Post-Cardiac Arrest Care (ROSC)
Return of spontaneous circulation is just the beginning. Post-cardiac arrest syndrome causes multi-organ dysfunction. Targeted temperature management, hemodynamic optimization, and finding the cause define survival and neurologic outcome.
🔍 Overview
Post-Cardiac Arrest Syndrome
After ROSC, four simultaneous processes occur:
- Post-arrest brain injury -cerebral edema, impaired autoregulation, seizures (10–35% of patients who do not follow commands) and myoclonus (18–34%)
- Post-arrest myocardial dysfunction -global systolic/diastolic dysfunction (peaks at 24–48 hrs, often reversible)
- Systemic ischemia/reperfusion injury -cytokine storm, coagulopathy, multi-organ dysfunction
- Precipitating cause -must identify and treat (ACS, PE, electrolytes, toxins)
Reversible Causes of Cardiac Arrest -The H's & T's
Mnemonic: 6 H's and 6 T's -Run through every H and T systematically during and after every arrest. If you don't find the cause, you can't fix it.
| H's | Clue / How to Find | Fix |
|---|---|---|
| Hypovolemia | Flat IVC on echo, trauma, GI bleed | Volume resuscitation, blood products, source control |
| Hypoxia | SpO₂, cyanosis, airway obstruction | Secure airway, ventilate, confirm ETT placement |
| Hydrogen ion (acidosis) | ABG -pH < 7.1, severe DKA, renal failure, sepsis | Bicarb, treat underlying cause |
| Hypo/Hyperkalemia | ECG (peaked T's, wide QRS, sine wave), BMP | Calcium, insulin/glucose, dialysis (hyperK). IV KCl, Mg (hypoK). |
| Hypothermia | Core temp < 30°C (86°F), exposure history, near-drowning | Active rewarming. "Not dead until warm and dead." |
| Hypoglycemia | Fingerstick glucose, diabetic on insulin/sulfonylureas, altered mental status pre-arrest | D50W 25–50 mL IV push (1–2 amps). Recheck glucose q15min. |
| T's | Clue / How to Find | Fix |
|---|---|---|
| Tension pneumothorax | Absent breath sounds, tracheal deviation, hypotension, distended neck veins | Needle decompression (adults: 4th-5th ICS just anterior to midaxillary line, ATLS 10th) → chest tube |
| Tamponade (cardiac) | Distended neck veins, muffled heart sounds, hypotension (Beck's triad). Echo: effusion + RV collapse. | Pericardiocentesis (subxiphoid, echo-guided) |
| Toxins | Med hx, pill bottles, toxidrome. Common: TCA, digoxin, beta-blockers, CCBs, opioids. | Specific antidotes. Bicarb (TCA), digibind, glucagon (BB), high-dose insulin (CCB), naloxone. |
| Thrombosis -PE | RV dilation on echo, history of immobility/DVT, PEA arrest | tPA 50mg IV push (if Category E PE during arrest), surgical/IR embolectomy |
| Thrombosis -MI | 12-lead ECG: STEMI or new LBBB, regional wall motion abnormality on echo | Emergent cath lab -even during CPR (in select patients) |
| Trauma | Mechanism of injury, external signs of hemorrhage, unstable pelvis, distended abdomen | Massive transfusion protocol, surgical intervention, pelvic binder. Address hemorrhagic shock. |
Immediately post-ROSC: 12-lead ECG (Class 1), then hunt for the cause. Persistent ST elevation → cath lab emergently, coma or not (neurologic status cannot be judged this early). Why "persistent": an ECG taken in the first minutes after ROSC over-calls STEMI; in one series about two-thirds of STEMI patterns on the immediate tracing had resolved on a repeat ECG roughly 30 minutes later, so repeat a very early tracing. Echo or POCUS → wall motion, RV dilation (PE), tamponade, EF (2025 AHA: may be reasonable, Class 2b). NEW 2025 Head-to-pelvis CT may be reasonable (Class 2b) when the cause is still unknown: it finds a clinically significant diagnosis or CPR injury in roughly 30% (intracranial hemorrhage, PE, aortic dissection, pneumothorax, liver or spleen laceration). Weigh that against the transport risk and the contrast load.
🚨 Management
What changed in the 2025 AHA post-cardiac arrest guideline (vs 2020). NEW 2025 Coronary angiography before discharge for a suspected cardiac cause is Class 1; routine emergency angiography without ST elevation is Class 3 (no benefit); immediate multivessel PCI in shock is Class 3 (no benefit); temporary mechanical circulatory support is not routine (Class 3) and only a Class 2b option in highly selected refractory shock; a dedicated myoclonus section (do not treat myoclonus that has no EEG correlate); neurofilament light (NfL) joins NSE; predictors of a favorable outcome; CT and echo or POCUS to find the cause (Class 2b). CHANGED Temperature control for at least 36 hours (2020: 24) across 32 to 37.5°C (89.6 to 99.5°F) (2020: 32 to 36°C (89.6 to 96.8°F)); SpO₂ 90–98% (2020: 92–98%); the blood pressure floor is MAP at least 65 alone, the SBP ≥ 90 target was removed. UNCHANGED Emergent angiography for ST elevation regardless of coma, PaCO₂ 35–45, prompt EEG, no seizure prophylaxis, multimodal prognostication no earlier than 72 hours after normothermia.
