The prognostic value of discontinuous EEG patterns in postanoxic coma
Introduction
Early outcome prediction of comatose patients after cardiac arrest remains a challenge. The EEG is a sensitive and reliable tool, especially within the first 24 h after cardiac arrest (Hofmeijer and Van Putten, 2016). A timely restoration of continuous, normal amplitude background activity is essential for good neurological recovery. Consistently among various studies, a continuous EEG pattern within 12 h after cardiac arrest predicts a good outcome, whereas a persistent isoelectric or low-voltage EEG at 24 h predicts a poor outcome (Sivaraju et al., 2015, Sondag et al., 2017, Spalletti et al., 2016).
In addition to the EEG background continuity, the degree of amplitude fluctuation seems to be a powerful indicator of the severity of postanoxic encephalopathy. EEG patterns with pronounced alternations in amplitudes, such as burst-suppression with identical bursts (Hofmeijer et al., 2013), and generalized periodic discharges (GPDs) on a suppressed background (Ruijter et al., 2015), have a strong association with poor outcome.
Although both background continuity and amplitude fluctuations play an important role in the visual assessment of the postanoxic EEG, they remain hard to quantify by a human observer. Continuity and relative amplitude are usually treated as categorical variables (Hirsch et al., 2013), for which interrater agreement is far from perfect (Gaspard et al., 2014, Westhall et al., 2015). Therefore, visual analysis alone has probably not been able to elucidate the discriminative power of these features.
The assessment of discontinuous EEG patterns can be improved by the use of quantitative EEG (qEEG). Quantitative measures do not suffer from interrater variability, can be applied by non-experts, and save reviewing time (Swisher and Sinha, 2016, Tjepkema-Cloostermans et al., 2017, Zubler et al., 2016). Quantitative measures for EEG background continuity have proven to be valuable for the prediction of outcome in postanoxic coma, but have only been applied to small case series (Noirhomme et al., 2014, Wennervirta et al., 2009) or selected EEG patterns (Ruijter et al., 2015), so far. The prognostic value of the amplitude ratio between bursts and suppressions has, to the best of our knowledge, never been investigated in any visual or quantitative analysis.
In this study, we relate outcome of postanoxic coma to quantitative measures of EEG background continuity and the amplitude ratio between bursts and suppressions. We will show that these straightforward measures are reliable predictors of outcome, with sensitivities higher than reported for visual EEG analysis. Since the presented features remain closely related to key visual observations, they will also provide insights to be of value in routine EEG assessment.
Section snippets
Patients
This study was designed as a prospective cohort study on continuous EEG monitoring for outcome prediction of comatose patients after cardiac arrest. Data were collected between June 2010 and April 2017 in Medisch Spectrum Twente and Rijnstate, two large teaching hospitals in the Netherlands. All consecutive adult patients that were admitted comatose after cardiac arrest to the Intensive Care Unit (ICU) were included. Parts of the EEG data up to December 2015 were used previously for studies on
Patients
EEG recordings were started in 582 patients. Twenty-three patients were lost to follow-up, leaving 559 patients for the analysis. Outcome was good in 46%. Baseline characteristics are listed in Table 2 and were as expected: patients with good outcome were younger (60 vs. 68 years, p < 0.001), more often had a primary cardiac cause of the arrest (94% vs. 76%, p < 0.001), ventricular fibrillation as initial rhythm (92% vs. 58%, p < 0.001), and required higher doses of propofol (2.95 vs.
Discussion
We show that the background continuity index and the burst-suppression amplitude ratio are straightforward, quantitative EEG-measures that assist in the prediction of outcome after cardiac arrest. Our results confirm that return of EEG background continuity within 24 h indicates a good outcome. Otherwise, a lack of background continuity predicts poor outcome without false positives, except for the first 10 h after cardiac arrest. An amplitude ratio between ‘bursts’ and suppressions of 6.12 and
Acknowledgements
We thank Carin Eertman, Astrid Glimmerveen, Monique Raaijmakers, Saskia Metz, Helene Vogelesang, and Yvonne Teitink for assistance with data aquisition. We thank the ICU staff and all clinical neurophysiology lab technicians from Medisch Spectrum Twente an Rijnstate hospital for the extensive support and constructive collaboration. Barry J. Ruijter was financially supported by the Dutch National Epilepsy Fund (Nationaal Epilepsie Fonds, grant reference NEF 14–18). The funders had no role in
Conflict of interest statement
M.J.A.M. van Putten is co-founder of Clinical Science Systems, which is a supplier of EEG systems for one of the participating sites (Medisch Spectrum Twente). Clinical Science Systems did not provide funding and was not involved in the design, execution, analysis, interpretation or publication of the study. The other authors do not report any conflicts of interest.
