Continuous EEG in the ICU
Continuous EEG monitoring (cEEG) has become standard of care in the neurologic ICU. It detects non-convulsive seizures (present in 10–35% of obtunded ICU patients), monitors the depth of anesthetic-induced burst-suppression during refractory status, tracks evolution of encephalopathy, and provides early warning of delayed cerebral ischemia after subarachnoid hemorrhage. Modern cEEG combines raw EEG review with qEEG dashboards, allowing real-time bedside monitoring by ICU staff supplemented by daily review by trained electrophysiologists. This page covers the cEEG indications, workflow, ACNS terminology, key patterns, and the integration of qEEG into ICU practice.
Indications for cEEG
Strong Indications
- Unexplained altered mental status in ICU patient.
- Convulsive status epilepticus (monitor for transition to non-convulsive after convulsions resolve).
- Subtle motor activity suggesting subclinical seizures (twitching, eyelid flutter).
- Patient on neuromuscular blockade who could be having unobserved seizures.
- Post-cardiac arrest hypothermia and rewarming periods.
- Subarachnoid hemorrhage with risk of delayed cerebral ischemia.
- Refractory status epilepticus during anesthetic-infusion treatment.
- Pediatric/neonatal ICU brain monitoring.
Reasonable Indications
- Acute stroke with unexpected level of consciousness change.
- Traumatic brain injury with suspected seizures.
- Encephalitis evaluation.
- Sepsis with unexplained encephalopathy.
- Withdrawal of anti-seizure medications.
Detection Yields
| Patient Population | Seizure Detection Yield |
|---|---|
| Convulsive SE (post-event) | ~50% |
| Comatose ICU patients (unexplained) | ~25% |
| Acute stroke + altered consciousness | ~15–25% |
| TBI + altered consciousness | ~20% |
| SAH (post-coil/clip) | ~15% |
| Post-cardiac arrest | ~20–30% |
| Sepsis-related encephalopathy | ~10% |
Duration of Recording
- Yield is time-dependent:
- 30 minutes: ~50% of seizures detected (in patients who have them).
- 24 hours: ~85%.
- 48 hours: ~95%.
- ACNS recommends ≥24 hours of cEEG for non-comatose patients.
- ≥48 hours for comatose patients.
- Longer (up to 72+ hours) for higher-risk populations.
Workflow
Setup
- Full 10-20 array (21 electrodes).
- Standard impedance check (≤5 kΩ).
- Continuous video synchronization.
- Quality assurance review by EEG technician.
Monitoring
- Real-time qEEG dashboard at bedside.
- Trained ICU nurse oversight of qEEG trends.
- EEG technician review of raw EEG periodically (every 2–4 hours).
- Daily review by neurologist or neurophysiologist.
- Alerts for: new seizures, evolving rhythmic patterns, background changes.
Reporting
- ACNS-standardized terminology.
- Daily progress notes summarizing changes.
- Specific recommendations for treatment adjustment.
ACNS Terminology Application in cEEG
- Background (continuous, nearly continuous, discontinuous, burst-suppression, suppression).
- Frequency (delta, theta, alpha, beta).
- Voltage (low, normal, high).
- Reactivity (reactive, unreactive).
- Sleep architecture (preserved, absent).
- Patterns: PDs (LPD, GPD, BIPD), RDA (GRDA, LRDA), SW, BS.
- Modifiers: +F (with fast features), +S (with sharp features), +R (with rhythmic features).
- Prevalence: continuous, abundant, frequent, occasional, rare.
- Seizures: present, suspected, absent.
Key Patterns to Recognize
Subclinical Seizures
- Rhythmic activity at 2 Hz or faster, evolving in frequency or distribution, lasting at least 10 seconds.
- Sometimes subtle: brief 5–30 second bursts with rhythmic onset.
- Must distinguish from periodic discharges (PD) and rhythmic delta activity (RDA).
- Real-time qEEG dashboard flags rhythmic patterns for review.
Generalized Periodic Discharges (GPDs)
- Symmetric, bilateral periodic discharges.
