Post-arrest neuroprognostication
multimodal prognosis · Critical Care / ICU
Physician-written · Verified by three sets of doctors · Reviewed November 2026
Bedside summary
Structured, multimodal assessment of neurological outcome in the comatose survivor of cardiac arrest. No single test should determine withdrawal of care; prognostication is delayed to avoid self-fulfilling prophecy and confounders.
Red flags
- Reliance on a single test to withdraw care
- Prognostication before 72 h or before confounders cleared
- Bilaterally absent pupillary/corneal reflexes plus absent N20
- Malignant EEG (suppression, burst suppression with identical bursts)
- Diffuse anoxic injury on imaging with loss of grey-white differentiation
Assessment
Differential diagnosis
- Residual sedation/neuromuscular blockade
- Hypothermia effect
- Metabolic derangement
- Non-convulsive status epilepticus
- Hypoxic-ischaemic brain injury
- Structural intracranial pathology
Investigations
- Clinical neuro exam (pupillary and corneal reflexes, motor response, myoclonus)
- EEG (burst suppression, status epilepticus, reactivity)
- Somatosensory evoked potentials (N20)
- Neuron-specific enolase serial levels
- Brain CT (grey-white differentiation, oedema)
- MRI (diffusion restriction) at 2-5 days
Management
- Delay definitive prognostication to at least 72 hours after return to normothermia
- Ensure confounders are excluded: clear sedation, neuromuscular blockers, hypothermia, hypotension, severe metabolic derangement
- Perform a careful clinical examination for bilaterally absent pupillary and corneal reflexes
- Quantify pupillary responses with automated pupillometry where available
- Assess for status myoclonus and capture EEG to distinguish benign from malignant patterns and detect NCSE
- Obtain somatosensory evoked potentials looking for bilaterally absent N20 responses
- Measure serial neuron-specific enolase, noting rising or high values and avoiding haemolysed samples
- Use brain imaging (CT early, MRI at 2-5 days) for diffuse anoxic injury
- Combine at least two concordant unfavourable predictors before concluding a poor prognosis
- Communicate uncertainty honestly to family and make decisions through a multidisciplinary team
- Continue active care when findings are indeterminate and re-evaluate over time
Drugs & doses
Short-acting sedation (e.g. propofol/remifentanil) stopped in time to allow clean examination
Antiseizure therapy (levetiracetam or valproate) for post-arrest seizures
Disposition
ICU with neurology input; structured multidisciplinary and family discussions; continue support while prognosis remains uncertain and reassess.
Complications & pitfalls
- Self-fulfilling prophecy from premature withdrawal
- Missed non-convulsive seizures
- Family distress from inconsistent messaging
- Over-optimistic or over-pessimistic decisions
Exam pearls
- Normothermic target temperature management should not accelerate prognostication timelines
- Myoclonus alone is not a reliable predictor; correlate with EEG pattern
- Avoid haemolysed samples when interpreting neuron-specific enolase
- Concordance of multiple modalities, not any single finding, supports a poor prognosis
Disclaimer: Educational decision support for qualified clinicians and students — not a substitute for clinical judgement, local protocols or a full drug reference. Confirm every dose and contraindication against your formulary before prescribing.