G Model

RESUS 6125 1–6 ARTICLE IN PRESS

Resuscitation xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Resuscitation
journal homepage: www.elsevier.com/locate/resuscitation

1 Clinical Paper

2 Prognostic value of electroencephalography (EEG) after

3 out-of-hospital cardiac arrest in successfully resuscitated patients
4 used in daily clinical practice
5 Q1 Helle Søholm a , Troels Wesenberg Kjær b , Jesper Kjaergaard a , Tobias Cronberg c ,
6 John Bro-Jeppesen a , Freddy K. Lippert d , Lars Køber a , Michael Wanscher e ,
7 Christian Hassager a,∗
a
8 Department of Cardiology 2142, The Heart Center, Copenhagen University Hospital Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen, Denmark
b
9 Department of Neurophysiology, Copenhagen University Hospital Rigshospitalet, Denmark
c
10 Department of Clinical Sciences, Division of Neurology, Lund University, Lund, Sweden
d
11 Emergency Medical Services, The Capital Region of Denmark, Denmark
e
12 Department of Thoracic Anaesthesiology, The Heart Center, Copenhagen University Hospital Rigshospitalet, Denmark
13

14
31 a r t i c l e i n f o a b s t r a c t
15
16 Article history:
17 Received 1 July 2014 Background: Out-of-hospital cardiac arrest (OHCA) is associated with a poor prognosis and predicting
18 Received in revised form 11 August 2014 outcome is complex with neurophysiological testing and repeated clinical neurological examinations
19 Accepted 22 August 2014
as key components of the assessment. In this study we examine the association between different
20 electroencephalography (EEG) patterns and mortality in a clinical cohort of OHCA-patients.
21 Keywords:
Methods: From 2002 to 2011 consecutive patients were admitted to an intensive-care-unit after resus-
22 Cardiac arrest
citation from OHCA. Utstein-criteria for pre-hospital data and review of individual patients’ charts for
23 Out-of-hospital cardiac arrest
24 Electroencephalography
post-resuscitation care were used. EEG reports were analysed according to the 2012 American Clinical
25 EEG Neurophysiology Society’s guidelines.
26 Survival Results: A total of 1076 patients were included, and EEG was performed in 20% (n = 219) with a median of
27 Prognosis 3(IQR 2–4) days after OHCA. Rhythmic Delta Activity (RDA) was found in 71 patients (36%) and Periodic
28 Prognostication Discharges (PD) in 100 patients (45%). Background EEG frequency of Alpha+ or Theta was noted in 107
29 Outcome patients (49%), and change in cerebral EEG activity to stimulation (reactivity) was found in 38 patients
30 Q3 Neurological outcome
(17%). Suppression (all activity <10 ␮V) was found in 26 (12%) and burst-suppression in 17 (8%) patients.
A favourable EEG pattern (reactivity, favourable background frequency and RDA) was independently
associated with reduced mortality with hazard ratio (HR) 0.43 (95%CI: 0.24–0.76), p = 0.004 (false positive
rate: 31%) and a non-favourable EEG pattern (no reactivity, unfavourable background frequency, and PD,
suppressed voltage or burst-suppression) was associated with higher mortality (HR = 1.62(1.09–2.41),
p = 0.02) after adjustment for known prognostic factors (false positive rate: 9%).
Conclusion: EEG may be useful in work-up in prognostication of patients with OHCA. Findings such as
Rhythmic Delta Activity (RDA) seem to be associated with a better prognosis, whereas suppressed voltage
and burst-suppression patterns were associated with poor prognosis.
© 2014 Published by Elsevier Ireland Ltd.

32 1. Introduction after resuscitation, while prediction of outcome in patients with 36

impaired consciousness after successful resuscitation is complex. 37

33 Out-of-hospital cardiac arrest (OHCA) is associated with a After the implementation of targeted temperature management 38

34 poor prognosis even though survival has increased in recent prognostication has become more challenging as sedatives, muscle 39

35 years.1 The best predictor of a favourable outcome is awakening relaxants and cooling in itself seem to make traditional pro- 40

gnostic indicators less reliable.2–5 C:\Users\franklandj\AppData\ 41

Local\Microsoft\Windows\Temporary Internet Files\Content.IE5 42

∗ Corresponding author.
\QE9QNGK1\Hypothermia – Current guidelines recommend that 43
Q2
E-mail address: helle.soholm@gmail.com (C. Hassager).
prognostication is not performed until 72 h after rewarming.6–9 44

http://dx.doi.org/10.1016/j.resuscitation.2014.08.031
0300-9572/© 2014 Published by Elsevier Ireland Ltd.

