Electroencephalographic Patterns and Outcomes After Hypoxic Ischaemic Brain Injury Secondary to Cardiac Arrest: A Systematic Review and Meta-Analysis
Abstract number :
3.419
Submission category :
16. Epidemiology
Year :
2019
Submission ID :
2422310
Source :
www.aesnet.org
Presentation date :
12/9/2019 1:55:12 PM
Published date :
Nov 25, 2019, 12:14 PM
Authors :
Kevin H. Perera, University of Calgary; Sundus Khan, University of Calgary; Shaily Singh, University of Calgary; Julie Kromm, University of Calgary; Tolulope Sajobi, University of Calgary; Nathalie Jette, Icahn School of Medicine; Samuel Wiebe, University
Rationale: Many persons successfully resuscitated from cardiac arrest (CA) remain initially comatose due to hypoxic-ischemic brain injury (HIBI). Electroencephalography (EEG) has been used to prognosticate recovery but its predictive value has yet to be systematically evaluated for this indication in intensive care settings. Hence, the objective of the study was to determine the prognostic utility of specific EEG patterns for predicting disability and mortality following HIBI secondary to CA. Methods: We searched Medline, Embase, and Cochrane Central up to March 2018. We included original research involving prospective and retrospective cohort studies relating specific EEG patterns to disability and mortality in comatose patients above the age 16 post cardiac arrest requiring admission to an intensive care setting. We evaluated study quality using the QUADAS-2 tool. Descriptive statistics were used to summarize patient and EEG characteristics. Estimates of sensitivity and specificity for specific EEG patterns defined a priori according to each outcome of interest were pooled using a random effects bivariate meta-analysis. Funnel plots were used to assess publication bias. Results: Of 3840 abstracts, 175 were reviewed in full text, of which 39 were included in the final dataset with 22 meta-analyzed from a total of 2433 patients. No reported EEG pattern was invariably associated with death or disability. Pooled specificity of status epilepticus, burst suppression and electrocerebral silence for disability or death were high (94-99%) but sensitivities were low (10-43%). Study quality varied depending on domain; patient flow and timing performed well in all while EEG interpretation was retrospective in 17/39 studies. Conclusions: Current evidence is inconsistent but indicates that EEG specificity far exceeds sensitivity indicating risk of false negative outcome attribution (incorrectly suggesting survival in HIBI patients following CA). Standardized study protocols and EEG definitions are required to better evaluate the prognostic utility of EEG in post CA populations to address existing between study heterogeneity. Funding: No funding
Epidemiology