ICU EEG: Prognosis in...

ICU EEG: ICU EEG: Prognosis in AdultsPrognosis in Adults

Susan T. Herman, MDSusan T. Herman, MD

Assistant Professor of NeurologyAssistant Professor of Neurology

Beth Israel Deaconess Medical CenterBeth Israel Deaconess Medical Center

Harvard Medical SchoolHarvard Medical School

Boston, MABoston, MA

DisclosuresDisclosures

None relevant to this presentationNone relevant to this presentation Scientific Advisory BoardScientific Advisory Board

Eisai Inc.Eisai Inc. Biotie, Inc.Biotie, Inc.

ResearchResearch UCB PharmaUCB Pharma Acorda TherapeuticsAcorda Therapeutics Epilepsy Therapy Development ProjectEpilepsy Therapy Development Project Sage PharmaceuticalsSage Pharmaceuticals NeuroPace, Inc.NeuroPace, Inc. PfizerPfizer

Utility of EEG in Critically IllUtility of EEG in Critically Ill

Objectively measure severity of alteration in Objectively measure severity of alteration in consciousnessconsciousness

Assess neurologic function in patients who are Assess neurologic function in patients who are pharmacologically paralyzedpharmacologically paralyzed

Narrow the differential diagnostic possibilities when Narrow the differential diagnostic possibilities when combined with appropriate clinical informationcombined with appropriate clinical information

Determine if nonconvulsive seizures are cause of altered Determine if nonconvulsive seizures are cause of altered consciousness and assess response to treatmentconsciousness and assess response to treatment

Follow progression / improvement with serial studiesFollow progression / improvement with serial studies Provide prognostic informationProvide prognostic information Confirm the diagnosis of brain deathConfirm the diagnosis of brain death

Diffuse EtiologiesDiffuse Etiologies

Metabolic, toxic, infectious encephalopathiesMetabolic, toxic, infectious encephalopathies Grade or degree of abnormalities correlates fairly well Grade or degree of abnormalities correlates fairly well

with clinical statuswith clinical status EEG changes may precede or lag clinical changesEEG changes may precede or lag clinical changes Serial studies may be usefulSerial studies may be useful

Etiology often plays larger role than EEG patternEtiology often plays larger role than EEG pattern

Diffuse EtiologiesDiffuse Etiologies Slowing of posterior dominant rhythmSlowing of posterior dominant rhythm Diffuse thetaDiffuse theta Diffuse polymorphic theta and delta Diffuse polymorphic theta and delta

Loss of faster frequencies and sleep transientsLoss of faster frequencies and sleep transients Abnormal arousalsAbnormal arousals Intermittent rhythmic delta activityIntermittent rhythmic delta activity

Continuous diffuse high amplitude polymorphic deltaContinuous diffuse high amplitude polymorphic delta Continuous diffuse low voltage monomorphic deltaContinuous diffuse low voltage monomorphic delta Burst suppressionBurst suppression Low voltage (<20 Low voltage (<20 μμV) unreactive deltaV) unreactive delta Electrocerebral inactivityElectrocerebral inactivity

GOODGOOD

BADBAD

Classification SystemClassification SystemGr Synek Scollo-Lavizzari Young

I Regular alpha, some theta

Normal alpha Delta-theta > 50% of record

II Predominant theta Alpha, theta/delta Triphasic waves

III Widespread delta, spindle coma

Theta/delta, no alpha

Burst suppression

IV Burst-suppression, alpha coma, theta coma, delta coma

≤ 20 μV

Delta, low voltage; burst-suppression, PEDs, alpha coma

Alpha / theta / spindle coma, unreactive

V ECI ≤ 2 μV Very low to ECI Epileptiform activity

VI Suppression ≤ 10μV

Synek VM. J Clin Neurophysiol. 1988; 5: 161-74Scollo-Lavizzari G, et al. Eur Neurol. 1987; 26: 161-70 Young GB, et al. Can J Neurol Sci 1997;24:320-325

