Post-Cardiac Arrest Care — Part 3: Neuroprognostication After Cardiac Arrest

1. Principles of Neuroprognostication

Neuroprognostication — the systematic assessment of likely neurologic recovery after cardiac arrest — is one of the most consequential clinical activities in post-arrest care. The accuracy of prognostication directly determines whether life-sustaining treatment is continued or withdrawn, making the stakes of this process uniquely high.^{1 2 3}

1.1 Why Multimodal Prognostication is Mandatory

No single test, biomarker, or clinical finding is sufficient to predict poor neurologic outcome with certainty. This principle, consistently emphasized by all major guideline bodies, reflects:

• The heterogeneity of hypoxic-ischemic brain injury patterns
• The confounding effects of sedation, hypothermia, metabolic derangements, and organ failure on clinical examination and diagnostic testing
• The documented occurrence of neurologic recovery in patients initially assessed as having “hopeless” prognosis (the “Lazarus phenomenon”)
• Historical evidence that premature withdrawal of life-sustaining treatment (WLST) has been a significant contributor to mortality in post-cardiac arrest patients — the “self-fulfilling prophecy” problem⁴

1.2 The Self-Fulfilling Prophecy

The self-fulfilling prophecy in neuroprognostication refers to the circular reasoning in which:

1. A patient is assessed as having a poor prognosis
2. Life-sustaining treatment is withdrawn based on that assessment
3. The patient dies, “confirming” the poor prognosis
4. The prognostic test is retrospectively validated by the outcome it helped create

This phenomenon has been identified as a significant limitation in virtually all neuroprognostication studies, including the HACA and TTM trials. It particularly affects the reported specificity of prognostic tests: if patients with positive indicators (suggesting poor prognosis) have WLST, the apparent specificity is artificially inflated because recovery is never given the chance to occur.

Mitigation strategies:

• Use multiple, independent prognostic modalities before reaching a final determination
• Delay prognostication until ≥72 hours after ROSC (or ≥72 hours after normothermia if TTM was used)
• Ensure that prognostic test results are not used in real-time WLST decisions before the designated prognostication timepoint
• Involve neurology consultation in all cases of uncertain prognosis
• Explicitly discuss the self-fulfilling prophecy risk with families and the care team

2. Timing of Neuroprognostication

2.1 Minimum Waiting Period

2.2 Practical Timing Considerations

• The ≥72-hour rule is a minimum, not an absolute threshold. If confounders (residual sedation, organ failure, metabolic derangements) are present at 72 hours, prognostication should be further delayed until these confounders are resolved or accounted for.
• Sedation clearance may require 24–72 additional hours after discontinuation, particularly after prolonged infusions (propofol, midazolam) and in the setting of renal or hepatic dysfunction.
• Neuromuscular blockade must be fully reversed before clinical examination for prognostication.
• Some prognostic tools (CT, MRI, EEG) can be obtained earlier to gather data, but the final integrated prognostic determination should not occur before the minimum timepoint.

3. Clinical Examination

The bedside neurologic examination remains a cornerstone of neuroprognostication, though it must be interpreted in the context of potential confounders and combined with other modalities.^{1 2 3}

3.1 Pupillary Light Reflex

Confounders affecting pupillary light reflex:

• Anticholinergic medications (atropine, ipratropium — especially if nebulized near eyes)
• Pre-existing ocular pathology (cataracts, prior surgery, prosthetic eye)
• Neuromuscular blocking agents (do NOT affect pupillary light reflex — this reflex is mediated by smooth muscle, not skeletal muscle)
• Local eye trauma from resuscitation
• High-dose barbiturates (may cause unreactive pupils but are rarely used in current TTM practice)

3.2 Corneal Reflexes

Assessment technique:

• Apply sterile saline drops or a wisp of cotton to the cornea (avoid tissue paper, which may cause corneal abrasion)
• Observe for bilateral eyelid closure (requires intact CN V afferent and CN VII efferent)
• Confounders: Sedation, neuromuscular blockade, facial nerve palsy, corneal edema from prolonged eye closure

3.3 Motor Response

GCS Motor Score reference:

3.4 Myoclonus and Status Myoclonus

Myoclonus in the post-cardiac arrest setting requires careful characterization, as its prognostic significance varies dramatically depending on its type, timing, and EEG correlation.^{1 2 5}

Critical distinction: The presence of early status myoclonus (within 48 hours) in a comatose patient with a highly malignant EEG background is a strong predictor of poor outcome. However, myoclonus occurring during or after neurologic recovery (Lance-Adams syndrome) is NOT a predictor of poor outcome and should not be used as a basis for treatment limitation.

