Neuroprognostication After Cardiac Arrest

 Neuroprognostication After Cardiac Arrest

A Grand Rounds Review for Postgraduate Trainees and Practicing Consultants

Dr Neeraj Manikath , claude.ai

Introduction: A Question That Haunts the ICU

A 58-year-old man is brought to your emergency department following a witnessed out-of-hospital cardiac arrest. Bystander CPR was initiated within two minutes; the rhythm was ventricular fibrillation, and return of spontaneous circulation (ROSC) was achieved after 22 minutes of resuscitation. He is intubated, haemodynamically stable on vasopressors, and transferred to the ICU for targeted temperature management (TTM). On day three, his pupils are sluggishly reactive, he withdraws to pain, and the EEG shows burst suppression. His wife, clutching his hand, asks you:

"Will he wake up? Will he be the same person I married?"

This scene replays thousands of times weekly in ICUs worldwide. Cardiac arrest survivors represent one of the most heterogeneous and prognostically challenging populations in all of acute medicine. Approximately 10–15% of out-of-hospital cardiac arrest patients survive to hospital discharge in high-income countries, yet among those who achieve ROSC, up to 70% will die or survive with severe neurological disability. The single most consequential decision a clinician makes in this setting is not which vasopressor to choose — it is whether, when, and how to prognosticate neurological outcome.

This review distils the current evidence into a practical, multimodal framework for neuroprognostication — one grounded in pathophysiology, sharpened by clinical pearls, and calibrated by hard-won bedside wisdom.

 

Pathophysiology: What You Need to Know to Prognosticate

Global cerebral ischaemia during cardiac arrest triggers a cascade of excitotoxicity, mitochondrial failure, and inflammatory injury that continues — and may worsen — after ROSC. This 'post-cardiac arrest brain injury' (PCABI) unfolds in temporally distinct phases, each with prognostic relevance.

Phase 1 (0–6 hours): Reperfusion injury dominates. Reactive oxygen species and glutamate-mediated excitotoxicity peak. Clinically, this window is unreliable for prognostication — even patients with seemingly intact neurological signs may deteriorate.

Phase 2 (6–72 hours): Delayed neuronal death, cerebral oedema, and apoptosis evolve. EEG abnormalities and elevated biomarkers (NSE, S100B) become diagnostically informative. TTM exerts its neuroprotective effect during this phase by attenuating inflammation and metabolic demand.

Phase 3 (>72 hours): Structural injury is largely fixed. This is the earliest reliable window for multimodal prognostication. The degree of cortical connectivity — reflected in SSEP responses, EEG reactivity, and biomarker trajectories — predicts functional recovery.

The key clinical take-away: any prognostic assessment performed before 72 hours after ROSC (or before 72 hours after TTM ends) risks both false pessimism and false optimism. Sedation, hypothermia, and ongoing organ dysfunction confound nearly every clinical sign in the early window.

 

🪙  CLINICAL PEARLS

Pearl 1: Absent pupillary reflexes are the most robust early sign — but interpret bilaterality strictly.

Bilateral absent pupillary light reflexes (PLR) at ≥72 h post-ROSC carry ~100% specificity for poor outcome. However, unilateral absence or sluggish responses are not sufficient for prognostication. Always examine both pupils, in a darkened room, with a bright light source. The quantitative pupillometer (NPi) reduces observer variability and identifies subtle responses invisible to the naked eye.

Pearl 2: Motor responses are unreliable after TTM.

Extension posturing (GCS motor = 2) was once considered a marker of poor prognosis. Post-TTM data shows that up to 15% of patients with good outcomes demonstrate extensor responses on day 3 — almost certainly a residual sedation effect. Never use motor response in isolation.

Pearl 3: Status myoclonus ≠ Lance-Adams syndrome.

Early (within 24–48 h) generalised myoclonus, especially if continuous and accompanied by a malignant EEG, predicts poor neurological outcome. This is distinct from Lance-Adams syndrome — action myoclonus emerging days-to-weeks later in a patient regaining consciousness — which is compatible with a good eventual outcome. Misclassifying these two entities causes both premature WLST and inappropriate reassurance.

 

🦪  OYSTERS — HIDDEN GEMS

Oyster 1: The 'self-fulfilling prophecy' bias is real and kills salvageable patients.