Please cite this article in press as: Søholm H, et al. Prognostic value of electroencephalography (EEG) after out-
of-hospital cardiac arrest in successfully resuscitated patients used in daily clinical practice. Resuscitation (2014),
http://dx.doi.org/10.1016/j.resuscitation.2014.08.031
 G Model
RESUS 6125 1–6 ARTICLE IN PRESS
2 H. Søholm et al. / Resuscitation xxx (2014) xxx–xxx

45 Multimodal prognostication in comatose OHCA-patients is rec- 2.3. Electroencephalograms 104

46 ommended with use of a combination of repeated neurological

47 examination by experienced neurologists, use of neuroimaging Electroencephalograms were recorded with electrode-caps 105

48 (cerebral computed tomography (CT) or magnetic resonance imag- using silver/silver-chloride electrodes placed according to the 106

49 ing (MR)), biomarkers such as neuron-specific enolase (NSE), international 10/20 system At least nineteen digitized channels 107

50 and neurophysiological testing (somatosensory evoked potentials were acquired with a band pass filter setting of at least 0.53–35 Hz, 108

51 (SSEP) and electroencephalography (EEG)).10–17 Bilateral absence but typical 0.1–70 Hz. Most data were recorded with Nicolet One 109

52 of N20 responses at SSEP has been suggested as a reliable predictor (Natus Medical Incorporated, San Carlos, CA), some with Cadwell 110

53 of poor prognosis in non-cooled patients, and to some extend in Easy II (Cadwell Laboratories, Inc., Kennewick, WA) or Stellate Har- 111

54 patients who have been rewarmed after target temperature man- monie (Stellate Systems, Inc., Montreal, Quebec, Canada). 112

55 agement. However, this modality is not available everywhere.18,19 The EEG was examined by reviewing of clinical response reports 113

56 EEGs in comatose patients is recognized as useful, however the written by a certified neurophysiologist (attending physician) at 114

57 EEG-pattern may be affected by sedatives as well as hypothermia, time of examination. The EEG-reports were reassessed according 115

58 and the different patterns indicating prognosis after cardiac arrest to selected rhythms and periodic patterns in the ACNS terminol- 116

59 have not been thoroughly investigated.20 Updated guidelines for ogy in critical care patients from 2012 blinded to clinical outcome 117

60 EEG reporting and terminology in critical care patients was recently and supervised by a professor in neurophysiology.21 The EEG was 118

61 published from the American Clinical Neurophysiology Society categorized according to ‘Main Term 2 with presence of rhythmic 119

62 (ACNS) and the validity of the new guidelines with regards to fea- or periodic patterns (Rhythmic Delta Activity (RDA) and Peri- 120

63 sibility and assessment of post-cardiac-arrest prognosis remains to odic Discharges (PD)) and ‘Background EEG’ with predominant 121

64 be established.21 background EEG frequency (Alpha+ (≥8 Hz), Theta (4–7.9 Hz) or 122

65 This study investigates the association between adverse Delta (<4 Hz)), ‘Reactivity’ (change in EEG activity to stimulation), 123

66 outcome and different electroencephalography (EEG) patterns ‘Voltage’ (suppressed, all activity <10 ␮V), and ‘Continuity’ (Burst- 124

67 assessed according to the updated ACNS guidelines in a clinical suppression).21 125

68 cohort of out-of-hospital cardiac arrest patients.

2.4. Outcome 126

69 2. Materials and methods

All-cause in-hospital and 30-day mortality was the primary end- 127

70 2.1. Patients and study area point of this study. Mortality data were acquired from the Danish 128

National Patient Registry, which holds vital status on all Danish citi- 129

71 Patients with OHCA were consecutively included in the study zens, all of whom are registered by a unique personal identification 130

72 from June 2002 through 2011.22 Patients with OHCA of all aeti- number. Outcome data were available with 99.5% completeness 131

73 ologies resuscitated by the Emergency Medical Services (EMS) and (missing in n = 5) due to invalid ID number. Neurological outcome 132

74 admitted to an intensive care unit (ICU) in hospitals in the greater was assessed with the Cerebral Performance Category score (CPC) 133

