How is a patient's electroencephalogram (EEG) affected in the prognosis of a patient after post cardiac arrest?

EEG in Post-Cardiac Arrest Prognosis

EEG is a critical tool for neurological prognostication after cardiac arrest, with specific patterns reliably predicting poor outcomes when assessed ≥72 hours post-ROSC, though no EEG findings within the first 24 hours can reliably predict outcome. 1

Timing of EEG Assessment for Prognostication

Critical timing principle: No electrophysiological study, including EEG, reliably predicts outcome during the first 24 hours after return of spontaneous circulation (ROSC). 1 The earliest reliable prognostication occurs at 72 hours post-cardiac arrest in patients not treated with targeted temperature management (TTM). 2

Key Temporal Considerations:

• 24 hours post-ROSC: EEG interpretation may assist with predicting poor outcome in normothermic patients without confounders, but reliability is limited 1
• 72 hours post-ROSC: This is the optimal timepoint for reliable prognostication in both TTM-treated and non-TTM patients 1
• One week post-arrest: Specific EEG findings become increasingly useful for predicting poor outcomes 1

EEG Patterns Predicting Poor Outcome

In Patients Treated with TTM (32-34°C):

During hypothermia:

• Burst-suppression pattern: 0-6% false positive rate (FPR), though some studies show up to 6% FPR 1
• Absence of EEG reactivity to external stimuli: 2% FPR (95% CI: 1-7%) 1

After rewarming (≥72 hours post-ROSC):

• Persistent burst-suppression: 0% FPR (95% CI: 0-5%) - highly reliable predictor 1
• Persistent absence of EEG reactivity: 0% FPR (95% CI: 0-3%) - most reliable predictor 1
• Intractable status epilepticus (>72 hours) without EEG reactivity: May reasonably predict poor outcome 1, 2

In Patients NOT Treated with TTM:

At 72 hours or more post-ROSC:

• Burst-suppression pattern: 0% FPR (95% CI: 0-11%) 1
• Generalized suppression to <20 μV: Associated with poor outcome (3% FPR, 95% CI: 0.9-11%) 1
• Burst-suppression with generalized epileptic activity: 3% FPR 1
• Diffuse periodic complexes on flat background: 3% FPR 1
• EEG grades 4-5: 0% FPR (95% CI: 0-8%) at ≤72 hours 1

Important Confounding Factors

EEG interpretation is only reliable in the absence of:

• Sedatives or neuromuscular blockers 1
• Hypotension 1
• Hypothermia (when assessing normothermic patterns) 1
• Seizures 1
• Hypoxemia 1

Critical caveat: The prognostic accuracy of malignant EEG patterns is less reliable in patients treated with hypothermia, particularly during the cooling phase. 1

EEG Patterns Associated with Potential Good Outcome

Favorable prognostic indicators:

• Continuous cortical background activity within 12 hours of ROSC is associated with favorable neurological outcome 1
• Reactive EEG background (response to external stimuli) suggests better prognosis 1
• Presence of sleep spindles on initial EEG associated with favorable outcome at 6 months 1
• Moderate encephalopathy (diffuse slowing with reactivity/variability) has better prognosis than severe patterns 3

Temporal evolution matters: Patients who develop epileptiform abnormalities >24 hours after ROSC are more likely to recover than those with immediate onset. 2, 4

Seizures and Epileptiform Activity

Detection and Prevalence:

• Seizures occur in 10-35% of post-cardiac arrest patients who remain unresponsive after ROSC 2
• Nonconvulsive seizures are particularly common, detected in approximately 20-25% of comatose survivors 1
• Continuous EEG monitoring is more sensitive than brief intermittent recordings due to the episodic nature of seizures 1, 4

Prognostic Implications:

• Status epilepticus in TTM-treated patients: 7-11.5% FPR for poor outcome 1
• Seizures with reactive EEG background: Some patients achieve good outcome despite seizures 1
• Seizures with unreactive or discontinuous background: Almost invariably poor outcome 1
• Unequivocal electrographic seizures (frequencies ≥2.5 Hz) or evolving patterns suggest worse prognosis 1

Treatment Recommendations:

• Treat clinically apparent seizures (Class 1 recommendation) 1, 2
• Treat nonconvulsive seizures detected by EEG (Class 2a recommendation) 1, 2
• Do NOT use prophylactic antiseizure medications (Class 3: No Benefit) 2

Myoclonus: Special Considerations

Critical distinction: Myoclonus alone should NOT be used to predict poor outcome due to high false positive rates (5-11% FPR). 1, 2

However, status myoclonus (persistent, generalized myoclonus during first 72-120 hours) combined with other poor prognostic indicators is reasonable for predicting poor outcome (0% FPR, 95% CI: 0-4%). 1

Types of myoclonus with different implications:

• Subcortical myoclonus (no EEG correlate): May not warrant aggressive antiseizure treatment 1
• Cortical myoclonus (lockstep with epileptiform activity): Consider treatment 1
• Myoclonus with continuous cortical background: Some patients may recover 1, 4

Practical Implementation Algorithm

Step 1: Timing of EEG Assessment

1. Perform EEG promptly in all comatose post-cardiac arrest patients who cannot follow commands 2, 5
2. Consider continuous EEG monitoring rather than single snapshot recordings 1, 2
3. Do not make prognostic decisions based on EEG <72 hours post-ROSC 1, 2

Step 2: Ensure Absence of Confounders

Before interpreting EEG for prognosis, verify:

• No active sedation or recent neuromuscular blockade 1
• Hemodynamically stable (no hypotension) 1
• If TTM used, patient has completed rewarming and achieved normothermia 1
• No ongoing hypoxemia 1

Step 3: Pattern Recognition at ≥72 Hours

Highly malignant patterns (predict poor outcome with high reliability):

• Persistent burst-suppression after rewarming 1
• Persistent absence of EEG reactivity to stimuli 1
• Generalized suppression <20 μV 1
• Intractable status epilepticus (>72 hours) without reactivity 1, 2

Potentially favorable patterns:

• Continuous background activity 1, 3
• Reactive EEG (responds to stimulation) 1
• Moderate encephalopathy with reactivity/variability 3

Step 4: Multimodal Integration

Never use EEG alone for prognostication. Combine with:

• Clinical neurological examination (pupillary/corneal reflexes at 72 hours) 1
• Somatosensory evoked potentials (bilateral absent N20 has 0-2% FPR) 1
• Neuroimaging (MRI/CT findings) 1
• Biomarkers (NSE, S100B) 1

Common Pitfalls to Avoid

1. Premature prognostication: Do not make withdrawal decisions based on EEG findings <72 hours post-ROSC 1, 2

2. Ignoring confounders: Sedation, hypothermia, and neuromuscular blockade can produce malignant-appearing EEG patterns that are reversible 1

3. Over-reliance on single modality: EEG should never be the sole criterion for prognostication 1, 2

4. Misinterpreting myoclonus: Presence of myoclonus alone has unacceptably high false positive rate for poor outcome prediction 1, 2

5. Technical factors affecting interpretation: Low-voltage EEG can result from scalp edema, extra-axial fluid, or technical issues rather than true brain injury 1

6. Self-fulfilling prophecy: Be aware that EEG findings may influence withdrawal of life-sustaining treatment decisions, potentially inflating apparent predictive accuracy 1

Pediatric Considerations

In children after cardiac arrest, EEG background patterns have similar prognostic value, with incrementally worse backgrounds (normal → slow-disorganized → discontinuous/burst-suppression → attenuated-featureless) associated with progressively worse outcomes. 1 However, EEG should not be used as the sole criterion for prognostication in pediatric patients. 1