Post-Arrest Bundle -ICU Protocol
Immediately Post-ROSC
12-lead ECG as soon as feasible (Class 1): persistent ST elevation → cath lab without delay, regardless of coma. ABG, labs, portable CXR, bedside echo or POCUS (Class 2b), and head-to-pelvis CT when the cause is unclear (Class 2b). Establish arterial line + central line. If still on scene -transport immediately. Without ST elevation, emergency angiography is NOT routine NEW 2025 (Class 3, no benefit): COACT, 2019 and TOMAHAWK, 2021 found no benefit from immediate cath in stable comatose patients, and TOMAHAWK a trend toward harm. The exceptions (shock, recurrent VT/VF, ongoing ischemia) are in the angiography table below.
Airway & Ventilation
Confirm ETT position. Give 100% oxygen only until the SpO₂ or PaO₂ reads reliably (Class 1), then titrate FiO₂ to SpO₂ 90–98% (PaO₂ 60–105 mmHg) (Class 2a) CHANGED (2020: 92–98%). Avoid hyperoxia (increases reperfusion injury) and never allow hypoxemia (Class 1), which extends the ischemic injury. Target PaCO₂ 35–45 mmHg (Class 1) UNCHANGED -avoid hypocapnia (causes cerebral vasoconstriction); confirm with a blood gas rather than trusting end-tidal CO₂ alone. Pulse oximetry can overread in darker skin and hide hypoxemia, so check a gas when the number and the patient disagree.
Hemodynamics
MAP at least 65 mmHg (Class 1) CHANGED: 2025 dropped the 2020 SBP ≥ 90 target and kept a MAP floor alone, because that is what the trials tested. Why not aim higher: four randomized trials (over 1,000 patients, including BOX, 2022) found no gain in survival or neurologic outcome from MAP targets above 70, and their lower arms averaged about 65. The safety of aiming lower (60–65) has not been established, so do not drift under 65: autoregulation is impaired and hypotension becomes secondary brain injury. No vasopressor is preferred (Class 2b, insufficient evidence); norepinephrine is the usual first choice by extrapolation from cardiogenic and septic shock, where epinephrine and dopamine cause more arrhythmia. Dobutamine if post-arrest myocardial dysfunction. Avoid aggressive fluid loading in pulmonary edema.
Targeted Temperature Management (TTM)
The goal is preventing fever, not inducing hypothermia. TTM-2 2021 found no benefit of 33°C (91.4°F) over normothermia, so routine cooling to 33°C (91.4°F) is no longer the default. CHANGED 2025 2025 AHA: hold one constant target anywhere from 32 to 37.5°C (89.6 to 99.5°F) (Class 1) for at least 36 hours of total temperature control (Class 2a), in every adult who does not follow commands after ROSC, whatever the rhythm or arrest location. The 2020 guideline said 32-36°C (89.6 to 96.8°F) for at least 24 hours; the 36-hour floor comes from BOX, 2023, which found 36 hours no worse than 72. Running the full 72 hours (the TTM-trial protocol, and the ERC-ESICM 2021 standard of preventing T > 37.7°C (99.9°F) for 72 hours) remains acceptable; 36 is the minimum, not a stopping rule. Use a controlled device rather than relying on antipyretics. Why the device stays on: post-rewarming rebound fever is common, and fever even after the first 24 hours is associated with worse outcome. If a hypothermic target is chosen, rewarm no faster than 0.5°C/h (0.9°F/h) (Class 2b; trials used 0.25-0.5°C/h (0.45 to 0.9°F/h)), and rapid cold IV saline before hospital arrival is not recommended (Class 3, no benefit); in hospital it is only an induction adjunct, never the way a target is held. Cooling blankets, Arctic Sun, or intravascular devices all work.
Seizures & Myoclonus
Get an EEG promptly, and have it read promptly, in anyone who does not follow commands (Class 1): seizures or status occur in 10–35% of these patients and many have no clinical correlate. Repeat or continuous monitoring is reasonable (Class 2a) because seizures can first appear days after ROSC. Treat clinically apparent seizures (Class 1) and EEG-only (nonconvulsive) seizures (Class 2a), using the same antiseizure drugs as for seizures of any other cause (Class 2b). No routine seizure prophylaxis (Class 3, no benefit) UNCHANGED: two randomized trials of prophylaxis found no fewer seizures and no better neurologic outcome. For rhythmic or periodic patterns on the ictal-interictal continuum, a therapeutic trial of a nonsedating antiseizure drug may be reasonable (Class 2b); nonsedating so that the exam and the prognostication clock stay interpretable, and only a trial because in TELSTAR, 2022 aggressive suppression of these patterns did not change outcome (poor outcome 90% vs 92%). NEW 2025 Myoclonus: get an EEG (Class 1), and do not treat myoclonus that has no EEG correlate (Class 3, no benefit). Details in the myoclonus section below.
Neurological Prognostication
Multimodal, never a single finding (Class 1), and never early (Class 1). Consolidate the assessment at a minimum of 72 hours after return to normothermia AND after sedatives are off (Class 2a) UNCHANGED. Why both clocks: hypothermia slows clearance of propofol, midazolam and opioids, so an exam 72 hours after rewarming can still be a drug effect. NEW 2025 Neurofilament light (NfL) alongside NSE, and predictors of a favorable outcome. Neuro exam, SSEP, EEG, CT/MRI brain and serum biomarkers each contribute; no single test is definitive. Full table below.