- Common in post-cardiac arrest, sepsis-associated encephalopathy, advanced metabolic.
- +F modifier (with fast activity): higher seizure risk.
- Treatment varies: not all GPDs require treatment.
Lateralized Periodic Discharges (LPDs)
- Unilateral periodic discharges at 0.5–2 Hz.
- Often after focal stroke, hemorrhage, herpes encephalitis, tumor.
- “PLEDs+” with embedded fast: more likely ictal.
- Sometimes part of ictal-interictal continuum.
Bilateral Independent Periodic Discharges (BIPDs)
- Independent periodic patterns over both hemispheres.
- Suggests bilateral disease (severe sepsis, anoxic injury, herpes encephalitis).
- Poor prognosis often.
Generalized Rhythmic Delta Activity (GRDA, formerly FIRDA)
- Rhythmic delta at 1.5–2.5 Hz, generalized, often anterior maximum.
- Sign of diffuse encephalopathy.
- Relatively benign compared to GPDs.
SIRPIDs (Stimulus-Induced Rhythmic, Periodic, or Ictal Discharges)
- Patterns triggered by external stimulation.
- Common with bedside exam, suctioning, position changes.
- Ictal status uncertain.
Quantitative EEG (qEEG) for cEEG
Compressed Spectral Array (CSA)
- Shows 12–24 hours of EEG on one screen.
- Time on x-axis, frequency on y-axis, color = power.
- Allows rapid scanning for seizure activity, periodic patterns, trends.
Alpha-Delta Ratio
- Continuously calculated ratio.
- Decreases hours before delayed cerebral ischemia in SAH.
- Used to trigger early intervention.
Burst-Suppression Ratio
- Percentage of recording in suppression.
- Continuous monitoring during anesthetic infusion.
- Target 50–80% for refractory status epilepticus treatment.
Seizure Detection Algorithms
- Modern software flags potentially ictal patterns for review.
- Sensitivity ~70–85%; specificity ~80–90%.
- Complement, not replace, human review.
Asymmetry Index
- Quantifies left-right power differences.
- Detects lateralized changes that visual review might miss.
cEEG in Specific Conditions
Post-Cardiac Arrest
- cEEG during targeted temperature management and rewarming.
- Look for: seizures (treat if confirmed), background changes (track for prognostication), specific patterns (myoclonic status epilepticus).
- Continue ≥72 hours post-rewarming for prognostication.
Status Epilepticus
- cEEG essential during anesthetic-infusion treatment.
- Monitor burst-suppression depth.
- Detect breakthrough seizures.
- Guide weaning when seizure-free for 24+ hours.
Subarachnoid Hemorrhage
- Daily alpha-delta ratio tracking.
- Detect delayed cerebral ischemia 6–12 hours before clinical decline.
- Continue for typical vasospasm window (days 4–14).
Stroke
- cEEG in patients with altered consciousness disproportionate to imaging.
- Detect subclinical seizures (~15–25%).
- Monitor for hemorrhagic conversion or extension.
Traumatic Brain Injury
- cEEG ≥24 hours for moderate-severe TBI with altered consciousness.
- Detect non-convulsive seizures (~20%).
- Treat subclinical SE to prevent secondary injury.
Sepsis-Associated Encephalopathy
- cEEG for unexplained encephalopathy.
- Detect subclinical seizures (~10%).
- Address reversible contributors.
Pediatric ICU cEEG
- Higher seizure detection rates in pediatric ICU.
- Different patterns in different age groups.
- Specialized pediatric electrophysiology review required.
- Neonatal: aEEG and full cEEG combined.
Practical Challenges
Artifact in ICU
- Ventilator artifact (rhythmic at ventilator rate).
- IV pump artifact.
- Electrical equipment noise.
- Staff movement artifact.
- Patient movement (intubation, suctioning, position changes).
- Mitigation: regular electrode checks, room layout planning, video correlation.
Long-Recording Volume
- 24 hours = 60,000+ pages of EEG.
- qEEG compression essential.
