post-cardiac arrest care

in review 7 min read Updated 2026-10-29
post-cardiac arrest care

After ROSC the injury continues: brain (hypoxic–ischaemic), heart (stunning), whole-body reperfusion, plus whatever caused the arrest. ECG now — STEMI → cath lab even if comatose; no ST elevation → delayed angiography unless shock, recurrent VT/VF or ongoing ischaemia. Then protect the brain: MAP ≥65, normoxia, normocapnia, glucose control, prevent fever (≤37.5 °C for ≥72 h), EEG for seizures. No prognosis before 72 h after normothermia with sedation off, and never from one test.

post-ROSC ICU bundle
  • ST elevation with suspected cardiac cause → emergency angiography, even if comatose
  • no ST elevation: emergency angiography only for shock, recurrent VT/VF or ongoing ischaemia; otherwise before discharge if cardiac cause
  • echo; CT head (± chest/abdomen/pelvis) if cause unclear or CPR trauma suspected
  • arterial line; MAP ≥65 mmHg (norepinephrine ± inotrope)
  • SpO₂ 90–98% (ERC: 94–98%; avoid hyperoxaemia); PaCO₂ 35–45 mmHg; Vt 6–8 mL/kg
  • not following commands: core temp ≤37.5 °C ≥72 h (AHA 32–37.5 °C ≥36 h); rewarm ≤0.5 °C/h
  • glucose 4–10 mmol/L
  • not following commands: early EEG; treat clinical and EEG seizures; no seizure prophylaxis
  • no prophylactic antibiotics
  • sedation minimised for neuro exam (titrate to ventilator synchrony and shivering)
  • no neuroprognosis <72 h after normothermia with sedation cleared; multimodal only
  • early family meeting on the timeline of uncertainty
  • survivors: cause work-up, ICD assessment, driving restrictions, cardiac rehab

post-cardiac arrest syndrome

  • brain injury: cerebral oedema, impaired autoregulation (needs adequate MAP), seizures; hyperthermia, hypoxaemia, hyperoxaemia, hypocapnia and hypoglycaemia all worsen secondary injury
  • myocardial dysfunction: global stunning, worst at 24–72 h, usually recovers → may need inotrope early; doesn’t by itself prove ACS
  • systemic ischaemia–reperfusion: vasoplegia, capillary leak, coagulopathy, AKI — looks like sepsis
  • precipitating cause: ACS, primary arrhythmia, PE, hypoxaemia, haemorrhage, electrolytes, toxins → find and treat

coronary angiography

situationtiming
ST elevation (suspected cardiac cause), comatose or notemergent primary PCI; fibrinolysis if timely PCI unavailable
no ST elevation + shock, recurrent VT/VF or ongoing ischaemiaemergent (AHA 2a)
no ST elevation, comatose, stabledelayed or selective — no benefit from immediate angiography (COACT (2019), TOMAHAWK (2021))
survivor with suspected cardiac cause (esp. shockable rhythm, unexplained LV dysfunction)before discharge (AHA I)
  • start temperature control alongside angiography, not after
  • STEMI after arrest → STEMI management; shock after ROSC → cardiogenic shock (routine MCS not recommended)
  • comatose STEMI with features predicting poor neurological outcome (unwitnessed, no bystander CPR, non-shockable, prolonged ROSC time, severe acidosis/lactate) → individualise

haemodynamics, oxygen and ventilation

  • MAP ≥65 mmHg (AHA I; CCS); higher targets no better — BOX (2022): MAP 77 vs 63 mmHg → same death or severe disability
  • vasopressor of choice: norepinephrine; add inotrope for low output → vasoactive drugs
  • oxygen: 100% until measured, then SpO₂ 90–98% (AHA: PaO₂ 60–105; CCS: 68–105, avoid <68 and >195 mmHg; ERC: SpO₂ 94–98%, PaO₂ 75–100); pulse oximetry can hide hypoxaemia in darker skin
  • PaCO₂ 35–45 mmHg — hypocapnia → cerebral vasoconstriction; check ABG, not just ETCO₂. Lung-protective Vt 6–8 mL/kg (ERC)