Synek: Prediction of Outcome Synek: Prediction of Outcome after Cardiac Arrestafter Cardiac Arrest

Good OutcomeGood Outcome Grade 1 Grade 1 48/61 48/61 79%79% Grade 2 Grade 2 45/88 45/88 51%51% Grade 3 Grade 3 11/43 11/43 26%26% Grade 4 Grade 4 0/138 0/138 0%0% Grade 5 Grade 5 0/70 0/70 0%0%

Severe Diffuse Slowing & Attenuation

Fp1-F3

F3-C3

C3-P3

P3-O1

Fp2-F4

F4-C4

C4-P4

P4-O2

Fp1-F7

F7-T 3

T 3-T 5

T 5-O1

Fp2-F8

F8-T 4

T 4-T 6

T 6-O2

Fz-Cz

Cz-Pz

LUC

R LC

E KG

Comment100 uV

1 sec

Burst Suppression: Barbiturate-Induced

Fp1-F3

F3-C3

C3-P3

P3-O1

Fp2-F4

F4-C4

C4-P4

P4-O2

Fp1-F7

F7-T 3

T 3-T 5

T 5-O1

Fp2-F8

F8-T 4

T 4-T 6

T 6-O2

Fz-Cz

Cz-Pz

E KG

Comment75 uV

5 sec

Electrocerebral Inactivity, ECI Montage

Fp1-C3

C3-O1

Fp1-T 3

T 3-O1

Fp2-C4

C4-O2

Fp2-T 4

F3-P3

Fz-Pz

F4-P4

F7-Fz

Fz-F8

T 3-Cz

Cz-T 4

T 5-Pz

Pz-T 6

LUC

R LC

E KG

Comment 30 uV

1 sec

2 uV/mm

Focal Structural EtiologiesFocal Structural Etiologies

Cause coma from herniation and compression/distortion Cause coma from herniation and compression/distortion of brainstem and diencephalonof brainstem and diencephalon

Focal asymmetriesFocal asymmetries Polymorphic delta activity: Subcortical white matterPolymorphic delta activity: Subcortical white matter Attenuation of faster frequencies: CortexAttenuation of faster frequencies: Cortex Intermittent rhythmic delta activity: Deep gray matter Intermittent rhythmic delta activity: Deep gray matter

structuresstructures May not be clear which hemisphere is more severely May not be clear which hemisphere is more severely

affected affected Slower frequencies and lower voltagesSlower frequencies and lower voltages

Other EtiologiesOther Etiologies

Brainstem lesionsBrainstem lesions Exception to relationship between EEG and clinical Exception to relationship between EEG and clinical

examexam Patient may be deeply comatosePatient may be deeply comatose Cortex (and therefore EEG) may be relatively Cortex (and therefore EEG) may be relatively

unaffectedunaffected Locked-in syndromeLocked-in syndrome

Psychogenic coma, catatoniaPsychogenic coma, catatonia EEG is normalEEG is normal

ReactivityReactivity

EEG change in response to sensory stimulationEEG change in response to sensory stimulation AuditoryAuditory VisualVisual SomatosensorySomatosensory

Light comaLight coma Generalized high voltage delta burstsGeneralized high voltage delta bursts

Deeper comaDeeper coma Diffuse attenuationDiffuse attenuation

Deep comaDeep coma No reactivity; poorer prognosisNo reactivity; poorer prognosis

Specific Coma Specific Coma PatternsPatterns

Alpha ComaAlpha Coma

Diffuse alpha frequency activity, 8-13HzDiffuse alpha frequency activity, 8-13Hz Often frontally dominantOften frontally dominant InvariantInvariant UnreactiveUnreactive

Transient pattern, evolves to other patternsTransient pattern, evolves to other patterns EtiologyEtiology