Coronary Angiography After ROSC: Who Goes Now, Who Waits
The ECG decides the timing, not the troponin and not the coma. Elevated troponin and nonspecific ischemic ECG changes are nearly universal after an arrest and do not identify who benefits from an emergency angiogram. An absent ST elevation does not exclude a culprit lesion either, which is why the question is when, not whether: survivors with a suspected cardiac cause still get angiography before discharge.
| Scenario | 2025 AHA | vs 2020 | Why |
|---|---|---|---|
| Comatose, no ST elevation, stable The most common post-arrest picture | Emergency angiography is NOT recommended over a delayed or selective strategy (Class 3, no benefit). Angiography after awakening, or sooner if instability develops. | NEW 2025 2020 had no recommendation against routine emergent angiography; it predated TOMAHAWK | COACT, 2019: 90-day survival 64.5% vs 67.2%. TOMAHAWK, 2021: 30-day death 54.0% vs 46.0% with immediate angiography. An early trip to the cath lab delays temperature control and ICU care for no gain. In the delayed arms up to 40% never underwent angiography because they did not awaken, and up to 17% crossed over to early angiography for instability, which is the safety net. |
| Persistent ST elevation | Emergent angiography, regardless of coma (Class 1). | UNCHANGED 2025 adds the word "persistent" | ST elevation means transmural ischemia with a high probability of acute occlusion, and an untreated occlusion drives re-arrest. The angiography-timing trials all excluded STEMI, so nothing in them argues for waiting. Repeat an ECG taken within minutes of ROSC before activating, since early tracings over-call STEMI. |
| No ST elevation, but cardiogenic shock, recurrent VT/VF, or significant ongoing ischemia | Emergency angiography is reasonable, regardless of coma (Class 2a). | UNCHANGED 2020: reasonable for electrically or hemodynamically unstable patients | Hemodynamic or electrical instability after arrest is often severe ischemia, and these patients re-arrest if it is left untreated. Every timing trial allowed crossover to emergent angiography for exactly these findings. "Severe ischemia" has no agreed definition; TOMAHAWK used a troponin above 70 times the upper reference limit. |
| Any survivor with a suspected cardiac cause | Coronary angiography before hospital discharge (Class 1), particularly with an initial shockable rhythm, unexplained LV systolic dysfunction, or evidence of severe ischemia. | NEW 2025 | Delayed does not mean skipped. The trials that support waiting performed angiography in every patient who woke up. An unrecognized ischemic substrate also changes the ICD decision (see the next section). |
| Multivessel disease found, patient in shock Multivessel CAD in up to 50% of post-arrest angiograms | Treat the infarct-related artery only. Immediate PCI of non-infarct lesions is NOT recommended (Class 3, no benefit); stage the rest. | NEW 2025 | CULPRIT-SHOCK, 2017: death or dialysis 45.9% with culprit-only vs 55.4% with immediate multivessel PCI, with the same harm signal in the cardiac arrest subgroup. More contrast and a longer procedure in a patient in shock buys nothing. |
Temporary Mechanical Circulatory Support After ROSC (IABP, Impella, VA-ECMO) NEW 2025
- Not routine in post-arrest cardiogenic shock (Class 3, no benefit). Why: across 14 trials routine MCS did not improve survival in cardiogenic shock (OR 1.17, 95% CI 0.97–1.42), and in the 816 resuscitated cardiac arrest patients pooled from 11 of them it did not either (OR 1.21, 95% CI 0.91–1.60), while severe bleeding and ischemic limb complications rose. ECLS-SHOCK, 2023: early VA-ECMO, 30-day death 47.8% vs 49.0%, bleeding 23.4% vs 9.6%.
- May be considered in highly selected patients with refractory shock (Class 2b). The one positive trial, DanGer Shock, 2024, enrolled STEMI shock and excluded comatose cardiac arrest survivors, so the candidate it supports is the awake STEMI patient whose shock persists despite vasopressors and an inotrope, not the comatose post-arrest patient on a single pressor.
- If a device is in, it needs a team that runs them regularly (Class 1). Responses to fluids, pressors and ventilator changes can be paradoxical on support, and centers with higher device volume have better survival. Selection and device choice are covered on the cardiogenic shock page.
Myoclonus After ROSC NEW 2025
Myoclonus is a reason to get an EEG. It is not a diagnosis, and by itself it is not a prognosis. It appears in roughly 18–34% of post-arrest patients. Up to half of them have electrographic seizures or status epilepticus, and no feature of the jerks (pattern, timing, distribution) tells you which half. "Status myoclonus" is only a bedside label that applies until the EEG has characterized it.
| EEG finding | What it is | 2025 AHA | Why |
|---|---|---|---|
| Jerks time-locked to an EEG correlate Cortical myoclonus | An electroclinical myoclonic seizure or status epilepticus. | Treat as a seizure (treatment of clinically apparent seizures is Class 1), with the same antiseizure drugs used for other causes (Class 2b). See the status epilepticus protocol. | Untreated seizure activity raises metabolic demand in a brain whose supply is already impaired, so it is treated despite the absence of trial proof that treatment changes outcome. |
| Jerks with NO EEG correlate Subcortical myoclonus | A movement generated below the cortex. Not a seizure. | Treatment to suppress it is NOT recommended (Class 3, no benefit). | No evidence links it to secondary brain injury, and no study shows that suppressing it helps. It is often refractory, so chasing it means stacking antiseizure and anesthetic drugs whose hypotension and sedation are real harms and which confound the neurologic exam you are waiting to do. The guideline acknowledges that myoclonus can interfere with ventilation and nursing care; its recommendation is against suppression as a goal in itself. |
| EEG unreadable from muscle artifact | Undetermined. | Get the EEG promptly (Class 1); a short period of neuromuscular blockade under close monitoring can unmask the tracing. | The decision to treat depends entirely on whether a correlate exists, and the clinical exam alone cannot answer that. |
Prognosis: CHANGED in 2020 status myoclonus within 72 hours could support a poor prognosis; in 2025 the usefulness of undifferentiated myoclonus as a predictor is uncertain, even within 72 hours (Class 2b). Do not let jerking on day 1 anchor the family conversation.