temperature control

  • all patients not following commands after ROSC, any rhythm, in or out of hospital
  • target: ERC-ESICM 2025 — prevent fever, ≤37.5 °C for ≥72 h; CCS 2024 — “reasonable” to target ≤37.5 °C and continue temperature control ≥72 h (its strong rec for 33–36 °C applies if hypothermic TTM is chosen); AHA 2025 — any temperature 32–37.5 °C, for ≥36 h (AHA I / 2a)
  • TTM2 (2021): n=1900, 33 °C vs normothermia (≤37.5 °C, active cooling at ≥37.8 °C) for 40 h → 6-month death 50.3% vs 48.2%, no difference in poor function; more haemodynamically significant arrhythmia with 33 °C (23.9% vs 16.5%) → fever prevention is the goal
  • start in hospital; prehospital cold IV fluids not recommended (pulmonary oedema, re-arrest)
  • continuous core temperature monitoring; surface or intravascular devices perform similarly (intravascular: more bleeding in one RCT); treat shivering
  • avoiding rewarming faster than 0.5 °C/h may be reasonable (AHA 2b)
  • uncertain whether sicker subgroups do better with hypothermia
  • spontaneous mild hypothermia (32–36 °C) after ROSC: don’t actively rewarm (ERC)

seizures, other treatments

  • EEG promptly and repeatedly/continuously in patients not following commands; nonconvulsive seizures common
  • treat clinical seizures and EEG-only seizures: levetiracetam or valproate first line plus sedation (ERC); no seizure prophylaxis; trial of a non-sedating drug for ictal–interictal patterns may be reasonable
  • generalized periodic discharges alone → unlikely to benefit from suppression (CCS); no seizure prophylaxis
  • antibiotics: no routine prophylaxis (AHA: uncertain benefit; CCS: reserve for suspected pneumonia)
  • steroids for post-arrest shock: uncertain; no proven neuroprotective drug
  • glucose: avoid <3.9 and >10 mmol/L

neuroprognostication

Most deaths after ROSC follow withdrawal of life-sustaining measures for a presumed poor neurological prognosis, so a false-positive prediction is the error that matters. Hence: wait, exclude confounders, and require agreement across modalities.

  • no definitive prediction in the first 24 h (CCS 2023, strong). Formal assessment at ≥72 h after ROSC in anyone not obeying commands (ERC-ESICM 2025), once sedation, NMBA, hypothermia and metabolic or organ-failure confounders are excluded. Delay further if confounded or if seizures are still being treated
  • withdrawal of life-sustaining measures needs adverse findings in ≥2 modalities (CCS 2023, strong; ERC-ESICM: ≥2 predictors)
modalitypredicts poor outcomecaveat
exambilaterally absent pupillary reflexes at ≥72 h (CCS strong); absent corneals at ≥72 h (weaker); quantitative pupillometry from ≥48 h (weak, thresholds unclear)motor M1–2 alone is not poor prognosis (high false-positive rate); motor score is an entry criterion only
myoclonusstatus myoclonus within 72 h (ERC) / first 7 days (CCS), only combined with other findingsget an EEG: some myoclonus (Lance-Adams, reactive background) is compatible with recovery
EEGhighly malignant patterns after 24 h (ERC; CCS weak, suppressed background only ≥72 h): suppressed background ± periodic discharges, burst suppression (AHA: burst suppression off sedation at ≥72 h)electrographic status epilepticus alone is not definitive → treat it, delay prognostication. Don’t use absent EEG reactivity
SSEPbilaterally absent N20 (from ≥24 h CCS/ERC; AHA ≥48 h)needs a comatose patient (no purposeful pain response); NMBA reduces artefact
imagingCT within 72 h: low grey–white ratio (quantitative, e.g. <1.15–1.20); MRI within 7 days if still uncertain: extensive diffusion restrictionrepeat CT at 72–96 h if the first one was early and normal
biomarkersNSE >60 µg/L at 48 and/or 72 h, or a rising trend (ERC-ESICM)CCS: not routinely (assay and threshold variability). Neurofilament light: not yet (ERC-ESICM); AHA 2025: high NSE or NfL within 72 h may be considered (2b)
  • normal SSEPs, a continuous reactive EEG or low NSE favour recovery and argue for more time
  • an indeterminate picture at 72 h → keep supporting and reassess; late awakening happens
  • regular, early, transparent family discussions about the timeline and the uncertainty
  • not the same as death determination → death determination and organ donation (prerequisites, 48 h rule after ROSC)

refractory arrest (persisting VF/pVT)

  • amiodarone or lidocaine after ≥1 failed shock, particularly witnessed arrest
  • after ≥3 failed shocks: double sequential or vector-change defibrillation if equipment and training available
  • eCPR (VA-ECMO): not routine; selected patients — witnessed arrest, limited comorbidity, refractory >5–15 min, arrest-to-cannulation <60 min and hospital arrival-to-cannulation <30 min, at centres with an established pathway (CCS)

after survival

Prepared 2026-10-29. Not yet verified — check every number against the sources.

Key references

+4 more sources

Suggest a correction post-cardiac arrest care