References (40)
Computerized epileptiform transient detection in the scalp electroencephalogram: Obstacles to progress and the example of computerized ECG interpretation
Clin Neurophysiol
(2009)- et al.
EEG in postanoxic coma: Prognostic and diagnostic value
Clin Neurophysiol
(2016) - et al.
Assessment of outcome after severe brain damage: A practical scale
Lancet
(1975) - et al.
The natural course of neurological recovery following cardiopulmonary resuscitation
Resuscitation
(1998) - et al.
Prognostic value of neonatal discontinuous EEG
Pediatr Neurol
(2002) - et al.
Synaptic damage underlies EEG abnormalities in postanoxic encephalopathy: A computational study
Clin Neurophysiol
(2017) - et al.
Propofol and the electroencephalogram
Clin Neurophysiol
(2010) - et al.
Prognostication in comatose survivors of cardiac arrest: An advisory statement from the European Resuscitation Council and the European Society of Intensive Care Medicine
Resuscitation
(2014) - et al.
Single electroencephalographic patterns as specific and time-dependent indicators of good and poor outcome after cardiac arrest
Clin Neurophysiol
(2016) - et al.
Triggering and execution of neuronal death in brain ischaemia: Two phases of glutamate release by different mechanisms
Trends Neurosci
(1994)
The use of conventional EEG for the assessment of hypoxic ischaemic encephalopathy in the newborn: A review
Clin Neurophysiol
Interrater variability of EEG interpretation in comatose cardiac arrest patients
Clin Neurophysiol
The origin of extracellular fields and currents - EEG, ECoG, LFP and spikes
Nat Rev Neurosci
Continuous electroencephalography monitoring for early prediction of neurological outcome in postanoxic patients after cardiac arrest: A prospective cohort study
Crit Care Med
Loss and recovery of functional connectivity in cultured cortical networks exposed to hypoxia
J Neurophysiol
C.C.E.M.R. Consortium. Interrater agreement for Critical Care EEG Terminology
Epilepsia
Poor survival after cardiac arrest resuscitation: A self-fulfilling prophecy or biologic destiny?
Crit Care Med
Predicting neurologic outcome after targeted temperature management for cardiac arrest: systematic review and meta-analysis
Crit Care Med
Sensitivity of quantitative EEG for seizure identification in the intensive care unit
Neurology
American Clinical Neurophysiology Society’s standardized critical care EEG terminology: 2012 version
J Clin Neurophysiol
Cited by (41)
Common patterns of EEG reactivity in post-anoxic coma identified by quantitative analyses
2022, Clinical NeurophysiologyEEG patterns and their correlations with short- and long-term mortality in patients with hypoxic encephalopathy
2021, Clinical NeurophysiologyEEG functional connectivity contributes to outcome prediction of postanoxic coma
2021, Clinical NeurophysiologyCitation Excerpt :The best performance for prediction of poor outcome in this study was achieved inspecting data at 12 hours after cardiac arrest, showing a sensitivity of 47% (42–51 95%CI) and a specificity of 100% (100–100 95%CI), which are similar results to those obtained with our classifiers trained only with functional connectivity. Regarding prediction of good outcome, functional connectivity achieves a higher sensitivity (69% (69–69)) at 95% specificity (95% (95–95)) than previous work in the field, which reported sensitivities of approximately 50% at 90% specificity (Tjepkema-Cloostermans et al. 2013, 2017; Sondag et al. 2017; Ruijter et al. 2018, 2019). The highest predictive values were achieved using data from both 12 and 48 hours after cardiac arrest, showing that this approach can be used for early detection of poor outcome patients.