18.2 Global Anoxia-Ischemia and Post-Resuscitation Encephalopathy (03.A.1)
Key Takeaways
- AHA 2023 adult ACLS focused update: after ROSC, unresponsive adults receive a deliberate temperature-control strategy, selecting and maintaining a constant temperature between 32 °C and 37.5 °C for at least 24 hours, with fever prevention; TTM2 found no mortality benefit of 33 °C versus 37.5 °C.
- Multimodal neuroprognostication is delayed until ≥72 hours after ROSC and after rewarming and other confounders are gone; never withdraw life-sustaining treatment on a single modality.
- A highly specific clinical pair at ≥72 hours is absent pupillary light reflexes AND absent corneal reflexes, interpreted with other tests — not in isolation.
- Other poor-outcome predictors used together include highly malignant EEG, bilaterally absent N20 SSEPs, NSE with assay-specific cutoffs (older ERC/ESICM teaching cited ≥60 ng/mL / µg/L), and diffuse anoxic injury on CT or MRI.
- Early myoclonus is not automatically futile: status myoclonus on a burst-suppressed EEG differs from Lance–Adams physiology with a continuous background and vertex spike-waves.
Global Anoxia-Ischemia and Post-Resuscitation Encephalopathy
Quick Answer: For adults who do not follow commands after ROSC, AHA 2023 recommends temperature control at a constant 32–37.5 °C for ≥24 hours, plus fever prevention. TTM2 showed 33 °C versus 37.5 °C did not change 6-month death (~50% vs 48%). Prognosticate multimodally at ≥72 hours after ROSC, after rewarming and off confounders. Never withdraw on one test. Absent pupils AND corneals at ≥72 hours, highly malignant EEG, bilateral absent N20, NSE (lab-specific; older ≥60 ng/mL), and diffuse DWI/ADC or lost gray–white are used together. Myoclonus is not one disease.
Cardiac arrest produces global hypoxic-ischemic brain injury: cortical and deep-gray vulnerability, watershed ischemia, and a post-resuscitation cascade of reperfusion, edema, seizures, fever, and secondary hypotension. Independent OpenExamPrep material here covers global anoxia-ischemia and post-resuscitation encephalopathy as listed among Neurocritical care diseases in the ABPN Content Specifications. This guide is not an ABIM or ABPN product.
Temperature control is not “must cool to 33”
The 2023 American Heart Association Focused Update on Adult Advanced Cardiovascular Life Support changed the language from older targeted temperature management (TTM) slogans. For adults who do not follow commands after return of spontaneous circulation (ROSC), regardless of arrest location or initial rhythm, use a deliberate temperature-control strategy. Select and maintain a constant temperature between 32 °C and 37.5 °C, generally for at least 24 hours after the target is reached. Hospitals should have a protocol. Fever prevention after that window remains part of care in patients who stay unresponsive. Patients who are already spontaneously hypothermic should not be rapidly rewarmed faster than about 0.5 °C per hour as a routine.
TTM2 (Dankiewicz and colleagues, NEJM 2021) randomized comatose adults after out-of-hospital arrest of presumed cardiac or unknown cause to 33 °C versus targeted normothermia with early fever treatment (device cooling if temperature reached 37.8 °C, target 37.5 °C). Death at 6 months was 465/925 (50%) with hypothermia versus 446/925 (48%) with normothermia (relative risk 1.04, P=0.37). mRS ≥4 at 6 months was 55% in both arms. Hemodynamically important arrhythmia was more common at 33 °C (24% vs 17%). That trial is why AHA raised the upper bound of the control range to 37.5 °C and why “everyone to 33 °C” is outdated teaching. It is also why “do nothing about temperature” is wrong: the control arm was active fever prevention, not permissive hyperthermia.
Supportive post-ROSC care still includes MAP support, oxygenation and ventilation without hypocapnic fashion statements, coronary angiography when an ACS phenotype is present, seizure detection, and glucose control in a moderate ICU band. Those topics sit in other chapters; they still change neurologic outcome more than an extra degree of hypothermia in an unselected TTM2-like patient.
Prognostication: wait, strip confounders, use more than one test
Early death after arrest is often cardiac or shock. Neurologic withdrawal of life-sustaining treatment (WLST) is the other large mortality pathway, and it is contaminated by a self-fulfilling prophecy if you treat an early, confounded exam as destiny. European Resuscitation Council / European Society of Intensive Care Medicine (ERC/ESICM) 2021 post-resuscitation guidance, which U.S. neurointensivists still use as the operational multimodal map, starts the poor-outcome algorithm in a still-comatose patient with motor response no better than stereotypic flexion (GCS motor ≤3) at ≥72 hours after ROSC, after major confounders are gone: residual sedation, neuromuscular blockade, hypothermia, and severe metabolic derangement. If you used cooling, the 72-hour clock for the clinical prognostic exam is after rewarming, not while the patient is 33 °C and on a propofol infusion.