- Targeted review of marked events.
- Automated alarms for changes.
Communication
- Daily rounds discussion of cEEG findings.
- Clear treatment recommendations.
- Documentation of when monitoring can be discontinued.
🔍 Did You Know?
The recognition that 10–35% of comatose ICU patients have non-convulsive seizures — patients whose seizures would be entirely missed without continuous EEG monitoring — has transformed the standard of care in neurologic ICU practice. Before widespread cEEG adoption, these patients were often treated for “metabolic encephalopathy” while their seizures continued, contributing to ongoing brain injury and prolonged ICU stays. cEEG studies in the 2000s demonstrated that treating subclinical seizures improves outcomes: reduced morbidity, shorter time to recovery of consciousness, lower mortality in some subgroups. The clinical impact is so substantial that the ACNS guideline now recommends cEEG (rather than serial routine EEGs) for any ICU patient with unexplained altered mental status, particularly with risk factors (acute brain injury, recent convulsive status, sepsis). The lesson is profound: routine EEG misses the majority of subclinical seizures because they’re temporally distributed, while cEEG captures them by sampling all 24 hours. For practicing intensivists and neurologists, the take-home is that asking “is there subclinical status?” requires more than a 30-minute EEG — it requires 24+ hours of monitoring. Modern ICUs with cEEG capability have dramatically improved care for the obtunded patient. The same principle drives the push for cEEG in resource-limited settings where it’s not yet standard — the diagnostic and therapeutic gap is real.
Pitfalls and Pearls
- cEEG indications: any unexplained altered consciousness in ICU; convulsive SE follow-up; post-cardiac arrest; SAH; TBI; sepsis.
- Detection yield: 50% at 30 min, 85% at 24 hr, 95% at 48 hr.
- Minimum duration: 24 hr non-comatose, 48 hr comatose.
- 10–35% of obtunded ICU patients: have subclinical seizures.
- Real-time qEEG dashboard: nurses monitor; technologist reviews periodically; neurologist daily.
- ACNS terminology: standardized vocabulary for reporting patterns.
- Compressed spectral array: 24 hours visualizable on one screen.
- Alpha-delta ratio: tracks SAH delayed cerebral ischemia.
- Burst-suppression ratio: tracks anesthetic-infusion depth.
- SIRPIDs: stimulus-induced; ictal status uncertain.
- GPD+F: higher seizure risk than pure GPD.
- Treat subclinical seizures: improves outcomes.
- Post-cardiac arrest: cEEG ≥72 hr post-rewarming for prognostication.
- Refractory SE: cEEG essential during anesthetic infusion.
- Pediatric cEEG: higher detection rates; specialized review.
- Artifact mitigation: regular electrode checks; video correlation.
- Multimodal monitoring: cEEG + clinical exam + neuroimaging + laboratory.
References
- Herman ST, Abend NS, Bleck TP, et al. Consensus statement on continuous EEG in critically ill adults and children, part I: indications. J Clin Neurophysiol. 2015;32(2):87-95.
- Claassen J, Mayer SA, Kowalski RG, Emerson RG, Hirsch LJ. Detection of electrographic seizures with continuous EEG monitoring in critically ill patients. Neurology. 2004;62(10):1743-1748.
- Hirsch LJ, Fong MWK, Leitinger M, et al. American Clinical Neurophysiology Society’s Standardized Critical Care EEG Terminology: 2021 Version. J Clin Neurophysiol. 2021;38(1):1-29.
- Vespa PM, Nuwer MR, Nenov V, et al. Increased incidence and impact of nonconvulsive and convulsive seizures after traumatic brain injury as detected by continuous electroencephalographic monitoring. J Neurosurg. 1999;91(5):750-760.
- Westhall E, Rossetti AO, van Rootselaar AF, et al. Standardized EEG interpretation accurately predicts prognosis after cardiac arrest. Neurology. 2016;86(16):1482-1490.
- Hirsch LJ, Brenner RP. Atlas of EEG in Critical Care. Wiley-Blackwell; 2010.