75 Copenhagen area were included. Patients younger than 18 years at hospital discharge and favourable neurological outcome was 134

76 and non-Danish residents were excluded since outcome data were defined as CPC 1 or 2, non-favourable as 3 or 4 and dead as CPC-score 135

77 not readily available for these patients. 5.25 136

78 The EMS in Copenhagen, the capital of Denmark (area of 675 km2

79 (260 Mi2 ), 1.2 million people), is staffed with a paramedic and
80 an attending physician. The EMS is dispatched to all patients 2.5. Statistics 137

81 with assumed OHCA with the treatment protocol adherent to the

82 advanced life support guidelines from the European Resuscitation Continuous variables are presented as mean ± standard devia- 138

83 Council.23,24 The attending physician uses an Utstein registration tion for normally distributed data and as median and interquartile 139

84 sheet as documentation and enters the data in the OHCA-database ranges (IQR) for non-normal distributed data. Differences were ana- 140

85 immediately after each dispatch.25–27 lysed with Student’s unpaired t-test or Wilcoxon rank sum test as 141

appropriate. Categorical variables are presented as number (n) and 142

per cent and differences are analysed with chi2 -test. A univariate 143

86 2.2. Post-resuscitation care logistic regression analysis was performed estimating the odds of 144

having an EEG performed in patients with myoclonic jerks and gen- 145

87 Patients were admitted for post-resuscitation care at an ICU at eralized seizures. Mortality is presented as Kaplan–Meier plots with 146

88 the nearest of eight hospitals in the greater Copenhagen area. If time 0 as time of admission to the hospital, and differences are 147

89 the diagnosis of ST-segment elevation myocardial infarction was tested using log rank test. Univariate and multiple COX-regression 148

90 suspected in the pre-hospital setting the patient was admitted at analysis for 30-day mortality were performed in patients examined 149

91 one of two tertiary heart centres for primary percutaneous coronary with an EEG estimating hazard ratios (HR) for each EEG pattern and 150

92 intervention. A single investigator reviewed all individual patient 95% confidence intervals (CI) adjusting for potential confounders 151

93 charts with focus on in-hospital post-resuscitation care. Pre-arrest (age, gender, rhythm, witnessed arrest, bystander CPR, time to 152

94 co-morbidity was estimated according to the Charlson Comorbidity ROSC, therapeutic hypothermia, cardiac aetiology, public arrest, 153

95 Index, which takes the severity of 22 conditions into account and is acute coronary angiography and calendar year) after checking for 154

96 commonly used to predict short-term mortality.28,29 Termination the underlying assumptions of proportionality and lack of inter- 155

97 of therapy was classified as due to circulatory failure, anoxic brain actions. The use of EEG has become more common during the 156

98 damage, living will/do not resuscitate wish, long time to ROSC or study period and the examination is therefore found to be time 157

99 due to extensive pre-existing comorbidity, and was determined in dependent. False-positive rates of in-hospital mortality and non- 158

100 the data collection according to the chart notes by the attending favourable neurological outcome at discharge were calculated for 159

101 physician. each EEG-pattern with 95% CI. All statistical analysis was carried 160

102 The regional ethics committee waived informed consent to the out in SAS Statistics version 9.3 (Cary, NC, USA) with a level of 161

103 study with the reference number: H-2-2012-053. significance defined as p < 0.05. 162

Please cite this article in press as: Søholm H, et al. Prognostic value of electroencephalography (EEG) after out-
of-hospital cardiac arrest in successfully resuscitated patients used in daily clinical practice. Resuscitation (2014),
http://dx.doi.org/10.1016/j.resuscitation.2014.08.031
 G Model
RESUS 6125 1–6 ARTICLE IN PRESS
H. Søholm et al. / Resuscitation xxx (2014) xxx–xxx 3

Table 1 Table 2
Demographic and clinical data in patients with out-of-hospital cardiac arrest admit- Electroencephalographic (EEG) patterns based on 2012 American Clinical Neuro-
ted to the hospital with successful resuscitation or ongoing CPR at arrival. physiology Society’s terminology.