AnoxiaAnoxia Brainstem strokesBrainstem strokes Traumatic brain injuryTraumatic brain injury Drug intoxication (benzodiazepines, tricyclic Drug intoxication (benzodiazepines, tricyclic

antidepressants)antidepressants)

Alpha Coma and PrognosisAlpha Coma and Prognosis

Meta-analysis, 335 casesMeta-analysis, 335 cases Etiology predicts outcomeEtiology predicts outcome

Etiology Mortality

Anoxia 88%

Brainstem infarct 90%

Drug intoxication 8%

Kaplan PW, et al. Clin Neurophysiol. 1999; 110: 205-13

Beta ComaBeta Coma

High amplitude (>30 High amplitude (>30 μμV) diffuse 12-16 Hz activityV) diffuse 12-16 Hz activity Often frontally maximalOften frontally maximal UnreactiveUnreactive

EtiologiesEtiologies Drug intoxicationDrug intoxication AnesthesiaAnesthesia

Prognosis usually determined by etiology rather than Prognosis usually determined by etiology rather than EEG patternEEG pattern

Diffuse Beta, Barbiturate-Induced

Fp1-F3

F3-C3

C3-P3

P3-O1

Fp2-F4

F4-C4

C4-P4

P4-O2

Fp1-F7

F7-T 3

T 3-T 5

T 5-O1

Fp2-F8

F8-T 4

T 4-T 6

T 6-O2

Fz-Cz

Cz-Pz

LUC

R LC

E KG

Comment 100 uV1 sec

Spindle ComaSpindle Coma

Diffuse exaggerated 12-14 Hz sleep spindlesDiffuse exaggerated 12-14 Hz sleep spindles Resembles stage 2 or 3 (N2 or N3) sleepResembles stage 2 or 3 (N2 or N3) sleep

May show some stage changes (vertex waves, K May show some stage changes (vertex waves, K complexes)complexes)

No REMNo REM Little or no reactivity to external stimuliLittle or no reactivity to external stimuli

EtiologiesEtiologies Traumatic brain injuryTraumatic brain injury AnoxiaAnoxia Brainstem lesionsBrainstem lesions Drug intoxicationDrug intoxication

Spindle Coma and PrognosisSpindle Coma and Prognosis

Etiology plays a roleEtiology plays a role Overall better prognosis than alpha coma, mortality 23%Overall better prognosis than alpha coma, mortality 23% Presence of normal sleep tranisents suggests that cortex Presence of normal sleep tranisents suggests that cortex

and diencephalon are more intactand diencephalon are more intact

Kaplan PW, et al. Clin Neurophysiol. 2000; 111: 584-90

Etiology Mortality

Structural / brainstem 73%

Hypoxia 33%

Trauma 15%

Drug intoxication 0%

Subarachnoid HemorrhageSubarachnoid Hemorrhage

116 / 756 SAH patients with CEEG and 3 mo mRS116 / 756 SAH patients with CEEG and 3 mo mRS 88% poor grade SAH (Hunt & Hess ≥ 3)88% poor grade SAH (Hunt & Hess ≥ 3) Overall 3 month outcomeOverall 3 month outcome

69% moderate-severely disabled or dead69% moderate-severely disabled or dead 34% dead34% dead

Multivariate analysisMultivariate analysis Poor admission Hunt & Hess grade (OR 7.0)Poor admission Hunt & Hess grade (OR 7.0) Older age (OR 1.0 per year > 65)Older age (OR 1.0 per year > 65) Intraventricular hemorrhage (OR 2.6)Intraventricular hemorrhage (OR 2.6) No effect of delayed cerebral ischemiaNo effect of delayed cerebral ischemia

Claassen J et al. Neurocrit Care 2006;4:103-112Claassen J et al. Neurocrit Care 2006;4:103-112Dennis LJ et al.. Neurosurgery 2002;51:1136-1143Dennis LJ et al.. Neurosurgery 2002;51:1136-1143

