Neuroprognostication After Cardiac Arrest and ROSC: Prediction of Poor Neurological Outcome

Biomarkers

 Last Full Review: ILCOR 2020

A systematic review and Consensus on Science with Treatment Recommendations (Berg et al. 2020, s92) was completed by International Liaison Committee on Resuscitation (ILCOR) in 2020 on the use of the biomarkers neuron-specific enolase (NSE), S-100B, glial fibrillary acidic protein (GFAP), serum tau protein, and neurofilament light chain (NFL) assessed within one week from cardiac arrest in adults who were comatose after resuscitation from in-hospital or out-of-hospital cardiac arrest, regardless of target temperature.

 

Evidence Summary

Outcomes included in a systematic review by ILCOR included prediction of poor neurological outcome defined as cerebral performance categories 3 to 5 or modified Rankin Score 4 to 6 at hospital discharge/1 month later (Berg et al. 2020, s92). Only observational studies were identified and deemed of moderate- to very low-certainty evidence. For neuron-specific enolase (NSE), an association was found in 13 observational studies between NSE of 33 to 120 ug/L within 72 hours and hospital discharge to 6 months with poor neurological outcome (Berg et al. 2020, s92). For S-100B protein, low-certainty evidence from three observational studies reported a wide variability of thresholds for 100% specificity for predicting poor neurological outcome following return of spontaneous circulation (ROSC) from 3 to 6 months (Berg et al. 2020, s92).

One study reported an association between GFAP and serum tau protein and predicted poor neurological outcome at 1 month and 6 months, respectively. Two studies reported an association between serum neurofilament light chain levels and predicted poor neurological outcome at 6 months (Berg et al. 2020, s92).

For adults who are comatose after cardiac arrest, ILCOR makes a weak recommendation to use NSE within 72 hours after ROSC, in combination with other tests, for predicting neurological outcome (Berg et al. 2020, s92).

A weak recommendation is made by ILCOR against using S-100B protein and against using serum levels of GFAP, serum tau protein, or NFL for predicting neurological outcome (Berg et al. 2020, s92).

Insights and Implications

Observational studies identified suggest an association between high levels of NSE measured at 24 to 72 hours after cardiac arrest and a poor neurological outcome (100% specificity). The evidence for S-100B, GFAP and serum NFL is very limited and while promising, does not support a recommendation for their use in prognostication.

 

Electrophysiology Tests

Last Full Review: ILCOR 2020

An International Liaison Committee on Resuscitation (ILCOR) systematic review evaluated the use of electrophysiology studies assessed within one week from cardiac arrest on adults who are comatose after return of spontaneous circulation (ROSC) in any setting, regardless of target temperature (Berg et al. 2020, s92).

 

Evidence Summary

Outcomes reviewed included prediction of poor neurological outcome. Very low-certainty evidence from a total of 20 observational studies evaluated the use of somatosensory evoked potentials (SSEPs), with four studies noting that a bilaterally absent negative 20 (N2O) SSEP wave within 24 hours from ROSC predicted poor neurological outcome from hospital discharge to 6 months with 100% specificity, and with sensitivity between 33% and 58% (Berg et al. 2020, s92). A bilaterally absent SSEP N2O wave at 24 to 96 hours predicted poor neurological outcome from hospital discharge to 6 months in 18 studies, with a wide range of sensitivity and specificity (Berg et al. 2020, s92).

An unreactive electroencephalogram (EEG) within 72 hours was reported in 12 studies to be associated with poor neurological outcome from hospital discharge to 6 months, with wide ranges in specificity and sensitivity (Berg et al. 2020, s92). Rhythmic/periodic discharges within 24 hours were reported in two studies to predict poor neurological outcome from 3 to 6 months with 100% specificity and sensitivity of 2.4% to 7.9%, and when occurring within 48 hours, four studies reported an association with poor neurological outcomes from 3 to 6 months with specificity of 97% to 100% and widely ranging sensitivity. Other observational studies reported the presence of rhythmic/periodic discharges at 48, 72, and 76 to 77 hours, or within 5 days as being associated with poor neurological outcome at 6 months with high specificity and low or wide ranges in sensitivity (Berg et al. 2020, s92).