Admitted ICU EEG performed EEG performed

n = 1076 n = 219 (20%) n = 219

Age, yrs. 63 ± 14 61 ± 12 Main term 2

Men 62 ± 14 61 ± 12 Rhythmic Delta Activity (RDA) 71 (36%)
Women 67 ± 15 62 ± 12 Periodic Discharges (PD) 100 (50%)
Men, n 780 (73%) 155 (71%) Background EEG frequency
Public arrest, n 411 (39%) 97 (45%) Alpha+ or Theta 107 (49%)
Witnessed cardiac arrest, n 875 (85%) 173 (82%) Delta 36 (16%)
Bystander CPR, n 541 (54%) 105 (51%) Other/unclassified 76 (35%)
On-going CPR, n 51 (5%) 14 (6%) Reactivity (change in EEG activity to stimulation)
Alarm to EMS arrival, min. 6 (5–9) 7 (5–10) Reactivity 38 (17%)
Arrest after arrival of EMS, n 38 (5%) 3 (2%) No reactivity 155 (71%)
Primary rhythm, n Other/unclassified 26 (12%)
VF/pVT 585 (54%) 120 (55%) Voltage
Asystole 234 (22%) 56 (26%) Suppressed (all activity <10 ␮V) 26 (12%)
PEA 184 (17%) 37 (17%) Continuity
Other 73 (7%) 6 (3%) Burst-suppression 17 (9%)
Alarm to DC, min. 5 (2–9) 6 (2–9)
Time to ROSC, min. 15 (9–22) 18 (12–24)
Admission to tertiary centre, n 702 (65%) 155 (71%)
Presumed cardiac aetiology, n 858 (80%) 179 (82%) Suppressed voltage (all activity <10 ␮V) was noted in 26 (12%), and 189

Acute coronary syndrome, n 465 (43%) 84 (39%) continuity with burst-suppression in 17 (9%), Table 2. 190
TH, n 665 (63%) 166 (78%)
TH (implemented and patient 665 (74%) 166 (85%)
not awake), n 3.3. Outcome 191
Charlson co-morbidity index, n
-0 435 (41%) 94 (43%) A total of 466 (43%) of patients admitted to the ICU where 192
1 279 (26%) 68 (31%)
discharged alive, however only 42 (19%) of patients with an EEG 193
2 166 (16%) 26 (12%)
≥3 185 (17%) 31 (14%) performed were discharged alive from hospital, p < 0.001. Patients 194

examined with EEG and not surviving to hospital discharge (n = 177, 195
Data are presented as mean ± standard deviation, number (percentage) or median
(interquartile range). 81%) died a median of 2 days (1–5) after the EEG was performed. 196

Abbreviations: CPR: cardiopulmonary resuscitation, EMS: emergency medical The primary reason for withdrawal of active treatment was pre- 197
service, VF: ventricular fibrillation, pVT: pulse less ventricular tachycardia, PEA: sumed poor neurological prognosis due to anoxic brain damage 198
pulse less electrical activity, DC: defibrillation (direct current), ROSC: return of (n = 161, 96%), whereas co-morbidity, organ failure and circulatory 199
spontaneous circulation, TH: therapeutic hypothermia.
failure were only noted in 1% each. The time from admission to 200

withdrawal of active treatment was significantly longer for patients 201

treated with TH (6 days (IQR: 4–10) vs. 5 (3–8), p = 0.04) compared 202
163 3. Results
to patients without cooling applied. 203

Fig. 1 shows Kaplan–Meier 30-day mortality plots according 204

164 3.1. Patient characteristics
RDA vs. PD, suppressed voltage and burst-suppression, background 205

frequency and reactivity. Reactivity, background frequency (Alpha+ 206

165 A total of 1076 patients suffered from OHCA and were admit-
or Theta) and RDA were associated with significantly lower mortal- 207
166 ted to the ICU. In 219 patients (20%) an EEG was performed in
ity rates compared to no reactivity, background frequency (Delta) 208
167 a median of 3 (IQR: 2–4, min. 0 max. 9 days) days after OCHA
and other rhythms than RDA, Fig. 1. All patients with suppressed 209
168 with significantly more EEGs performed in recent years with 77%
voltage (n = 26, 12%) or burst-suppression (n = 17, 9%) died during 210
169 (n = 169) of the EEGs performed after 2006 (p = 0.02). Demographic,
the hospital stay. 211
170 pre-hospital and data on post-resuscitation care are shown in
By multivariable models adjusting for known prognostic factors, 212
171 Table 1. Myoclonic jerks were seen in 62% (n = 128) vs. 13% (n = 107)
RDA and background frequency (Alpha+ or Theta) were inde- 213
172 (p < 0.001) and generalized seizures in 39% (n = 80) vs. 9% (n = 69)
pendently associated with lower 30-day mortality, whereas PD, 214
173 (p < 0.001) of patients with and without an EEG performed. The
suppressed voltage and burst-suppression were found indepen- 215
174 odds of having an EEG performed was 14.9 (9.9–19.5), p < 0.001
dently associated with higher mortality rates, Table 3. A favourable 216
175 in patients with myoclonic jerks and 8.4 (5.8–12.0), p < 0.001 in
EEG-pattern (reactivity, background frequency (Alpha+ or Theta) 217
176 patients with generalized seizures.
and RDA) was independently associated with lower mortality, 218