Studies from the TTM trial era demonstrate that early withdrawal of life-sustaining treatment (WLST) — before multimodal assessment — accounts for a substantial proportion of ICU deaths after cardiac arrest. When clinicians prognosticate early and communicate pessimism, families consent to WLST, making the prognosis appear correct. The antidote is a structured, timed, multidisciplinary prognostication protocol that defers WLST decisions.

Oyster 2: NSE is powerful — but laboratory variation destroys its utility.

Neuron-specific enolase (NSE) is the most validated serum biomarker post-arrest. A value >60 µg/L at 48–72 h has high specificity for poor prognosis. However, haemolysis (from red cell lysis) raises NSE by up to 40 µg/L. Always request a concurrent haemolysis index. An NSE result without haemolysis correction is diagnostically worthless.

Oyster 3: A normal EEG on day 1 is not reassuring in the way most clinicians think.

Continuous EEG monitoring in the first 24 hours primarily serves to detect non-convulsive seizures — a treatable secondary injury — rather than to confirm good prognosis. An EEG that is normal at 12 hours may show malignant patterns at 36 hours as sedation clears. A single snapshot EEG is insufficient; 24-hour continuous monitoring is the standard of care in comatose post-arrest patients.

 

⚡  CLINICAL HACKS & TIPS

Hack 1: Use the '3-3-3 rule' for prognostication timing.

Prognosticate no earlier than: 3 days after ROSC, and at least 3 days after TTM ends, using at least 3 independent modalities. This triple-three heuristic encodes the ERC 2021 guidelines into a bedside memory aid.

Hack 2: The 'traffic light' framework for bedside communication.

Use a three-tier verbal framing with families: RED = multiple poor prognostic signs consistently pointing to severe injury; AMBER = mixed signals, uncertainty is honest; GREEN = reassuring signs emerging. This prevents binary thinking (will he live or die?) and aligns families with the temporal reality of neurological recovery.

Hack 3: Quantitative pupillometry — the NPi threshold to remember.

A Neurological Pupil index (NPi) <2 bilaterally at 72 h has specificity approaching 100% for CPC 4–5 outcome. If your unit has a pupillometer (and it should), document NPi serially from day 1. An NPi that falls from 3 to <2 over 24 hours is a clinically significant deterioration signal.

 

🔬  STATE-OF-THE-ART UPDATES

Update 1: TTM2 trial — the death of '33 is better than 36'.

The landmark TTM2 trial (Dankiewicz et al., NEJM 2021, n=1900) demonstrated no difference in all-cause mortality between TTM at 33°C versus normothermia (targeted at ≤37.8°C). This overturned a decade of 33°C dogma. The take-home: fever prevention (>37.7°C strictly avoided) remains essential; active cooling to 33°C is no longer mandated but remains an option.

Update 2: EEG standardisation — the ACNS terminology revolution.

The American Clinical Neurophysiology Society (ACNS) standardised critical care EEG terminology has transformed post-arrest EEG interpretation. 'Malignant' EEG patterns — suppression, burst-suppression, isoelectric trace, or absence of EEG reactivity — are now reproducible descriptors with validated prognostic weight. Training in this terminology is no longer optional for intensivists managing cardiac arrest survivors.

Update 3: GFAP and NfL — the next-generation biomarkers.

Glial fibrillary acidic protein (GFAP) and neurofilament light chain (NfL) are emerging as superior biomarkers to NSE. NfL in particular shows excellent discrimination for poor neurological outcome at 48–72 h and is unaffected by haemolysis. While not yet in routine clinical use, expect these to enter guidelines within 2–3 years.

 

Diagnostic Nuances: Separating Good from Great Clinicians

The multimodal prognostication algorithm recommended by the ERC 2021 guidelines stratifies assessment into four complementary domains:

1. Clinical neurological examination: At ≥72 h post-ROSC (after sedation washout confirmed by pharmacokinetic assessment), evaluate: (a) PLR — bilaterally absent is a 'major' predictor; (b) corneal reflex — bilaterally absent; (c) motor response — ≤2 alone is insufficient but contributes to the constellation. Crucially, document the timing of last sedative/paralytic dose and calculate expected clearance. A pragmatic bedside test: can the patient follow commands to 'open eyes' or 'squeeze my hand'? Consistent command-following is a favourable sign even without full consciousness.

2. Electrophysiology — SSEP: Bilateral absence of the N20 cortical response on short-latency somatosensory evoked potentials (SSEPs) is the single most specific prognostic test available (specificity ~100%, FPR <1% with TTM-corrected interpretation). However, sensitivity is only ~45%, meaning a present N20 does not guarantee good outcome. SSEPs should ideally be performed at ≥24 h after normothermia restoration.