CEEG Risk Factors in SAHCEEG Risk Factors in SAH

Claassen J et al. Neurocrit Care 2006;4:103-112Claassen J et al. Neurocrit Care 2006;4:103-112Dennis LJ et al.. Neurosurgery 2002;51:1136-1143Dennis LJ et al.. Neurosurgery 2002;51:1136-1143

EEG Finding Poor outcome

With RF (%)

Poor outcome Without RF (%)

OR 95% CI

Lateralized periodic discharges 91 66 18.8 1.6 - 214.6

Any periodic discharges 90 63 9.0 1.7 - 49.0

Absent sleep, 1st 24 hrs 74 29 10.4 1.4 - 78.1

Absent sleep, entire EEG 89 47 4.3 1.1 - 17.2

Absent reactivity, n = 9 * 100 0 - -

NCSE within 24 hrs, n = 4 * 100 0 - -

NCSE, entire EEG, n = 12 92 8 - -

GPEDs or BiPLEDs, n = 17 * 100 0 - -

* = Specificity and PPV for poor outcome = 100

Intracerebral HemorrhageIntracerebral Hemorrhage

Predictors of poor outcomePredictors of poor outcome Generalized periodic dischargesGeneralized periodic discharges Lateralized periodic dischargesLateralized periodic discharges Stimulus-induced rhythmic, periodic, or ictal Stimulus-induced rhythmic, periodic, or ictal

discharges (SIRPIDs)discharges (SIRPIDs)

Claassen J et al. Neurology 2007;69:1356-1365Claassen J et al. Neurology 2007;69:1356-1365Vespa PM et al. Neurology 2003;60:1441-1446Vespa PM et al. Neurology 2003;60:1441-1446

Cardiac ArrestCardiac Arrest

Therapeutic hypothermiaTherapeutic hypothermia 4 randomized clinical trials4 randomized clinical trials Comatose patients within 6 hrs of arrestComatose patients within 6 hrs of arrest Ventricular fibrillation or pulseless ventricular Ventricular fibrillation or pulseless ventricular

tachycardiatachycardia Mild TH (32-34º C) for 24 hrsMild TH (32-34º C) for 24 hrs Decreased mortality by 20%Decreased mortality by 20% Decreased poor neurologic outcome by 27%Decreased poor neurologic outcome by 27%

Cheung KW et al. Can J Emergency Med. 2006;8:329-337


AAN Practice Parameter: Prediction of outcome in AAN Practice Parameter: Prediction of outcome in comatose survivors after cardiopulmonary resuscitationcomatose survivors after cardiopulmonary resuscitation

Clinical factor Timing Level

Absent pupillary response 3 days A

Absent corneal reflexes 3 days A

Absent motor responses 3 days A

Myoclonic status epilepticus 24 hrs B

Serum NSE > 33µg/L 1-3 days B

Bilateral absent cortical SSEP 3 days B

Wijdicks EF et a;. Neurology 2006;67:203-210Wijdicks EF et a;. Neurology 2006;67:203-210


AAN Practice Parameter: Prediction of outcome in AAN Practice Parameter: Prediction of outcome in comatose survivors after cardiopulmonary resuscitationcomatose survivors after cardiopulmonary resuscitation

EEG finding Timing Level

Generalized suppression ≤ 20µV Any C

Burst-suppression Any C

GPEDs on flat background Any C

Wijdicks EF et a;. Neurology 2006;67:203-210Wijdicks EF et a;. Neurology 2006;67:203-210

Predictors of Poor Outcome: No THPredictors of Poor Outcome: No TH

Myoclonic status epilepticusMyoclonic status epilepticus EEG usually shows burst suppression and/or GPDsEEG usually shows burst suppression and/or GPDs Rare (<5%) with good cognitive outcomeRare (<5%) with good cognitive outcome

Usually treated with high-dose cIV-AEDsUsually treated with high-dose cIV-AEDs Preserved brainstem reflexesPreserved brainstem reflexes Intact cortical SSEP responsesIntact cortical SSEP responses Reactive EEG backgroundReactive EEG background

