaged 1/h for QEEG-only and 0.5/h for QEEG + raw. Mean sensitivity of seizure probability software was 26.2%-26.7%, with FPR of 0.07/h. Epochs with the highest sensitivities contained frequent, intermittent seizures. Lower sensitivities were seen with slow-frequency, low-amplitude seizures and epochs with rhythmic or periodic patterns. Median review times were shorter for QEEG (6 minutes) and QEEG + raw analysis (14.5 minutes) vs raw EEG (19 minutes; p = 0.00003).

CONCLUSIONS

A panel of QEEG trends can be used by experts to shorten EEG review time for seizure identification with reasonable sensitivity and low FPRs. The prevalence of false detections confirms that raw EEG review must be used in conjunction with QEEG. Studies are needed to identify optimal QEEG trend configurations and the utility of QEEG as a screening tool for non-EEG personnel.

CLASSIFICATION OF EVIDENCE REVIEW

This study provides Class II evidence that QEEG + raw interpreted by experts identifies seizures in patients in the ICU with a sensitivity of 63%-68% and FPR of 0.5 seizures per hour.