3. Neuroimaging — CT and MRI: Non-contrast CT within 24–48 h assesses grey-white matter ratio (GWR). A GWR <1.2 in the basal ganglia region predicts poor prognosis with high specificity. Brain MRI at 2–5 days using DWI/ADC mapping is more sensitive: diffuse cortical or deep grey matter restriction indicates severe hypoxic injury. An important pitfall: a normal early CT does not exclude significant injury — MRI is the definitive imaging modality.

4. Biomarkers — NSE and emerging tests: NSE at 48 h >33 µg/L (moderate concern) and >60 µg/L (high specificity for poor outcome) should be interpreted with haemolysis index. Serial rising values are more concerning than a single elevated reading. GFAP and NfL are promising adjuncts where available.

 

The Prognostication Algorithm at a Glance

Step 1: Confirm adequate observation period (≥72 h post-ROSC, ≥72 h post-TTM end). Step 2: Exclude confounders (residual sedation, metabolic encephalopathy, haemodynamic instability). Step 3: Apply clinical examination. Step 4: Obtain EEG (continuous 24 h) and SSEP. Step 5: Request neuroimaging (CT or MRI). Step 6: Measure NSE at 48 and 72 h. Step 7: Synthesise findings across modalities. Step 8: Convene multidisciplinary family meeting with palliative care input.

 

Management Intricacies: Drugs, Doses, and Pitfalls

Temperature management: Target 32–36°C for 24 hours, then controlled rewarming at 0.25–0.5°C per hour to 37°C. Following TTM2, normothermia protocols target ≤37.7°C using active cooling for a minimum of 72 h post-ROSC. Avoid fever aggressively — even a single hour of temperature >38°C has been associated with worsened neurological outcomes in observational data.

Sedation and analgesia: During TTM, use short-acting agents: propofol and remifentanil or fentanyl are preferred for their rapid offset. Avoid benzodiazepines where possible — their prolonged effect confounds neurological assessment. Conduct daily sedation holds with clinical reassessment once normothermia is established.

Seizure management: Continuous EEG should guide treatment. Non-convulsive status epilepticus (NCSE) occurs in up to 30% of comatose survivors. Treat with IV levetiracetam (loading dose 60 mg/kg, max 4,500 mg) as first-line, followed by valproate or lacosamide. Prophylactic anticonvulsants are not routinely recommended. Distinguish NCSE from myoclonic status (which is often refractory and carries a poor prognosis regardless of treatment).

Haemodynamic targets: Target MAP ≥65–70 mmHg; avoid hypotension (MAP <65 mmHg) scrupulously. Some centres target MAP 80–100 mmHg for the first 24 h based on cerebrovascular autoregulation loss post-arrest. SpO2 94–98%; PaCO2 35–45 mmHg (avoid hypocapnia — cerebral vasoconstriction worsens ischaemia). Hyperoxia (PaO2 >300 mmHg) should be avoided.

 

🚨  WHEN TO ESCALATE / WHEN TO WATCH

Escalate to neurology/neurocritical care immediately if:

• Continuous EEG shows NCSE or malignant patterns requiring expert interpretation

• Bilateral absent N20 on SSEP — requires specialist confirmation before any WLST discussion

• Refractory myoclonic status epilepticus

• Family requesting second opinion on prognosis

Watch and wait (do NOT prognosticate early) if:

• Sedation/paralytic clearance not confirmed pharmacokinetically

• Metabolic derangement present (uraemia, hyponatraemia, hepatic failure)

• Only 48 h has elapsed since ROSC — even with apparently catastrophic signs, wait

• Mixed prognostic signals (e.g., absent PLR but present N20 — these are conflicting and demand reassessment)

• Signs of neurological improvement emerging (spontaneous eye opening, tracking, purposeful movement)

 