while a non-favourable EEG included no reactivity, background 219

177 3.2. EEG patterns frequency (Delta or 0) and PD, suppressed voltage or burst- 220

suppression. The false positive rates of prediction in-hospital 221

178 Targeted temperature management was implemented during mortality and poor neurological outcome for each EEG-pattern are 222

179 the study period in the different admission hospitals from 2004 to given in Table 4. Burst-suppression and suppressed voltage was the 223

180 2006 and the EEGs were performed significantly later after imple- only variables with false positive rates of 0, but a non-favourable 224

181 mentation of therapeutic hypothermia 3 (IQR: 2–4) days vs. 2 (1–4) EEG was associated with a low false positive rate of 9% and PD with 225

182 days (p = 0.01) compared to the period before implementation. a false positive rate of 13%, Table 4. 226

183 Table 2 shows the EEG-patterns based on the ACNS terminol- Patients examined with EEG were discharged alive from hos- 227

184 ogy. According to the main term 2, Rhythmic Delta Activity (RDA) pital (n = 42, 19%) with a favourable neurological outcome in 43% 228

185 was seen in 71 patients (36%) and Periodic Discharges (PD) in 100 of cases (n = 18) and a non-favourable (CPC 3 or 4) in 57% (n = 24). 229

186 patients (50%). Background frequency (Alpha+ or Theta), perceived EEG showing suppressed voltage and burst-suppression were asso- 230

187 as more favourable, was noted in 107 patients (49%) and change in ciated with death in all cases whereas the remaining rhythms were 231

188 EEG activity to stimulation (reactivity) was seen 38 patients (17%). found with higher false positive rates, Table 4. 232

Please cite this article in press as: Søholm H, et al. Prognostic value of electroencephalography (EEG) after out-
of-hospital cardiac arrest in successfully resuscitated patients used in daily clinical practice. Resuscitation (2014),
http://dx.doi.org/10.1016/j.resuscitation.2014.08.031
G Model
RESUS 6125 1–6 ARTICLE IN PRESS
4 H. Søholm et al. / Resuscitation xxx (2014) xxx–xxx

Fig. 1. Kaplan–Meier 30-day mortality for out-of-hospital cardiac arrest patients examined with electroencephalography (EEG). (A) RDA vs. PD, suppressed voltage and
burst-suppression. (B) Background frequency. (C) Change in cerebral EEG activity to stimulation (reactivity). Abbreviations: RDA: Rhythmic Delta Activity, PD: Periodic
Discharges.

Please cite this article in press as: Søholm H, et al. Prognostic value of electroencephalography (EEG) after out-
of-hospital cardiac arrest in successfully resuscitated patients used in daily clinical practice. Resuscitation (2014),
http://dx.doi.org/10.1016/j.resuscitation.2014.08.031
 G Model
RESUS 6125 1–6 ARTICLE IN PRESS
H. Søholm et al. / Resuscitation xxx (2014) xxx–xxx 5

Table 3
Univariate and multivariable models of 30-day mortality for successfully resuscitated patients (ROSC or ongoing CPR at hospital arrival) suffering from out-of-hospital cardiac
arrest admitted to the intensive care unit and examined with an electroencephalography (EEG). One EEG pattern was included in each model.