4. Electrophysiology

4.1 Electroencephalography (EEG)

EEG is a critical component of the multimodal neuroprognostication approach. It provides information about cortical function, seizure activity, and the degree of background suppression, each of which carries prognostic significance.^{1 2 6}

4.1.1 EEG Timing

4.1.2 EEG Classification for Prognostication

The standardized terminology from the 2021 consensus classifies post-arrest EEG patterns into prognostic categories:^{2 6}

Highly malignant patterns (strong predictors of poor outcome):

Malignant patterns (associated with poor outcome but with higher uncertainty):

Benign/indeterminate patterns:

4.1.3 Key Principles for EEG Interpretation in Neuroprognostication

1. Highly malignant EEG patterns at ≥24 hours after ROSC (or after rewarming and sedation clearance) are strong predictors of poor outcome, but should be combined with at least one other prognostic modality before determining prognosis.
2. EEG reactivity should always be tested and documented; absence of reactivity is concerning but not independently sufficient to predict poor outcome.
3. Sedation is a major confounder. High-dose propofol and midazolam can produce burst-suppression patterns indistinguishable from those caused by severe hypoxic-ischemic injury. EEG should ideally be interpreted after at least 24 hours of sedation clearance.
4. Serial EEGs are more informative than a single recording. An evolving pattern (e.g., from burst-suppression to continuous background) suggests recovery; a persistent highly malignant pattern is more concerning.
5. Continuous EEG monitoring (rather than spot EEGs) is recommended to detect non-convulsive seizures and to track background evolution.

4.2 Somatosensory Evoked Potentials (SSEPs)

SSEPs assess the integrity of the somatosensory pathway from the peripheral nerve (typically the median nerve at the wrist) through the dorsal columns of the spinal cord, brainstem, and thalamus to the primary somatosensory cortex. The cortical N20 component is the key prognostic target.^{1 2 7}

4.2.1 Technique

• Stimulation: Median nerve at the wrist, square-wave pulse, 3–5 Hz, sufficient intensity to produce thumb twitch
• Recording: Scalp electrodes over the contralateral somatosensory cortex (C3’ or C4’, referenced to Fz)
• Key component: N20 — the negative cortical potential occurring approximately 20 milliseconds after stimulation, representing thalamocortical activation
• Control: The cervical N13 or Erb’s point potential must be present to confirm adequate peripheral nerve stimulation; absence of N20 is only meaningful if the peripheral and subcortical responses are intact

4.2.2 Prognostic Significance

4.2.3 Advantages and Limitations of SSEPs

5. Biomarkers

5.1 Neuron-Specific Enolase (NSE)

NSE is a glycolytic enzyme found predominantly in neurons and neuroendocrine cells. Elevated serum NSE levels after cardiac arrest reflect the magnitude of neuronal injury and have been extensively studied as a prognostic biomarker.^{1 2 8}

5.1.1 Recommended Sampling Protocol

5.1.2 Prognostic Thresholds

5.1.3 Confounders and Limitations

5.2 S-100B Protein

S-100B is a calcium-binding protein found in astrocytes and Schwann cells. Elevated serum levels reflect astrocytic injury and blood-brain barrier disruption.^{1 9}

5.3 Emerging Biomarkers

6. Neuroimaging

6.1 Computed Tomography (CT)

6.1.1 CT Head — Timing and Purpose

6.1.2 Gray-White Matter Ratio (GWR)

The GWR quantifies the loss of differentiation between gray and white matter on non-contrast CT, which occurs as a result of cytotoxic edema in hypoxic-ischemic brain injury.^{1 2 10}