Neuroprognostication: The 2025 Multimodal Table
Three rules before any test result is read. (1) Never a single finding (Class 1): every predictor below is Class 2b and counts only alongside others. (2) Not before 72 hours after normothermia AND off sedatives (Class 2a), and later still when clearance is slow (hypothermia, renal or hepatic injury), because a transiently poor exam and residual drug effect both imitate devastating injury. (3) Talk to the family early, regularly and transparently about the time course (Class 1), so that day 3 is an expected milestone and not an ambush. Withdrawing support on an early impression is a self-fulfilling prophecy.
| Modality | Supports an UNFAVORABLE outcome All Class 2b, only with other tests | Supports a FAVORABLE outcome NEW 2025 | Timing and traps |
|---|---|---|---|
| Clinical exam | Bilaterally absent pupillary light reflex at ≥ 72 h; quantitative pupillometry at ≥ 72 h; bilaterally absent corneal reflexes at ≥ 72 h. | A withdrawal motor response or better (Class 2b). Pupillometry for favorable outcome is not well established. | Preserved pupillary or corneal reflexes do NOT predict a good outcome (Class 3, no benefit): brainstem reflexes survive injuries the cortex does not. An absent or extensor motor response is of uncertain value (Class 2b), as is undifferentiated myoclonus: both are too sensitive to sedation and too often seen in patients who recover. |
| Serum biomarkers | High NSE or NfL within 72 h NEW 2025 (NfL added). NfL is more sensitive than NSE at similarly high specificity in head-to-head studies. | Normal NSE within 72 h (Class 2b; below 17 mcg/L between 24 and 72 h in the systematic review). | There is no universal cutoff: published NSE thresholds for poor outcome run from 17 to 106 mcg/L, NfL assays differ between manufacturers, and the optimal sampling time is unknown, so only very high values carry weight. Hemolysis falsely raises NSE (red cells contain it). Other biomarkers (S100B, GFAP, tau, UCH-L1): uncertain. |
| EEG | Status epilepticus persisting ≥ 72 h; burst suppression off sedating drugs at ≥ 72 h. | A continuous background without discharges within 72 h (Class 2b). | Do NOT use absent EEG reactivity within 72 h (Class 3, no benefit). Seizures as such, and rhythmic or periodic discharges, are of uncertain prognostic value. Burst suppression is also a propofol and midazolam effect, so the sedation history matters as much as the tracing. |
| SSEP | Bilaterally absent cortical N20 at ≥ 48 h. | N20 amplitude for favorable outcome is not well established. | The modality least affected by sedation, which is why it is ordered first when the exam is confounded. It needs a skilled operator: muscle artifact and electrical interference in the ICU produce false "absent" responses, so consider neuromuscular blockade during the recording. |
| Imaging | Reduced gray-white ratio on CT; extensive restricted diffusion or reduced ADC on MRI at 2 to 7 days. | No restricted diffusion on MRI at 2 to 7 days (Class 2b). Gray-white ratio for favorable outcome is not well established. | MRI diffusion changes evolve: a scan before day 2 can look falsely reassuring and one after day 7 pseudonormalizes. No validated gray-white ratio threshold exists, so a qualitative read of "loss of differentiation" is weak evidence alone. |
Why the favorable column matters at the bedside: prognostication used to be a search for reasons to stop. A patient who withdraws to pain, has a normal NSE, a continuous EEG background and a clean MRI at day 3 has four independent reasons to continue full support, even if still comatose. Absence of the unfavorable signs was always reassuring; in 2025 the guideline names the favorable ones.
Secondary Prevention ICD & Disposition
Every VT/VF survivor needs ICD evaluation before discharge unless the arrest was due to a completely reversible cause. Missing this step is a common disposition error and can be fatal, since recurrence risk is substantial.
Secondary prevention ICD is a Class I indication for any survivor of sudden cardiac arrest (VF or hemodynamically unstable sustained VT) NOT due to a transient or reversible cause. Trial basis: AVID 1997, CIDS 2000, CASH 2000. Only AVID was significant on its own; CIDS and CASH showed nonsignificant trends in the same direction, and the pooled Connolly meta-analysis, 2000 gives a 28% reduction in death with an ICD vs amiodarone (HR 0.72, 95% CI 0.60–0.87), driven by a halving of arrhythmic death.
