Aim:To develop an electroencephalography (EEG) based machine learning model to identify diffuse cerebral edema in post-cardiac arrest patients and to evaluate its ability to predict edema prior to radiographic detection. Methods:We performed a retrospective, single-center cohort study of adult patients resuscitated from cardiac arrest (2016-2024) who underwent both neuroimaging and EEG monitoring as part of routine clinical care. Machine learning models using Transformer and Long Short-Term Memory architectures were trained to detect diffuse cerebral edema from 4- and 8-hour EEG segments obtained >24 hours after arrest. The best performing detection model was then evaluated for its ability to predictdiffuse cerebral edema using EEG segments preceding radiographic recognition in patients who ultimately developed edema, compared with matched referents without edema (matched on age, sex, witnessed arrest, and EEG timing). Model performance was assessed using Area Under the Curve (AUC), accuracy, sensitivity, and specificity. Results:Among 124 patients in the detection model, median age was 53 years, and 74 (59.7%) were male. Sixty-five patients (52.4%) developed diffuse cerebral edema. The best-performing detection model, a Transformer using 4-hour EEG segments, achieved strong performance (AUC 80.0%, accuracy 80.0%, sensitivity 80.0%, specificity 90.0%). In a secondary analysis of 19 patients with diffuse cerebral edema and 19 matched referents, the top-performing prediction model used 8-hour EEG segments (AUC 92.2%, accuracy 90.6%, sensitivity 100%, specificity 88.9%). Conclusion:Diffuse cerebral edema can be identified in survivors of cardiac arrest using machine learning models applied to routine EEG data. In this study, a Transformer based approach demonstrated superior performance for both detection of established edema and identification of EEG patterns that preceded radiographic recognition than LSTM. With validation in larger and independent cohorts, this strategy may enable earlier recognition of evolving cerebral edema during standard EEG monitoring and support timely interventions to mitigate secondary brain injury.

Accurate neuroprognostication of cardiac arrest survivors who are initially comatose after restoration of spontaneous circulation is crucial for guiding patient management. Because hypoxic-ischaemic injury is typically diffuse, damage to a network of brain regions is likely involved in the patient's disorder of consciousness. To quantify these complex brain network changes, graph theoretical methods were applied. We hypothesize that structural connectivity metrics may provide insights into which patients will likely recover consciousness. Eighteen comatose patients (50 ± 22 years, 44% male) and four healthy participants (40 ± 20 years, 50% male) underwent multi-shell high angular diffusion MRI as part of a prospective study. Structural connectivity matrices were constructed using probabilistic tractography to measure the likelihood of connections between anatomical regions. Network topology alterations were quantified using clustering coefficient, global efficiency and degree. Hub index analysis was performed to explore the impact of anoxic injury on high-degree hubs. Network parameters were compared between patients with arousal recovery (AR, eye-opening to auditory or noxious stimulation) and without arousal recovery (No AR). Analyses were repeated for AR patients who achieved emergence from the minimally conscious state (EMCS) within one-year post-cardiac arrest and AR patients who did not achieve EMCS (AR'). Significant differences were observed between the Controls, AR and No AR for all four metrics (Kruskal-Wallis Tests, P < 0.05). Worsening disorders of consciousness were associated with decreasing brain complexity (Kendall's tau, P<0.01). Post-hoc testing showed Control values were significantly greater than No AR for all metrics (Wilcoxon rank sum, P < 0.05). Control values were greater than AR for all metrics (P < 0.05), except the clustering coefficient (P = 0.36). AR was significantly greater than No AR for all metrics (P < 0.05), except for the hub index (P = 0.12). Notable differences between AR' and Controls were observed for all metrics (P < 0.05), except clustering coefficient (P = 0.11). No significant differences were found between AR' and No AR groups. In contrast, for all metrics, EMCS values were not significantly different compared with the Controls but were significantly different than the No AR cohort values (P < 0.05). The hub index analysis revealed disproportionate damage to high-degree nodes such as the thalamus, putamen and precuneus, further linking topological disruption to the severity of outcomes. This study highlights the potential of graph theoretical measures of structural connectivity to guide decisions in the care of comatose cardiac arrest patients. By bridging structural connectivity with clinical outcomes, this research provides valuable insights into the neural mechanisms underlying consciousness and recovery after cardiac arrest.

