British Society for Clinical Neurophysiology

...to promote and encourage for the public benefit the science and practice of clinical neurophysiology and related sciences

Abstract Details

EEG and SSEP for prognostication following cardiac arrest at the Bristol Royal Infirmary, 2012-2017

Introduction Prognostication following cardiac arrest continues to be a diagnostic challenge. Our neurophysiology department has provided neuroprognostication support with electroencephalogram (EEG) and Somatosensory Evoked Potentials (SSEPs) since 2012. We present an analysis of 105 cases where neurophysiology assessment was undertaken following cardiac arrest in a tertiary centre between January 2012 and July 2017. Methods Every EEG and/or SSEP investigations carried out on ICU following cardiac arrest was extracted and analysed. The system proposed by Synek to stratify EEG severity was used (Guerit et al., 2009). Results During the 5 year period approximately 700 patients were admitted following cardiac arrest and 105 underwent neurophysiological assessment. The general cardiac arrest population admitted to our unit had an average age of 63 years, 74% were male. ICU and hospital mortality were 49% and 55% respectively. Patients in the neurophysiology cohort had an average age of 61 years, 79% were male. Hospital mortality was 94%. Neurophysiology was undertaken following rewarming, median 3 days post admission. Of the 102 EEGs performed none were benign (Synek grade 0-2). 16 were grade 3, 72 grade 4, and 14 grade 5. Malignant EEG (Grade 4/5) had a sensitivity of 86% at predicting mortality and a specificity of 50%. AUROC for malignant EEG predicting death was 0.68. Of 97 SSEPSs 48% demonstrated bilaterally present signal, 42% were absent. 9% were non-diagnostic. Absent SSEPs had a specificity of 100% and sensitivity of 51% for predicting mortality. AUROC was 0.75 for absent SSEP predicting death. No patient with absent SSEPs survived. Hospital mortality despite present SSEPs was 87%. Combining neurophysiology findings showed that no patient with absent SSEP survived, regardless of EEG finding (40/40 mortality). Patients with present SSEPs and Synek grade 3 EEG had a hospital mortality of 66.7% (6/9). Those with Synek grade 4 or 5 and present SSEPs had a 91.7% (33/36) mortality. Summary In our case series, 16% of cardiac arrests underwent neurophysiology. All patients with a malignant EEG (grade 4 or 5) and absent SSEPs died. All survivors had bilaterally present SSEPs but 87% of patient with present SSEPs did not survive. Combing EEG and SSEPs findings identified those with the best survival. The use of observational data, particularly the absence of SSEP is problematic given the self-fulfilling prophecy of neuroprognostication. We propose to incorporate clinical examination findings into our model to assess the accuracy of multimodal prognostication. 

TitleForenamesSurnameInstitutionLead AuthorPresenter
DrAggieSkorkoBristol Royal Infirmary, UHB NHS Foundation Trust, Bristol
DrMattThomasBRI
Dr TimGouldBRI
DrMPachuckiBRI
Dr NickKaneNorth Bristol NHS Trust
Reference
JM Guerit et al. (2009) ‘Consensus on the use of neurophysiological tests in the intensive care unit (ICU): Electroencephalogram (EEG), evoked potentials (EP), and electroneuromyography (ENMG).’ Clinical Neurophysiology; 39: pp 71—83.