When ICD IS indicated (Class I)
- Survivor of VF or hemodynamically unstable VT NOT from a clearly reversible cause
- Sustained VT with structural heart disease (ischemic or non-ischemic cardiomyopathy), whether or not hemodynamically tolerated
- Syncope with inducible sustained VT on EP study in a patient with structural heart disease
- Genetic arrhythmia syndromes with prior cardiac arrest: long QT, Brugada, ARVC, CPVT, HCM with prior arrest
When ICD is NOT indicated (reversible causes)
- Acute MI with complete revascularization and no residual ischemia, VF within 24-48h of MI is often attributable to acute ischemia alone and doesn't by itself mandate secondary ICD
- Electrolyte derangement (severe hypokalemia, hypomagnesemia), fully corrected, no structural substrate
- Drug-induced QT prolongation, offending drug stopped and QTc normalized off the agent
- Commotio cordis (blunt chest impact causing VF), structurally normal heart, no long-term risk
- Severe acid-base or metabolic disturbance, cause identified and corrected
- Illicit drug-induced arrest (cocaine, methamphetamine) with structurally normal heart and commitment to abstinence
Timing rules before implantation:
- Post-MI primary prevention: wait ≥ 40 days if EF persistently ≤ 35% (DINAMIT 2004, IRIS 2009, no benefit from earlier ICD). But secondary prevention ICDs can go in sooner if the arrest was NOT attributable to the acute MI itself.
- Post-revascularization (CABG/PCI): wait ≥ 90 days before primary prevention ICD assessment, EF often recovers with revascularization.
- New cardiomyopathy (ischemic or non-ischemic): ≥ 3 months of maximally tolerated GDMT before primary prevention ICD assessment, 30-40% of patients recover EF above 35% on GDMT.
- Neurologically devastated survivors: defer ICD, prognostication must first confirm meaningful neurologic recovery. Implanting in a patient who will not wake up is not appropriate.
Before-Discharge Checklist
- EP consultation to decide ICD vs no ICD
- Echocardiogram to document EF (primary vs secondary prevention logic differs)
- Coronary angiography if not already done: before discharge for any suspected cardiac cause (2025 AHA, Class 1), because an ischemic substrate changes both revascularization and the ICD decision
- Cardiac MRI if etiology unclear, detects ARVC, infiltrative disease, myocarditis, scar burden
- QTc measurement off all QT-prolonging drugs
- Toxicology screen and medication review
- Family history + genetic testing if channelopathy or ARVC suspected; family screening
- GDMT initiation if structural heart disease (ARNI/ACEi/ARB, BB, MRA, SGLT2i)
- Wearable cardioverter-defibrillator (LifeVest) bridge if ICD decision deferred (pending GDMT trial, revascularization recovery, neurologic prognostication), VEST 2018 evidence mixed; used pragmatically in many centers.
- Return precautions and AED-at-home counseling for the family
💊 Medications & Doses
Medications
| Drug | Dose | Role | Notes |
|---|---|---|---|
| Norepinephrine (Levophed) | 0.1–1 mcg/kg/min | Vasopressor (usual first choice) | Keep MAP at least 65 post-ROSC (2025 AHA, Class 1); higher MAP targets did not improve outcome in four trials. The guideline finds insufficient evidence to prefer any vasopressor after arrest (Class 2b); norepinephrine is chosen by extrapolation from cardiogenic and septic shock, where epinephrine and dopamine cause more arrhythmia. Post-arrest myocardial dysfunction is common -monitor CO. |
| Dobutamine (generic) | 2–15 mcg/kg/min | Post-arrest cardiogenic shock | Add if MAP adequate but echo shows severely reduced EF. Titrate to echo/clinical response. |
| Propofol (Diprivan) | 5–50 mcg/kg/min | Sedation during TTM | Reduces shivering, facilitates temperature control. Monitor for propofol infusion syndrome. |
| Meperidine / Buspirone | Meperidine 25–50 mg IV PRN | Anti-shivering during TTM | Shivering increases metabolic demand and raises temperature. Treat aggressively. Magnesium also helps. |
| Aspirin + Heparin | Per ACS protocol | If ACS precipitant | Do not withhold antiplatelet/anticoagulation for neurologic concerns alone. Treat the cause. |
| Insulin infusion | Target BG 140–180 mg/dL | Glycemic control | 2025 AHA: avoid both hypoglycemia (below 70 mg/dL) and hyperglycemia (above 180 mg/dL) (Class 2b). Tight control (< 110) increases hypoglycemia and worsens outcomes, and an injured brain tolerates low glucose poorly. |
| Levetiracetam / Valproate (Keppra / Depakote) | Per the status epilepticus protocol | Clinical or EEG seizures after ROSC | 2025 AHA: same antiseizure drugs as for seizures of any other cause (Class 2b); no prophylaxis (Class 3, no benefit). Prefer nonsedating agents so the neurologic exam and the 72-hour prognostication clock stay interpretable. Do not use them to chase myoclonus that has no EEG correlate (Class 3, no benefit). |
📋 On Rounds
On Rounds
Pimp Questions
Why do we avoid hyperoxia post-ROSC?
Hyperoxia (PaO₂ > 300 mmHg) worsens cerebral reperfusion injury through free radical generation. Multiple observational studies Kilgannon 2010 show increased mortality with hyperoxia post-ROSC. The harm is at the extremes: in BOX, 2022, a PaO₂ of 68–75 vs 98–105 mmHg made no difference. 2025 AHA: 100% oxygen only until the saturation reads reliably, then SpO₂ 90–98% (PaO₂ 60–105), never 100% by default.
When can you prognosticate neurological outcome after cardiac arrest?
No earlier than 72 hours after return to normothermia AND after sedatives have cleared (2025 AHA, Class 2a). Both clocks matter: hypothermia slows clearance of propofol, midazolam and opioids, so an exam 72 hours after rewarming can still be a drug effect. If no hypothermic target was used, count from ROSC, again off sedation. Sedation, hypothermia, and metabolic derangements all suppress neurological function and mimic poor prognosis. Early withdrawal based on exam alone leads to self-fulfilling prophecy.