Measurement technique:

• Select representative axial CT slices (typically at the level of the basal ganglia and the centrum semiovale)
• Place regions of interest (ROIs) in:
  • Gray matter: Caudate nucleus, putamen, cortical gray matter (insular cortex is commonly used)
  • White matter: Posterior limb of the internal capsule, corpus callosum, centrum semiovale white matter
• Calculate GWR = mean gray matter density (HU) / mean white matter density (HU)
• Normal GWR is approximately 1.2–1.3 (gray matter is denser than white matter)
• In severe hypoxic-ischemic injury, the GWR approaches 1.0 or lower as gray matter becomes edematous and isodense or hypodense relative to white matter

Prognostic significance:

Limitations:

• GWR measurement is not standardized across centers (different ROI placement methods, different slice selection)
• Early CT (within 6 hours) may be normal even in patients who will develop severe injury
• CT is less sensitive than MRI for detecting early hypoxic-ischemic changes

6.2 Magnetic Resonance Imaging (MRI)

MRI, particularly diffusion-weighted imaging (DWI), is the most sensitive neuroimaging modality for detecting hypoxic-ischemic brain injury after cardiac arrest.^{1 2 11}

6.2.1 Optimal Timing

6.2.2 Key MRI Findings

6.2.3 Distribution Patterns

6.2.4 Practical Considerations for MRI in Post-Arrest Patients

• Safety: MRI requires removal of all ferromagnetic materials; ensure cooling devices, monitoring equipment, and vascular access are MRI-compatible or removed
• Hemodynamic stability: Patients must be hemodynamically stable enough for transport to MRI and the duration of the scan (45–60 minutes)
• Timing of sedation clearance: MRI for prognostication is most informative when obtained at 3–5 days post-ROSC; does not require complete sedation clearance (unlike clinical examination)
• MRI is not required for prognostication if other modalities (clinical exam, EEG, SSEPs, biomarkers, CT) are concordant; however, it is the most sensitive test and should be obtained when prognosis is uncertain

7. Confounders in Neuroprognostication

All prognostic modalities are subject to confounders that may reduce their reliability. Recognition and documentation of these confounders is essential for accurate prognostication.^{1 2 3}

7.1 Summary of Confounders by Modality

7.2 Sedation as a Confounder — Detailed Considerations

Sedation is the most pervasive and impactful confounder in neuroprognostication. All commonly used ICU sedatives affect the neurologic examination and EEG.³

Key principle: Pupillary light reflex and SSEPs are the least affected by sedation and remain reliable in the presence of standard ICU sedation. EEG and motor examination are the most affected and should be interpreted with extreme caution during or immediately after sedation.

8. Multimodal Prognostication Algorithm

The following algorithm synthesizes the current evidence into a systematic, stepwise approach to neuroprognostication.^{1 2 3}

8.1 Step 1 — Prerequisites (Before Initiating Formal Prognostication)

8.2 Step 2 — Clinical Examination

Perform a comprehensive neurologic examination assessing:

8.3 Step 3 — Ancillary Testing

If clinical examination suggests poor prognosis (bilateral absent PLR, M1-M2, absent corneal reflexes), obtain ancillary testing to confirm or refute the clinical impression.

The goal is to identify ≥2 concordant predictors of poor outcome from DIFFERENT modalities:

8.4 Step 4 — Integration and Decision

8.5 Prognostication Summary Table — Test Performance

9. Outcome Scales

9.1 Cerebral Performance Category (CPC) Scale

The CPC scale is the most widely used outcome measure in post-cardiac arrest research and clinical practice.¹

9.2 Modified Rankin Scale (mRS)

The mRS provides a more granular assessment of functional outcome and is increasingly used in post-arrest research (particularly in TTM2 and subsequent trials).¹²

Good outcome is typically defined as mRS 0–3; poor outcome as mRS 4–6. Some studies use mRS 0–2 as the threshold for good outcome.

References