Predictors of Poor Outcome: No THPredictors of Poor Outcome: No TH

Background EEGBackground EEG Burst-suppressionBurst-suppression DiscontinuityDiscontinuity Generalized voltage attenuation (< 20µV)Generalized voltage attenuation (< 20µV) Alpha / theta / spindle coma without reactivityAlpha / theta / spindle coma without reactivity

Lack of reactivityLack of reactivity Periodic dischargesPeriodic discharges

Generalized periodic discharges on attenuated Generalized periodic discharges on attenuated backgroundbackground

EEG after Cardiac Arrest: No EEG after Cardiac Arrest: No THTH

Sensitivity 94%Sensitivity 94% Specificity 63%Specificity 63%

4 patients with malignant recovered awareness4 patients with malignant recovered awareness

Benign Malignant

Delta / theta > 50% of recording, with or without reactivity

Triphasic waves

Burst-suppression, with or without epileptiform activity

Alpha / theta / spindle coma, without reactivity

Generalized suppressioin

Thenayan EA et al. J Crit Care 2010;25:300-304Thenayan EA et al. J Crit Care 2010;25:300-304

Predictors of Outcome after Predictors of Outcome after Cardiac Arrest: No THCardiac Arrest: No TH

Meta-analysis of 50 studiesMeta-analysis of 50 studies 2828 adult patients, comatose after cardiac arrest2828 adult patients, comatose after cardiac arrest Outcomes assessed by Cerebral Performance Category Outcomes assessed by Cerebral Performance Category

(CPC)(CPC) CPC 4-5 vs. 1-3CPC 4-5 vs. 1-3 CPC 3-5 vs. 1-2CPC 3-5 vs. 1-2 Variable timing: hospital discharge to 12 mosVariable timing: hospital discharge to 12 mos

Sandroni C, Cavallaro F, Callaway CW, et al. Predictors of poor Sandroni C, Cavallaro F, Callaway CW, et al. Predictors of poor neurological outcome in adult comatose survivors of cardiac arrest: A neurological outcome in adult comatose survivors of cardiac arrest: A systematic review and meta-analysis. Part 1: Patients not treated with systematic review and meta-analysis. Part 1: Patients not treated with therapeutic hypothermia. Resuscitation 2013.therapeutic hypothermia. Resuscitation 2013.


Finding Timing Sensitivity FPR 95% CI

Quality

Myoclonus, n=471 24-48 hrs 9 0 0-3 Low

Bilateral absent SSEP, n = 293

24-72 hrs 45-46% 0 0-9 Low

Absent pupillary response, n = 382

72 hrs 18 0 0-8 Low

NSE, S-100B Variable -- -- Very low



EEG Finding Timing Sensitivity FPR 95% CI

Quality

Grade III-V (Edgren), n=46

24 hrs 36 0 0-22 Very low

Grade IV-V (Synek), n=40

≤ 48 hrs 42 0 0-19 Very low

Grade IV-V (Bassetti), n=59

≤ 72 hrs 42 0 0-24 Very low

Low voltage EEG, ≤ 20 µV, n=355

24-72 hrs

28 0 0-6 Low

Alpha coma --


Predictors of Poor Outcome: THPredictors of Poor Outcome: TH

Background EEGBackground EEG Burst-suppressionBurst-suppression DiscontinuityDiscontinuity Generalized voltage attenuation (< 10µV)Generalized voltage attenuation (< 10µV) Lack of reactivity Lack of reactivity

Periodic dischargesPeriodic discharges Generalized periodic discharges on attenuated Generalized periodic discharges on attenuated

backgroundbackground

Predictors of Poor Outcome: THPredictors of Poor Outcome: TH

Prospective, 111 adult survivors of cardiac arrestProspective, 111 adult survivors of cardiac arrest Unreactive EEG background strong predictor of mortality Unreactive EEG background strong predictor of mortality

and poor long-term neurologic recovery (FP = 7%)and poor long-term neurologic recovery (FP = 7%) Motor response to pain (FP = 24%)Motor response to pain (FP = 24%) 2+ risk factors = specificity 1.0; PPV 1.02+ risk factors = specificity 1.0; PPV 1.0