What is targeted temperature management (TTM) and what does current evidence say?
TTM2, 2021: targeted hypothermia at 33°C (91.4°F) did NOT improve survival or neurologic outcome compared to normothermia (targeting ≤ 37.5°C (99.5°F) and actively preventing fever). This changed practice -the benefit is from preventing fever, not inducing hypothermia. Current approach (2025 AHA): one constant target between 32 and 37.5°C (89.6 to 99.5°F), held with a device for at least 36 hours of total temperature control; many units still run 72 h, which was the TTM-trial protocol. If using active cooling, 33°C (91.4°F) × 24h followed by fever prevention is still acceptable. Rewarm slowly (no faster than 0.5°C/hr (0.9°F/hr); 0.25°C/hr (0.45°F/hr) is common).
Which findings support an unfavorable neurologic prognosis after cardiac arrest, which support a favorable one, and what changed in 2025?
No single test is sufficient (Class 1); every predictor is Class 2b and counts only alongside others. Unfavorable: bilaterally absent pupillary or corneal reflexes at ≥ 72h; bilaterally absent N20 on SSEP at ≥ 48h; status epilepticus persisting ≥ 72h or burst suppression off sedation at ≥ 72h; high NSE or NfL within 72h (NfL is new in 2025); reduced gray-white ratio on CT; extensive restricted diffusion on MRI at days 2–7. Favorable (new in 2025): withdrawal or better motor response, normal NSE, a continuous EEG background without discharges, no restricted diffusion on MRI. An absent or extensor motor response and myoclonus are of uncertain value, and absent EEG reactivity should not be used.
A comatose patient develops generalized myoclonic jerks 12 hours after ROSC. What do you do, and what do you NOT do?
Get an EEG promptly (Class 1). Myoclonus follows 18–34% of arrests, up to half of those patients have electrographic seizures or status, and the exam cannot tell which. If the jerks are time-locked to an EEG correlate it is a seizure: treat it. If there is no EEG correlate, do NOT treat to suppress it (Class 3, no benefit; new in 2025): nothing links it to secondary brain injury, it is often refractory, and the sedating drugs needed cause hypotension and confound the neurologic exam. Myoclonus alone is of uncertain prognostic value, even within 72 hours.
ROSC after a VF arrest. The patient is comatose and hemodynamically stable, and a repeat ECG shows no ST elevation. Cath lab now?
No. Emergency angiography is not recommended over a delayed or selective strategy (2025 AHA, Class 3, no benefit): COACT, 2019 and TOMAHAWK, 2021 found no survival benefit, and TOMAHAWK a trend toward harm (30-day death 54.0% vs 46.0%). Go emergently for persistent ST elevation (Class 1) or, without it, for cardiogenic shock, recurrent ventricular arrhythmias or significant ongoing ischemia (Class 2a). A survivor with a suspected cardiac cause still gets angiography before discharge (Class 1).
Clinical Examples
Case 1: ROSC After VF Arrest with STEMI
Presentation: 58M found unresponsive at home. EMS: VF, defibrillated × 2, ROSC after 18 minutes of CPR. Arrives intubated. 12-lead ECG shows inferior STEMI.
Priorities: Cath lab activated immediately, PCI does not wait for neurologic prognostication. Post-PCI, initiate temperature control: a constant target ≤ 37.5°C (99.5°F) with a device for at least 36 h (72 h in many protocols) and active fever prevention per TTM2, 2021. Titrate FiO₂ to SpO₂ 90–98%, avoid hyperoxia and hypoxia. MAP ≥ 65 with Norepinephrine; add Dobutamine if echo shows post-arrest low EF.
Teaching point: STEMI post-arrest goes to the cath lab regardless of coma. Post-arrest myocardial stunning is reversible, echo EF often normalizes by 48–72h. Rewarm slowly (0.25°C/hr (0.45°F/hr)) to avoid rebound hyperthermia. If the angiogram shows multivessel disease in a patient in shock, stent the culprit only: immediate multivessel PCI is Class 3 (no benefit) in the 2025 guideline on CULPRIT-SHOCK, 2017 data. Persistent shock on a pressor and an inotrope is when temporary mechanical support is discussed (Class 2b, highly selected), not before.
Case 2: Neurologic Prognostication, Day 3 Post-Arrest
Presentation: 65F, out-of-hospital VF arrest, ROSC after 12 minutes. Temperature control completed (normothermia protocol with a feedback device). Now day 4: more than 72 hours after normothermia and 48 hours off propofol and fentanyl, still comatose. GCS 4 (E1V1M2, extensor posturing). Absent pupillary reflexes bilaterally. EEG shows burst suppression with no sedation running. Family asking about prognosis.
Workup: SSEP (valid from ≥ 48 h): bilateral N20 waves absent, the most specific single test and the one least affected by sedation, but never used alone. MRI brain on day 4: extensive restricted diffusion. NSE markedly elevated, NfL high (NfL was added beside NSE in the 2025 AHA guideline: more sensitive at similar specificity, no universal cutoff, assays differ). CT head: no hemorrhage, diffuse cerebral edema with loss of gray-white differentiation.
