S24. The combination of electroencephalogram (EEG) and somatosensory evoked potentials (SSEP) for prognostication following anoxic brain injury – A pilot study from an Academic Hospital in Ireland

Abstract

Introduction Clinical evaluation for prognosis of comatose patients following hypoxic ischaemic encephalopathy in the intensive care setting is difficult. Electroencephalogram (EEG) and somatosensory evoked potentials (SSEP) both hold value as predictive tools when assessing outcome following cardiac arrest (CA) and hypoxic brain injury. EEG may be used for diagnosis and prognostication on a poorly responsive patient following catastrophic brain injury. There is evidence that loss of EEG reactivity is a strong risk factor for mortality or poor neurological recovery following cardiac arrest, however, bilaterally absent SSEPs remain the most reliable predictor of poor outcome. SSEPs have not been widely available in Ireland as a component of neurophysiologic investigation following hypoxic ischaemic encephalopathy. The aim of this review is to assess and compare the predictive value of EEG and SSEP separately as single modality testing and then subsequently combine both tests for a multimodal approach to predict poor outcome following anoxic brain injury. Is a multimodal approach mandatory? Methods Seventeen cardiac arrest survivors were included in this review. We adopted a multimodal approach to neurophysiologic investigations in the ICU setting with prospective evaluation of patients of hypoxic ischaemic encephalopathy after CA. EEG and SSEPs were performed at normothermia. EEG background reactivity to painful stimulation was tested. Results Of the 17 patients included in the review, two survived to discharge and 15 subsequently died during hospitalisation. EEG was utilised as a marker for poor outcome combining malignant EEG patterns and absent reactivity of background rhythms. 11 out of 17 fitted into the “poor outcome” category for EEG, all of whom died. Bilateral absent N20 cortical responses were utilised for poor outcome, 9 out of 17 fitted into the “poor outcome” category, all of whom went on to die. However, there were eight patients included showing present N20s, this is not prognostic of a favourable outcome. A combination of EEG and SSEP was needed to predict outcome. The EEG showed malignant EEG patterns and unreactive background in 2 of the present N20 patients who subsequently went on to die. Conclusion The EEG represents a useful tool for prognostic assessment. Unreactive EEG background and malignant EEG patterns are incompatible with good long term recovery. In settings without access to SSEPs this can be helpful, however, a bilaterally absent somatosensory evoked potential is still the most reliable predictor for poor outcome. Throughout our review a multimodal approach is shown to be favourable when assessing Hypoxic Ischemic Encephalopathy. This approach, as described above, was necessary to predict outcome in cases where the N20 cortical response was present, thus optimising prognostic accuracy when assessing hypoxic brain injury in the ICU. Early identification of patients without the potential for recovery of brain function may result in unnecessary prolongation of medical therapy.