Sporadic, nonrhythmic/periodic discharges within 24, 48, 48 to 72, or 96 to 120 hours were reported in five observational studies to be associated with poor neurological outcome at various time points with good specificity and low sensitivity (Berg et al. 2020, s92). Seizures on EEG within 120 hours were reported in five studies to be associated with poor neurological outcome from hospital discharge to 6 months with 100% specificity and low sensitivity. Status epilepticus likewise was reported to be associated with poor neurological outcome from hospital discharge to 6 months with high specificities and lower sensitivities (Berg et al. 2020, s92).

Burst suppression within 24 hours (one study) and at 24 to 76 hours (four studies) was reported to be associated with pool neurological outcome at 6 months with high specificity and low sensitivity. Highly malignant EEG patterns within 36 hours were reported in nine studies to be associated with poor neurological outcome from hospital discharge to 6 months with specificities ranging from 91% to 100% and sensitivities from 0.4% to 97%. Other studies extended the time for the presence of highly malignant EEG patterns in predicting poor neurological outcome at 6 months to 120 hours (Berg et al. 2020, s92).

Recommendations by ILCOR (Berg et al. 2020, s92) for adult patients who are comatose after cardiac arrest include suggestions:

• To use a bilaterally absent N2O SSEP wave in combination with other indices to predict poor outcome.
• Against using EEG background reactivity alone to predict poor outcome.
• To use the presence of epileptiform activity on EEG to predict poor outcome.
• Against using seizures on EEG or status epilepticus to predict poor outcome.
• To use burst suppression on EEG to predict poor outcome in adults who are off sedation after cardiac arrest and comatose.
• To use highly malignant EEG patterns to predict poor outcome in adults who are comatose and off sedation after cardiac arrest.

 

Insights and Implications

The electrophysiological tests that are suggested as part of a multimodal approach to predicting neurological outcome of adults who are comatose after cardiac arrest include bilaterally absent N2O wave of somatosensory evoked potential, the presence of seizure activity on EEG, and burst suppression on EEG. It is suggested to not use the absence of EEG background reactivity alone to predict poor outcome.

 

Clinical Examination

Last Full Review: ILCOR 2020

A 2020 International Liaison Committee on Resuscitation (ILCOR) systematic review and Consensus on Science with Treatment Recommendations evaluated the use of clinical examination on adults who were comatose following resuscitation from either in-hospital or out-of-hospital cardiac arrest (Berg et al. 2020, s92).

 

Evidence Summary

The 2020 ILCOR review evaluated the use of pupillary light reflex (PLR), pupillometry, corneal reflex, myoclonus and status myoclonus, as assessed within 1 week from cardiac arrest, on adults who were comatose following resuscitation from either in-hospital or out-of-hospital cardiac arrest, and regardless of target temperature (Berg et al. 2020, s92).

Outcomes included prediction of poor neurological outcome. For the standard pupillary light reflex, a total of 24 observational studies were included. These studies were reported to show an association between absent standard pupillary light reflex immediately after return of spontaneous circulation (ROSC), at less than 24 hours, at 36 to 72 hours, at 48 to 72 hours, at 72 hours, and at 72 hours to 7 days and poor neurological outcome at various time points from hospital discharge, with a wide range of specificities and sensitivities (Berg et al. 2020, s92). Similar associations were reported in observational studies for automated pupillometry (qPLR) at 24 hours, 48 hours and 72 hours for poor neurologic outcome from hospital discharge to 12 months, and for automated pupillometry (Neurologic Pupil index; NPi) from hospital discharge to 3 months (Berg et al. 2020, s92).

Very low-certainty evidence from a total of 14 observational studies of absent corneal reflex at time points between immediately after ROSC and 72 hours to day 7 were reported to be associated with poor neurologic outcome to discharge or from hospital discharge to 12 months, with a wide range of sensitivities and specificities (Berg et al. 2020, s92).