Management: The timing rule is satisfied only because both clocks have run: at least 72 h after normothermia and sedatives cleared; with renal or hepatic injury, or after a hypothermic target, wait longer. Convene multidisciplinary meeting (neurology, ICU, palliative). Absent bilateral N20 SSEPs + absent pupillary reflexes at ≥ 72 h + burst suppression off sedation + MRI and biomarkers = convergent unfavorable prognosis across modalities. Note what does not count in 2025: her extensor motor response and any myoclonus are of uncertain value, and absent EEG reactivity should not be used (Class 3). Family meeting with honest, compassionate goals-of-care discussion; the guideline asks for early, regular, transparent updates about the time course (Class 1), not a single day-3 verdict.
Teaching point: No single test predicts outcome, multimodal assessment required. Early withdrawal is a self-fulfilling prophecy. The mirror image matters too: a withdrawal-or-better motor response, a normal NSE, a continuous EEG background or an MRI without restricted diffusion each support a favorable outcome (new in 2025) and are reasons to keep going.
Case 3: Hypoxic Cardiac Arrest from Opioid OD
Presentation: 45M found unresponsive, suspected fentanyl overdose. EMS: PEA, ROSC after 8 minutes of CPR and Naloxone administration. Non-shockable rhythm. Arrives intubated, GCS 6.
Priorities: No cardiac cause, PCI not indicated. ECG: no STEMI. Bedside echo: normal EF, no wall motion abnormality. CT head to rule out intracranial pathology. Target SpO₂ 90–98%, normocapnia (PaCO₂ 35–45). Fever prevention protocol initiated. EEG promptly (Class 1): treat clinical or electrographic seizures, give no prophylactic antiseizure drugs (Class 3), and do not treat myoclonus that has no EEG correlate.
Goals of care: 8 minutes of CPR is relatively short, prognosis may be better than prolonged arrest. Defer prognostication until at least 72h after normothermia and off sedation (opioids on board make this patient's early exam especially unreliable). If no meaningful neurologic recovery expected on reassessment, early goals-of-care conversation with family. Address substance use disorder, consider Buprenorphine initiation prior to discharge if recovery occurs.
Teaching point: Hypoxic arrests (OD, asphyxia) receive the same post-arrest bundle, TTM, avoid hyperoxia, EEG monitoring. The cause determines whether you go to the cath lab, not the post-arrest protocol itself.
📋 Sample Presentation
"Mr. Perez is a 55-year-old who suffered an out-of-hospital VF arrest with 18 minutes of CPR before ROSC. He was brought to our ICU intubated and not following commands. Initial and repeat ECG showed persistent inferior ST elevation -he underwent emergency PCI with drug-eluting stent to the RCA; a moderate LAD lesion was left for staged assessment because he is in shock. He is on our temperature control protocol at a constant 33°C (91.4°F), hour 14; any constant target from 32 to 37.5°C (89.6 to 99.5°F) meets the 2025 guideline. He is on NE 0.2 mcg/kg/min, MAP 68. Echo shows EF of 30% with inferior wall hypokinesis -on dobutamine 5 mcg/kg/min; no mechanical support unless the shock becomes refractory. SpO₂ 96% on FiO₂ 0.4, PaCO₂ 41. Continuous EEG shows no seizure activity and no myoclonus; no antiseizure prophylaxis. BG 152, on insulin drip. We will hold 33°C (91.4°F) for 24 hours, rewarm at 0.25°C per hour (0.45°F per hour), never faster than 0.5°C per hour (0.9°F per hour), then continue device-based fever prevention to 72 hours from ROSC; the guideline floor is 36 hours of total temperature control. Neurological prognostication is deferred until at least 72 hours after return to normothermia and off sedation, and it will be multimodal."
Monitoring Parameters
| Parameter | Frequency | Target / Action |
|---|---|---|
| Vitals | q4h floor, q1–2h ICU | HR, BP, RR, SpO₂, Temp -notify for significant deviations |
| Labs (BMP, CBC) | Daily AM or as indicated | Trend Cr, K⁺, WBC, Hgb -adjust treatment based on trajectory |
| Disease-specific markers | Per clinical context | See Overview and Management tabs for condition-specific targets |
| I&Os | Strict if volume-sensitive | UOP ≥ 0.5 mL/kg/hr. Net fluid balance guides diuresis or resuscitation. |
| Telemetry | Continuous if indicated | Arrhythmia detection. Discontinue when no longer indicated (reduces alarm fatigue). |
| Clinical response | Each assessment | Symptom improvement, functional status, appetite, mental status -the exam matters more than labs |
Don't just order labs -act on them. Every lab should have a clear clinical question. If you wouldn't change management based on the result, don't order it.
⚡ Summary
Summary
ROSC Is the Start, Not the End
Post-cardiac arrest syndrome is a whole-body ischemia-reperfusion injury: brain injury, myocardial dysfunction, systemic inflammation resembling sepsis, and the persisting precipitating illness. Neurologic outcome, not survival to ICU, is the outcome that matters, and it is largely determined in the first 24 h.
ECG Immediately, Then Find the Cause
Persistent ST elevation means the cath lab now, even in a patient who looks neurologically devastated (Class 1), since prognostication that early is unreliable and the occluded artery is the treatable cause. Without ST elevation, emergency angiography is not routine (Class 3, no benefit; COACT, TOMAHAWK) unless there is cardiogenic shock, recurrent ventricular arrhythmia or ongoing ischemia (Class 2a); a survivor with a suspected cardiac cause still gets angiography before discharge (Class 1). Echocardiography gives wall motion, RV dilation suggesting pulmonary embolism, tamponade and ejection fraction in minutes at the bedside, and head-to-pelvis CT is reasonable when the cause is still unknown.