Summary: Multimodal Prognostication at a Glance

Domain	Key Points
TTM	Target 32–36 °C for 24 h; avoid fever (>37.7 °C) for ≥72 h post-arrest
Earliest safe prognostication	≥72 h after ROSC (or ≥72 h after TTM ends)
Pupillary reflexes	Bilateral absence = poor prognosis (Se 20%, Sp ~100%)
SSEP N20	Bilateral absence = poor prognosis (Sp ~100% when TTM-corrected)
EEG	Burst suppression / malignant patterns at 24 h → poor prognosis
NSE	>60 µg/L at 48–72 h → poor prognosis (use with other modalities)
CT/MRI Brain	GWR <1.2 (CT) or DWI restriction (MRI) → poor prognosis
Clinical signs to avoid	Motor response alone (confounded by TTM and sedation)
Self-fulfilling prophecy	Avoid early WLST before multimodal assessment
Family communication	Use 'waiting for certainty' framing; avoid premature prognostication

 

Mnemonic: PROGNOSE

●       P — Pupils (bilaterally absent PLR = strong poor predictor)

●       R — Reflex (corneal, oculocephalic — absent bilaterally adds weight)

●       O — Only after 72 hours (never prognosticate early)

●       G — GFAP/NSE (biomarkers with haemolysis correction)

●       N — Neuroimaging (CT GWR <1.2; MRI DWI restriction)

●       O — Only multimodal (no single test decides)

●       S — SSEP (absent N20 = most specific poor predictor)

●       E — EEG (malignant patterns, absence of reactivity)

 

References

1. Nolan JP, Sandroni C, Böttiger BW, et al. European Resuscitation Council and European Society of Intensive Care Medicine guidelines 2021: post-resuscitation care. Intensive Care Med. 2021;47(4):369–421.

2. Sandroni C, D'Arrigo S, Callaway CW, et al. The rate of brain death and organ donation in patients resuscitated from cardiac arrest: a systematic review and meta-analysis. Intensive Care Med. 2016;42(6):942–51.

3. Westhall E, Rossetti AO, van Rootselaar AF, et al. Standardized EEG interpretation accurately predicts prognosis after cardiac arrest. Neurology. 2016;86(16):1482–90.

4. Oddo M, Sandroni C, Citerio G, et al. Quantitative versus standard pupillary light reflex for early prognostication in comatose cardiac arrest patients: an international prospective multicenter double-blind study. Intensive Care Med. 2018;44(12):2102–11.

5. Dragancea I, Rundgren M, Englund E, Friberg H, Cronberg T. The influence of induced hypothermia and delayed prognostication on the mode of death after cardiac arrest. Resuscitation. 2013;84(3):337–42.

6. Moseby-Knappe M, Mattsson-Carlgren N, Stammet P, et al. Serum markers of brain injury can predict good neurological outcome after out-of-hospital cardiac arrest. Intensive Care Med. 2021;47(6):984–94.

7. Callaway CW, Donnino MW, Fink EL, et al. Part 8: post-cardiac arrest care: 2015 American Heart Association guidelines update for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation. 2015;132(18 Suppl 2):S465–82.

8. Rossetti AO, Tovar Quiroga DF, Juan E, et al. Electroencephalography predicts poor and good outcomes after cardiac arrest: a two-center study. Crit Care Med. 2017;45(7):e674–82.

9. Nielsen N, Wetterslev J, Cronberg T, et al. Targeted temperature management at 33°C versus 36°C after cardiac arrest. N Engl J Med. 2013;369(23):2197–206.

10. Dankiewicz J, Cronberg T, Lilja G, et al. Hypothermia versus normothermia after out-of-hospital cardiac arrest. N Engl J Med. 2021;384(24):2283–94.

11. Sandroni C, Cariou A, Cavallaro F, et al. Prognostication in comatose survivors of cardiac arrest: an advisory statement from the European Resuscitation Council and the European Society of Intensive Care Medicine. Intensive Care Med. 2014;40(12):1816–31.

12. Bongiovanni F, Romagnosi F, Barbella G, et al. Standardized EEG analysis to reduce the uncertainty of outcome prognostication after cardiac arrest. Intensive Care Med. 2020;46(5):963–72.

13. Tsetsou S, Novy J, Oddo M, Rossetti AO. Multimodal outcome prognostication after cardiac arrest and targeted temperature management: analysis at 36°C. Neurocrit Care. 2018;28(1):104–9.

14. Sandroni C, D'Arrigo S, Cacciola S, et al. Prediction of good neurological outcome in comatose survivors of cardiac arrest: a systematic review. Intensive Care Med. 2022;48(4):389–413.

15. Cronberg T, Brizzi M, Liedholm LJ, et al. Neurological prognostication after cardiac arrest — recommendations from the Swedish Resuscitation Council. Resuscitation. 2013;84(7):867–72.