Never decide on prognosis early. At ≥ 72 h after ROSC (and after sedation clears), combine examination, EEG, SSEP, NSE and imaging; poor outcome is likely only when two or more predictors agree.
At the bedside
Timed steps- 1
Do not prognosticate
Day 0-3Exclude confounders: sedatives, neuromuscular block, hypothermia, severe hypotension, hypoglycaemia, uraemia, hepatic failure.
- 2
Collect data
Day 1-3Daily neurological exam (GCS motor, pupils, corneals, myoclonus). EEG (continuous or intermittent). NSE at 24, 48 and 72 h. SSEP at ≥ 24 h. CT brain early; MRI at 2-7 days.
- 3
Assess unconscious patient
≥ 72 hConsider prognostication in any patient not awake and obeying commands (GCS motor < 6) at ≥ 72 h, with no confounders.
- 4
Look for ≥ 2 poor-outcome predictors
≥ 72 hBilaterally absent pupillary and corneal reflexes; bilaterally absent N20 SSEP; highly malignant EEG (suppressed background ± periodic discharges, or burst-suppression) > 24 h; NSE > 60 µg/L at 48 or 72 h; status myoclonus ≤ 72 h; diffuse anoxic injury on CT/MRI.
- 5
Indeterminate if < 2
InterpretContinue support and reassess; look for favourable signs (motor withdrawal or better, NSE < 17 µg/L, continuous/reactive EEG, normal MRI).
- 6
Shared decision
DecideMultidisciplinary, documented, with family; separate brain death testing (module 21) from prognostication.
Predictors of poor neurological outcome (ERC-ESICM 2025)
| Test | Timing | Poor-outcome finding |
|---|---|---|
| Pupillary + corneal reflexes | ≥ 72 h | Bilaterally absent (quantitative pupillometry preferred) |
| SSEP | ≥ 24 h | Bilaterally absent N20 waves |
| EEG | > 24 h | Highly malignant pattern: suppressed background with or without periodic discharges, or burst suppression |
| NSE | 48 and 72 h | > 60 µg/L |
| Myoclonus | ≤ 72 h | Status myoclonus (with EEG correlate) |
| CT / MRI | CT early; MRI 2-7 days | Diffuse, extensive anoxic injury (loss of grey-white differentiation, diffusion restriction) |
References
- 1
- 2
- 3
ERC-ESICM prognostication strategy
Step 1 of ~4
Decision: Confounders excluded (sedation, temp, metabolic)?
Based on Nolan JP, Sandroni C, Cariou A, et al. Educational use; follow local protocol.
Evidence
- ERC-ESICM 2025 extend prognostication to all unconscious patients (GCS motor ≤ 5, previously ≤ 3) and keep the multimodal strategy at ≥ 72 h with ≥ 2 concordant predictors of poor outcome; NSE > 60 µg/L at 48/72 h; favourable signs include NSE < 17 µg/L and continuous EEG background.
- AHA 2025 endorses multimodal prognostication no earlier than 72 h after normothermia and sedation clearance; NSE and NfL may be used.
Controversies
- NSE thresholds vary between analysers and studies; specificity is never 100%.
- Withdrawal of life-sustaining therapy and organ donation pathways differ greatly between countries.
🇮🇳 India notes
- SSEP, cEEG and NSE are unavailable in many Indian ICUs; rely on serial clinical examination, CT/MRI, and time. When tests are missing, be more conservative and allow longer observation.
- Families may need extended counselling; involve senior staff early. Indian law (Common Cause 2018 and 2023 modifications) allows withdrawal of life support with medical board process; see module 39.
References
- 1
- 2
- 3
Test yourself
Module quiz · score 70% to complete the module
Educational reference for trained clinicians. Check doses against your hospital protocol and the patient in front of you.