Run the H's and T's Systematically
Hypoxia, hypovolemia, hydrogen ion (acidosis), hypo- and hyperkalemia, hypothermia, hypoglycemia; toxins, tamponade, tension pneumothorax, thrombosis (coronary and pulmonary), trauma. If you do not find the cause you cannot fix it, and the arrest will recur.
Do Not Hyperoxygenate or Hyperventilate
Target a saturation of 90 to 98% (2025 AHA; 2020 said 92 to 98%): hyperoxia worsens reperfusion injury and hypoxemia extends the ischemic one. Target normocapnia, a PaCO₂ of about 35 to 45: hypocapnia causes cerebral vasoconstriction and worsens ischemia, and hypercapnia raises intracranial pressure. Wean the FiO₂ down deliberately rather than leaving the patient on 100% after ROSC.
Control Temperature Actively
Temperature control with one constant target between 32 and 37.5°C (89.6 to 99.5°F) for at least 36 hours (2025 AHA), and above all avoid fever. TTM2 found 33°C (91.4°F) was not superior to normothermia with active fever prevention, so the emphasis is on preventing hyperthermia with a device rather than on deep cooling; many units still run the 72 hours used in the trials. Treat shivering, which raises metabolic demand and defeats the target.
Support the Hemodynamics and the Glucose
Keep the MAP at least 65 (Class 1): higher targets did not improve outcome in four randomized trials, and the safety of going lower has not been established. No vasopressor has proven better than another, so norepinephrine is used by extrapolation from other shock states, with an inotrope for the stunned myocardium, which typically recovers over 24 to 72 h. Temporary mechanical support is not routine (Class 3) and is reserved for highly selected refractory shock. Keep glucose between 70 and 180 mg/dL and treat hypoglycemia promptly.
Do Not Prognosticate Too Early
Wait at least 72 h after return to normothermia AND until sedatives have cleared, since cold slows drug clearance and a sedated exam imitates a devastating injury. Use multiple modalities, never one finding: clinical examination, EEG, somatosensory evoked potentials, neuron-specific enolase or neurofilament light (added in 2025) and imaging. The 2025 guideline also names favorable signs: a withdrawal-or-better motor response, normal NSE, a continuous EEG background, no restricted diffusion on MRI. Early pessimism becomes a self-fulfilling prophecy, since it leads to withdrawal of the support that would have permitted recovery.
EEG Early; Treat Seizures, Not Jerks
Anyone who does not follow commands gets a prompt EEG (Class 1), because 10 to 35% seize and many seizures are visible only on the tracing. Treat clinical and electrographic seizures; give no prophylaxis (Class 3), which has never reduced seizures or improved outcome. Myoclonus with no EEG correlate should not be treated (Class 3, new in 2025): it is not known to injure the brain, and the sedating drugs it takes to suppress it cause hypotension and cloud the exam you are waiting to perform.
Plan Secondary Prevention Before Discharge
An ICD is a Class I indication for any survivor of sudden cardiac arrest from VF or hemodynamically unstable sustained VT not due to a transient or reversible cause (AVID, CIDS, CASH). The reversible-cause exception matters: an arrest from hyperkalemia or an acute STEMI treated by revascularization may not need a device. Arrange cardiology follow-up, family screening where an inherited channelopathy or cardiomyopathy is possible, and rehabilitation.
📄 One Pager
Text version
📄 One Pager
ICU · One Pager
Post-Cardiac Arrest Care (ROSC)
ROSC is the beginning, not the end. ECG → echo → TTM → avoid hyperoxia → find the cause. No prognosis until 72h after normothermia AND off sedation.
Post-ROSC Bundle
1
12-lead ECG -persistent ST elevation → cath lab now, even if comatose (Class 1). No ST elevation and stable: emergent cath is NOT routine (Class 3)
2
Bedside echo -EF, wall motion, tamponade, RV dilation (PE)
3
Avoid hyperoxia -target SpO₂ 90–98%, NOT 100% (free radical injury) Kilgannon 2010
4
MAP ≥ 65 (Class 1) -no proven best pressor, NE usual. Dobutamine if post-arrest low EF. Mechanical support is not routine.
5
Temperature control: one constant target 32-37.5°C (89.6 to 99.5°F) for ≥ 36h (2025 AHA; 72h still common), never let them get febrile. Routine cooling to 33°C (91.4°F) is no longer the default TTM-2 2021
6
Prompt EEG -10–35% seize, often nonconvulsive. No prophylaxis; do not treat myoclonus that has no EEG correlate
6 H's & 6 T's
- Hypovolemia, Hypoxia
- Hydrogen ion, Hypo/hyperK
- Hypothermia, Hypoglycemia
- Tension PTX, Tamponade
- Toxins, Thrombosis (PE/MI)
- Trauma
Prognostication
- ≥ 72h after normothermia AND off sedatives
- Neuro exam + SSEP + EEG
- MRI (day 2–7) + NSE or NfL
- No single test is definitive
- Favorable signs count too (2025)
Pitfalls
- Hyperoxia (FiO₂ 1.0 left on)
- Early prognostication
- Missing post-arrest NCSE
- Delaying PCI for STEMI
- Routine emergent cath without ST elevation
- Treating myoclonus with no EEG correlate