Poor outcome is considered likely when two or more of the following predictors are present in that setting (tests may be recorded earlier; they are acted on at the delayed assessment):
| Predictor | Usual timing for interpretation | What “poor” looks like | Critical caveat |
|---|---|---|---|
| Clinical exam | ≥72 h after ROSC, off confounders | No pupillary light reflex AND no corneal reflex | Motor score alone is sensitive but not specific; pupils or corneals in isolation are weaker than the pair |
| EEG | Recorded >24 h; used at the delayed assessment | Highly malignant pattern: suppressed background (<10 µV) ± periodic discharges, or burst-suppression (including identical bursts) | Sedation and hypothermia produce burst-suppression; read after those fade |
| SSEP | Often from ≥24 h | Bilateral absence of the cortical N20 after median-nerve stimulation | Technical failure, cervical injury, and neuromuscular blockade (use a twitch monitor) can erase N20s |
| NSE | 48 h and/or 72 h in ERC/ESICM 2021 | Older citation >60 µg/L (ng/mL) on specified platforms | Assay- and lab-specific; hemolysis falsely raises NSE; do not import 60 as a universal number |
| Imaging | CT often after 24 h; MRI when feasible | Loss of gray–white differentiation on CT; diffuse cortical and deep-gray DWI restriction / very low ADC | Early CT can be falsely reassuring; a single “dark cortex” slice is not a protocol |
| Status myoclonus | Within 72 h, lasting ≥30 min | Continuous generalized myoclonus, usually with a malignant EEG | See Lance–Adams exception below |
Never single-modality WLST. An isolated high NSE, an isolated “bad EEG” on day 1, or myoclonus in the first night is not a withdrawal note. Combine concordant tests, wait, and say “indeterminate” when the cluster is incomplete. Quantitative pupillometry can support the pupillary exam but does not replace multimodal reasoning.
Neuron-specific enolase (NSE) deserves extra humility. ERC/ESICM 2021 cited >60 µg/L at 48 and/or 72 hours on the platforms used in the prognostic literature (often Roche or Thermo-Fisher in later validation work). Different immunoassays do not share a cutoff. Serial rising values are more worrisome than a single modest elevation. Hemolyzed samples are garbage. Treat NSE as one vote, labeled with your laboratory’s method.
Myoclonus: status myoclonus versus early Lance–Adams
Post-anoxic myoclonus is common and is not a single prognostic entity.
Status myoclonus (sometimes called myoclonus status epilepticus in older papers) is continuous, generalized, often synchronous jerking for 30 minutes or more, typically in the first 24–72 hours, on an EEG with suppression-burst and high-amplitude polyspikes locked to the jerks (Elmer “Pattern 1”). That cluster is a poor-outcome predictor in the ERC/ESICM list — still one predictor, not a solo death sentence if everything else is unexpectedly reassuring, but it is not “just twitching.”
Lance–Adams syndrome is action (intention) myoclonus as consciousness returns, often after a hypoxic (including respiratory) arrest. Cognition may recover while the jerks disable the limbs. The acute EEG clue that this physiology can appear early, even before obvious awakening, is a continuous background with narrow vertex/midline spike-waves locked to jerks (Elmer “Pattern 2”), often with preserved N20s and reactivity. Those patients can remain unresponsive for 1–2 weeks and still recover awareness. Treat the myoclonus (levetiracetam, valproate, benzodiazepines, later agents such as perampanel or cannabinoids in refractory clinic care) and wait. Do not WLST because the first night was myoclonic if the EEG is continuous and SSEPs are present.
Worked stem
ROSC at 02:10. Cooled to 33 °C for 24 hours, rewarmed by 08:00 on day 2. On day 2, off sedation, GCS motor 2, pupils and corneals present, EEG is continuous and reactive, NSE pending. A colleague wants to withdraw because “anoxic brain injury, GCS 4.” The correct move is continue supportive care and complete multimodal testing at ≥72 hours. Presence of pupils and a continuous EEG already forbid a single-exam withdrawal.
Exam traps
Cooling every patient to 33 °C as mandatory after TTM2/AHA 2023. Prognosticating at 24 hours on propofol. Using NSE 60 ng/mL from a paper as if every hospital assay matched it. Calling all myoclonus status myoclonus. Withdrawing for bilateral N20 absence alone without the rest of the cluster and a valid study. Independent practice at /practice/abim-neurocritical-care should test the 72-hour, two-predictor rule.
A comatose adult after OHCA is not following commands. Which temperature plan matches the AHA 2023 focused update and TTM2?
When may a poor neurologic outcome be judged likely enough to support a withdrawal discussion after cardiac arrest?
Which finding is used as a poor-outcome predictor only as part of a multimodal assessment, not as sole grounds for WLST?
On night 1 after a respiratory arrest, a patient has frequent myoclonic jerks. EEG shows a continuous background with narrow vertex spike-waves; N20 potentials are present. What is the most accurate interpretation?