OBJECTIVES:To develop a comprehensive checklist, define critical actions, and establish a minimal passing standard for adult and pediatric critical care clinicians as well as other clinicians to facilitate formative and summative assessment of brain death/death by neurologic criteria (BD/DNC) determination. DESIGN:A prespecified three-round modified Delphi consensus process to define checklist items followed by a modified Angoff standard setting process to determine critical actions and item average ratings. SETTING:Electronic surveys. SUBJECTS:Selected authors of the 2023 Pediatric and Adult BD/DNC Consensus Practice Guideline, World Brain Death Project, and experts recommended by these authors ( n = 16) participated in the Delphi panel. Neurocritical Care United Council for Neurologic Subspecialties and Accreditation Council for Graduate Medical Education examination committee members ( n = 13) participated in Angoff standard setting. INTERVENTIONS:None. MEASUREMENTS AND MAIN RESULTS:A total of 98 unique checklist items related to assessment of prerequisites (23 items), performance of the clinical examination (28 items), apnea testing (36 items), and ancillary testing (11 items) were retained by the Delphi panel. Seven items were designated as critical actions based upon Angoff panelist consensus. The remaining 91 items were assigned item average ratings. The minimum passing score for an assessment including all noncritical items was set at 89%. CONCLUSIONS:These guideline-concordant consensus checklist items, including critical actions and noncritical actions with their assigned item average ratings, may be applied selectively to simulated cases of BD/DNC determination for adults and children to determine a minimum passing score and readiness for independent practice, mitigating the risk of inaccurate BD/DNC determination among critical care clinicians. Our process for systematically defining critical actions on a behavior checklist may be replicated for simulation-based summative assessment of learners in other critical care scenarios.

Background:Large middle cerebral artery (MCA) infarctions can result in life-threatening cerebral edema. Quantitative brain atrophy may improve risk stratification for severe edema. We examined whether quantitative brain atrophy is associated with severe midline shift after large ischemic stroke and whether incorporating atrophy improves prediction beyond established clinical and radiographic predictors. Methods:This was a retrospective observational cohort study of patients with ≥½ MCA ischemic infarction, presentation within 24 hours of last known well, and at least one follow-up head CT, admitted to two academic hospitals with comprehensive stroke centers between 2006 and 2024. The study was approved by the institutional review boards of both centers. Brain atrophy was quantified as the inverse of standardized brain volume on admission head CT. The primary outcome was severe radiographic mass effect, defined as midline shift ≥5 mm on follow-up CT. The secondary outcome was in-hospital mortality. Multivariable regression models assessed associations between quantified atrophy and outcomes. Incremental prognostic value was evaluated by comparing models with and without atrophy using measures of goodness of fit, calibration, and discrimination. Results:Among 565 patients (mean age 67.5±15.7 years; 49.9% female), 223 (39.5%) developed severe mass effect. Greater atrophy was associated with lower odds of midline shift ≥5 mm (OR 0.44, 95% CI 0.34-0.58), but not with in-hospital mortality. Incorporation of atrophy significantly improved prediction of severe mass effect compared to the baseline model (likelihood ratio test χ² (1) = 41, p <0.001; AIC 703 vs. 741; BIC 733 vs. 767; AUC 0.68 vs. 0.60). Conclusions:Quantified brain atrophy is independently associated with a reduced risk of severe mass effect after large MCA stroke and improved the performance of established predictive models. Incorporation of this imaging biomarker may enhance early risk stratification, monitoring, and intervention planning for patients at risk of life-threatening cerebral edema.

Cerebral edema is a life-threatening complication of large ischemic stroke. Imaging assessment of global and hemispheric cerebrospinal fluid (CSF) volumetrics quantifies edema progression, while quantitative pupillometry provides real-time bedside assessment of neurologic decline. However, the relationship between the two and their combined value for predicting neurologic deterioration remains unclear. We conducted a retrospective study of patients with large middle cerebral artery strokes admitted to Boston Medical Center between 2019 and 2024. Eligible patients had ≥ 1 head computed tomography (CT) and ≥ 3 pupillometry measurements. Total and hemispheric CSF volumes were extracted using an automated image analysis pipeline. Average pupillometry variables, including the Neurological Pupil index (NPi) and dilation velocity, were aligned to imaging within ± 1 h and within the subsequent 24 h of each image. Associations between pupillometry and CSF volumetrics were evaluated using Spearman’s correlations and linear mixed-effects models adjusted for age, sex, and standardized baseline brain volume. Cox proportional hazards models with time-dependent covariates were used to assess the predictive value of CSF and pupillometry markers for time-to-neurologic deterioration. We compared model performance using likelihood ratio tests and time-dependent area under the curve (AUC) metrics. A total of 71 patients (mean age 66 ± 16 years; 59