HR (95% CI) p-Value Adjusted HR (95% CI)a p-Value

Main term
RDA 0.55 (0.39–0.77) <0.001 0.53 (0.33–0.86) 0.01
PD 1.26 (0.92–1.72) 0.2 1.66 (1.10–2.51) 0.02
Background EEG frequency (Alpha+ or Theta) 0.64 (0.47–0.89) 0.008 0.64 (0.42–0.96) 0.03
Reactivity (yes) 0.63 (0.40–0.97) 0.03 0.76 (0.43–1.35) 0.3
Voltage
Suppressed (all activity <10 ␮V) 2.77 (1.79–4.28) <0.001 1.76 (1.02–3.02) 0.04
Continuity
Burst-suppression 2.84 (1.70–4.73) <0.001 2.62 (1.31–5.27) 0.007
EEG favourableb (n = 48) 0.59 (0.40–0.87) 0.007 0.43 (0.24–0.76) 0.004
EEG non-favourablec (n = 76) 1.81 (1.33–2.47) <0.001 1.62 (1.09–2.41) 0.02
a
Adjusted for: age, gender, primary rhythm, therapeutic hypothermia, cardiac aetiology, witnessed arrest, bystander CPR, acute coronary angiography, time to ROSC,
public arrest and calendar year.
b
EEG favourable indicates: RDA, background frequency: Alpha+ or Theta and reactivity.
c
EEG non-favourable indicates: PD, suppressed (all activity <10 ␮V) or burst-suppression, no reactivity and background frequency: Delta.
Abbreviations: ROSC: Return of Spontaneous Circulation, CPR: Cardiopulmonary Resuscitation, HR: Hazard Ratio, RDA: Rhythmic Delta Activity, PD: Periodic Discharges, EEG:
Electroencephalography.

Table 4 recovery of brainstem reflexes and bilateral absence of early cortical 252
False positive rates of relation of electroencephalographic patterns and in-hospital
responses (N20) at SSEP at 72 h as predictors with very low false- 253
mortality and non-favourable neurological outcome in survivors (CPC-score 3 or
4) at hospital discharge, respectively, after successful resuscitation from out-of- positive rates for poor outcome.17,19 Prognostication has become 254

hospital cardiac arrest. more challenging after the implementation of targeted tempera- 255

ture management as it may affect the prognostic parameters due 256

False positive rates False positive rates (95%
(95% CI)In-hospital CI)Non-favourable
to residual effects of sedation and cooling.2,32 257

mortality neurological outcome The best predictor of a good neurological outcome is early awak- 258

ening, but the time to awakening even with good neurological 259
Main term
RDA 0.28 (0.18–0.40) 0.45 (0.26–0.66) outcome is highly variable.33,34 Current guidelines recommend a 260

PD 0.13 (0.07–0.22) 0.31 (0.13–0.58) multimodal approach with prognostication not performed until 261
Background frequency 72 h after rewarming including repeated neurological examination, 262
(Alpha+ or Theta) 0.23 (0.16–0.33) 0.36 (0.20–0.55)
neuroimaging such as cerebral CT and MR, biomarkers (NSE), and 263
Reactivity (yes) 0.34 (0.21–0.5 0) 0.62 (0.37–0.82)
Voltage neurophysiological testing (EEG and SSEP).6–19 In this study the EEG 264

Suppressed (all 0.00 (0.00–0.16) (all dead) was performed significantly later after the implementation of TH in 265
activity <10 ␮V) a median of three days. More importantly the time from admission 266
Continuity to withdrawal of active treatment was likewise significantly longer 267
Burst-suppression 0.00 (0.00–0.23) (all dead)
for patients treated with TH (median 6 days) compared to patients 268
EEG favourablea 0.31 (0.19–0.46) 0.47 (0.25–0.70)
EEG non-favourableb 0.09 (0.04–0.19) 0.43 (0.16–0.80) without cooling applied (median 5 days). 269

a
The use of EEG in comatose patients has increased in recent 270
EEG favourable indicates: reactivity, background frequency: Alpha+ or Theta and
RDA. years. The EEG-pattern may be affected by sedatives as well as 271
b
EEG non-favourable indicates: no reactivity, background frequency: Delta or 0, cooling, and the different EEG-patterns that indicate prognosis 272
and PD or Suppressed (all activity <10 ␮V) or burst-suppression. after OHCA are not thoroughly investigated according to the newly 273
Abbreviations: CPC: Cerebral Performance Category, RDA: Rhythmic Delta Activity, published guidelines on critical care EEG terminology.21 In this clin- 274
PD: Periodic Discharges, EEG: Electroencephalography.
ical observational study with use of EEG in a daily clinical setting 275

of OHCA-patients burst-suppression and suppressed voltage were 276

associated with false positive rate of 0%, whereas a non-favourable 277

233 4. Discussion EEG was associated with false positive rate of 9% and PD with 13%, 278

however no other EEG pattern was strongly associated with out- 279

234 EEG is performed in 20% of patients in our consecutive clin- come (Table 4). Previous studies have shown that a non-reactive 280