Bilaterally absent cortical SSEPBilaterally absent cortical SSEP Unreactive EEGUnreactive EEG Early myoclonusEarly myoclonus Incomplete recovery of brainstem reflexesIncomplete recovery of brainstem reflexes

Rossetti AO et al. Ann Neurol 2010;67:301-307Rossetti AO et al. Ann Neurol 2010;67:301-307

Predictors of Outcome after Predictors of Outcome after Cardiac Arrest: THCardiac Arrest: TH

Meta-analysis of 37 studiesMeta-analysis of 37 studies 2403 adult patients, comatose after cardiac arrest2403 adult patients, comatose after cardiac arrest Outcomes assessed by Cerebral Performance Category Outcomes assessed by Cerebral Performance Category

(CPC)(CPC) CPC 4-5 vs. 1-3CPC 4-5 vs. 1-3 CPC 3-5 vs. 1-2CPC 3-5 vs. 1-2 Variable timing: hospital discharge to 12 mosVariable timing: hospital discharge to 12 mos

Sandroni C, Cavallaro F, Callaway CW, et al. Predictors of poor Sandroni C, Cavallaro F, Callaway CW, et al. Predictors of poor neurological outcome in adult comatose survivors of cardiac neurological outcome in adult comatose survivors of cardiac arrest: A systematic review and meta-analysis. Part 2: Patients arrest: A systematic review and meta-analysis. Part 2: Patients treated with therapeutic hypothermia. Resuscitation 2013.treated with therapeutic hypothermia. Resuscitation 2013.


Finding Timing Sensitivity FPR 95% CI

Quality

Bilateral absent SSEP

During TH

28 0 0-2 Moderate

Bilateral absent SSEP

After TH 42 0 0-4 Low

Absent pupillary + absent corneal + motor response ≤ extension, n = 103

72 hrs 15 0 0-8 Very low

NSE, S-100B Variable -- -- Very low



EEG Finding Timing Sensitivity FPR 95% CI

Quality

Burst-suppression During TH

37 0 0-5 Low

Burst-suppression After TH 18 0 0-5 Low

Status epilepticus from burst-suppression

Any time 42 0 0-5 Low

Nonreactive background

After TH 62 0 0-3 Low


ConfoundersConfounders

Sedating medicationsSedating medications Propofol, midazolam, pentobarbitalPropofol, midazolam, pentobarbital

Presence of multiple etiologies (e.g. post-arrest + hepatic Presence of multiple etiologies (e.g. post-arrest + hepatic or renal failure)or renal failure)

ArtifactArtifact Shivering / EMGShivering / EMG Electrode artifactElectrode artifact

ConclusionConclusion Use EEG for patients with altered mental statusUse EEG for patients with altered mental status

Objective measure for encephalopathyObjective measure for encephalopathy Narrow differential diagnosis when etiology unknownNarrow differential diagnosis when etiology unknown

Serial or continuous studies may be helpfulSerial or continuous studies may be helpful EEG can help with prognostication when etiology is knownEEG can help with prognostication when etiology is known

Better at predicting poor outcomeBetter at predicting poor outcome Early inaccurate prognostication may result in self-fulfilling Early inaccurate prognostication may result in self-fulfilling

prophecy: early withdrawal of careprophecy: early withdrawal of care Large prospective studies needed to determine prognostic Large prospective studies needed to determine prognostic

value of CEEG across multiple etiologies and severity of value of CEEG across multiple etiologies and severity of illnessillness

ICU EEG: Prognosis in...

Documents

Transcript of ICU EEG: Prognosis in...