STUDY OBJECTIVES:Temperature control for survivors of cardiac arrest is a complex bundled intervention with poorly defined optimal parameters. We defined high-quality temperature control initiation and evaluated the association between quality and clinical outcomes. METHODS:In this retrospective single academic center study between January 1, 2014, and July 19, 2024, consecutive out-of-hospital cardiac arrest patients treated with temperature control were identified. Patients were assigned a temperature control quality score (range 0 to 6) based on the time from hospital arrival to temperature control device initiation and the use of adjunctive pharmacologic treatment for shivering thermogenesis within 6 hours from hospital arrival. Based on the nonlinear relationship between temperature control quality and outcomes, quality was binarized into low quality (less than 3) and high quality (more than 3). The primary outcome was survival to hospital discharge and the secondary outcome was favorable neurologic outcomes, defined as a Cerebral Performance Category score of 1 to 2. We assessed the association between primary and secondary outcomes and temperature control quality using logistic regression. A sensitivity analysis using inverse probability treatment weighting, created using a propensity score, was performed to minimize measurable confounding. Standardized mean difference was used to quantify the difference between groups. RESULTS:Of the 421 patients treated with temperature control, 194 (46.1%) received high-quality temperature control. Demographic factors, arrest-related details, and postresuscitation management were similar between the low-quality and high-quality groups with an overall small effect size, except for the time from cardiac arrest to achievement of target temperature, which occurred faster in patients with high-quality temperature control (median [interquartile range] 5.4 [4.1, 8.6] versus 8.8 [7.3, 11.0] hours; standardized mean difference=0.79). High-quality temperature control was associated with increased survival to hospital discharge and favorable neurologic outcomes before and after inverse probability treatment weighting (adjusted odds ratio [95% confidence interval] 2.75 [1.53 to 5.04] and 2.05 [1.10 to 3.91] versus 2.13 [1.37 to 3.34] and 1.94 [1.16 to 3.30], respectively). CONCLUSION:In our single-center study of out-of-hospital cardiac arrest patients, high-quality temperature control was associated with improved survival and good neurologic outcomes. Temperature control parameters are likely important and may influence the neuroprotective benefit of temperature control. Prospective multicenter studies are warranted to evaluate the effect of temperature control quality on patient outcomes.

BACKGROUND AND PURPOSE:Absent pupillary light reflex (PLR) has been implicated as an indicator of poor prognosis following cardiac arrest; however, few studies report on pupil size. We evaluated the association between pupil size and reactivity immediately following return of spontaneous circulation (ROSC) and post-arrest illness severity and clinical outcomes. METHODS:In this retrospective, single-center study between 2018 and 2024, out-of-hospital cardiac arrest (OHCA) patients with early pupil size and reactivity data were identified. Dilated pupils were defined as ≥4 mm in diameter. Pupil size and reactivity were subjectively assessed by the clinical team and documented as part of the standardized note template. We evaluated the association between post-ROSC pupil size, reactivity and illness severity, clinical outcomes. RESULTS:In our cohort of 382 OHCA patients, 50.8 % had absent PLR (n = 194). Patients with absent PLR were younger, had less premorbid medical conditions, worse cardiac arrest features, and higher illness severity. The false positive rate for absent PLR was 10.9 % for mortality and 5.9 % for poor neurologic outcome. Amongst those with absent PLR, dilated pupils were present in 68 % (n = 132). Patients with absent PLR and dilated pupils were younger, had less premorbid medical conditions, were less likely to have a shockable rhythm arrest, had higher illness severity scores and worse post-arrest labs. Patients with absent PLR and dilated pupils had a higher incidence of brain death (34.8 % vs. 9.7 %, p < 0.001). Amongst patients with absent PLR, the presence of dilated pupils improved the prediction of brain death [ AUC (CI) 0.733 (0.672-0.793) vs. 0.683 (0.628-0.738), p = 0.026]. CONCLUSION:Absent PLR immediately following ROSC is associated with poor outcomes but does not preclude good outcome. Pupil size and reactivity immediately post-ROSC may help to differentiate brain injury phenotypes. Prospective work using quantitative pupillometry is important to validate our findings.

This essay describes the author’s experience with insights gained through vision loss.

AIMS:Changes in ventricular repolarisation, observed as QTc prolongation, are frequently observed following cardiac arrest. The T-peak to T-end (TpTe) interval represents a period of increased susceptibility to ventricular arrhythmia. We posit that TpTe prolongation may be associated with adverse clinical outcomes in patients resuscitated from cardiac arrest. METHODS AND RESULTS:We included patients aged ≥18 years with both out-of-hospital and in-hospital cardiac arrest following return of spontaneous circulation (ROSC) who had an electrocardiogram (ECG) obtained within 24 h following ROSC. The first ECG obtained was evaluated to determine the QTc and TpTe intervals. Hierarchical logistic regression was used to evaluate the association between prolongation of the QTc and TpTe intervals and clinical outcomes (in-hospital mortality and favourable neurologic outcome at hospital discharge). We included 443 patients, with a median age of 61 years (IQR: 50-72 years), 60.5% male, 65.7% OHCA, and 29.8% with initial shockable rhythm. Overall, 310 patients had QTc prolongation (70.0%), and 284 had TpTe prolongation (64.1%). Patients with TpTe prolongation had a greater incidence of initial shockable rhythm (35.6% vs. 19.5%, P < 0.001) and higher initial lactate (8.6 vs. 7.4 mmol/L, P = 0.03). QTc prolongation was not associated with in-hospital mortality [odds ratio (OR):1.27, 95% confidence interval (CI): 0.75-2.14, P = 0.37] or favourable neurologic outcome (OR: 0.88, 95% CI: 0.50-1.54, P = 0.65). TpTe prolongation was independently associated with in-hospital mortality (OR: 1.69, 95% CI: 1.01-2.85, P = 0.05) but not favourable neurologic outcome (OR: 0.78, 95% CI: 0.45-1.37, P = 0.39). CONCLUSION:TpTe interval prolongation, but not QTc interval prolongation, was associated with increased in-hospital mortality in patients resuscitated from cardiac arrest.