Eight observational studies with very low-certainty evidence reported an association between the presence of myoclonus within 96 hours and poor neurological outcome from hospital discharge to 6 months. Specificities ranged between 77.8% and 97.8%, while sensitivity ranged from 18.2% to 39.6% (Berg et al. 2020, s92).

Weak recommendations are made by ILCOR (Berg et al. 2020, s92) for predicting neurological outcome of adults who are comatose after cardiac arrest using:

• Pupillary light reflex at 72 hours or later after ROSC.
• Quantitative pupillometry at 72 hours or later after ROSC.
• Bilateral absence of corneal reflex at 72 hours or later after ROSC.
• Presence of myoclonus or status myoclonus within 96 hours after ROSC.

It is also suggested to record electroencephalogram (EEG) in the presence of myoclonic jerks to detect an associated epileptiform activity (Berg et al. 2020, s92).

Insights and Implications

The highest specificity for prediction of poor neurological outcome was noted in this review to be at 72 hours or later after cardiac arrest, possibly partly as a result of the earlier effect of sedatives used for targeted temperature management or ventilation. Despite limited evidence, pupillometry using NPi was associated with a 100% specificity for prediction of poor neurological outcome as early as 24 hours after cardiac arrest. The recommendation of 72 hours was based on evidence for standard manual PLR, a lower likelihood of effects at 72 hours from sedation, and increasing specificity noted for a qPLR between 24 and 72 hours. The recommended use of corneal reflex at 72 hours is also chosen because of potential confounding from initial use of sedatives or relaxants. Finally, although presence of myoclonus is included in the recommendations, reviewers noted that definitions were absent or inconsistent in most of the included studies and myoclonus may be associated with epileptiform activity on an EEG (Berg et al. 2020, s92).

 

Brain Imaging

Last Full Review: ILCOR 2020

An updated 2020 International Liaison Committee on Resuscitation (ILCOR) systematic review (Berg et al. 2020, s92) evaluated imaging for prognostication in comatose adult survivors of cardiac arrest, regardless of target temperature.

 

Evidence Summary

Imaging studies were assessed within 1 week from cardiac arrest, and outcomes included prediction of poor neurological outcome at hospital discharge/1 month or later. Observational studies were identified evaluating grey matter to white matter ratio (GWR)-average, GWR-basal ganglia, GW-putamen/corpus callosum, GWR-simplified (putamen/posterior limb of internal capsule), GWR caudate nucleus/posterior limb of internal capsule, GWR-cerebrum, GWR-thalamus/corpus callosum, GWR-caudate nucleus/corpus callosum, GWR in cardiac versus non-cardiac etiology, diffusion-weighted imaging (DWI), and apparent diffusion coefficient (ADC). Some of these studies suggested an association between the imaging performed following return of spontaneous circulation (ROSC) at various intervals and poor neurological outcome at time intervals ranging from 1 hour to 6 months post-ROSC (Berg et al. 2020, s92).

A weak recommendation is made by ILCOR to use GWR ratio on brain computed tomography scan for predicting neurological outcome of adults who are comatose after cardiac arrest. No GWR threshold for 100% specificity can be recommended (Berg et al. 2020, s92). It is suggested by ILCOR to use DWI and/or ADC on brain magnetic resonance imaging (MRI) for predicting neurological outcome of adults who are comatose after cardiac arrest (Berg et al. 2020, s92).

Insights and Implications

All studies included were of very low certainty and marked heterogeneity was noted for measurement techniques for GWR. Brain edema following cardiac arrest in patients who are unconscious predicts poor outcome. GWR is a means to provide a quantitative assessment of brain edema. Both DWI and ADC were described by the ILCOR reviews as having potential for predicting poor neurological outcome following cardiac arrest, but it was noted that the definition of a positive DWI MRI after cardiac arrest was inconsistent or not included in some studies, and there was wide heterogeneity of measurement techniques for ADC across studies (Berg et al. 2020, s92).