235 ical cohort of patients resuscitated from OHCA and admitted to EEG, no background frequency and burst-suppression pattern are 281

236 the ICU. Based on the 2012 guidelines from the American Clinical associated with a poor prognosis, which corresponds well with our 282

237 Neurophysiology Society the EEG-patterns ‘reactivity’, ‘background study.4,15,35 283

238 frequency’ (Alpha+ or Theta) and the ‘RDA’ were associated with Neurophysiological tests, including the EEG examinations in this 284

239 lower 30-day mortality whereas ‘PD’, ‘burst-suppression’ and study, are performed as prognostic tools as a support for clinical 285

240 ‘suppressed voltage were associated with a poor prognosis.21 Forty- decision-making. In this study on the prognostic value of the EEG 286

241 three percent of patients admitted to the ICU were discharged alive, and in other retrospective studies not all patients are therefore 287

242 whereas 19% of patients examined with an EEG survived to dis- examined. We found that the odds of having an EEG-performed 288

243 charge and of those were 43% discharged with a good neurological was 14.9 for patients with myoclonic jerks and 8.4 for patients with 289

244 outcome. generalized seizures, which underlines the use of the EEG in daily 290

245 Predicting which patients will achieve a good neurological clinical practice in this study. The prognostication should according 291

246 outcome especially in the comatose is challenging.30 The Levy to guidelines not be performed until 72 h after rewarming from TH 292

247 Criteria, published in 1985, with absent pupillary reflexes or or 72 h from OHCA in normothermic patients. Patients with early 293

248 absent/extensor motor response at 72 h after OHCA was previously awakening are commonly not examined with an EEG. The pre- 294

249 used as predictors of a poor prognosis.31 The American Academy of sumed poor neurological prognosis based on neurophysiological 295

250 Neurology published in 2006 updated guidelines with myoclonus, testing may lead to withdrawal of life supporting therapies, which 296

251 extending posturing, no motor response to pain, incomplete leads to subsequent death and it may therefore represent a 297

Please cite this article in press as: Søholm H, et al. Prognostic value of electroencephalography (EEG) after out-
of-hospital cardiac arrest in successfully resuscitated patients used in daily clinical practice. Resuscitation (2014),
http://dx.doi.org/10.1016/j.resuscitation.2014.08.031
 G Model
RESUS 6125 1–6 ARTICLE IN PRESS
6 H. Søholm et al. / Resuscitation xxx (2014) xxx–xxx

298 self-fulfilling prophecy.7 In this study, withdrawal of active ther- 11. Cronberg T, Rundgren M, Westhall E, et al. Neuron-specific enolase corre- 362

299 apy was performed due to assumed anoxic brain damage in 96% of lates with other prognostic markers after cardiac arrest. Neurology 2011;77: 363
623–30. 364
300 patients. 12. Greer DM, Yang J, Scripko PD, et al. Clinical examination for prognostication in 365
301 This study is limited in its retrospective design with a potential comatose cardiac arrest patients. Resuscitation 2013;84:1546–51. 366

302 selection bias as not all OHCA-patients were examined with an EEG 13. Hahn DK, Geocadin RG, Greer DM. Quality of evidence in studies evaluating 367
neuroimaging for neurologic prognostication in adult patients resuscitated from 368
303 and as all EEGs were ordered as part of clinical decision-making. No cardiac arrest. Resuscitation 2014;85:165–72. 369
304 data on sedation was available, which could have been relevant as 14. Kamps MJ, Horn J, Oddo M, et al. Prognostication of neurologic outcome in car- 370
305 sedating drugs such as propofol may alter the EEG pattern. How- diac arrest patients after mild therapeutic hypothermia: a meta-analysis of the 371
current literature. Intensive Care Med 2013;39:1671–82. 372
306 ever, the EEG was performed in a median of 3 days after OHCA and 15. Oddo M, Rossetti AO. Predicting neurological outcome after cardiac arrest. Curr 373
307 as the anaesthetic effect on the EEG pattern is normally sufficiently Opin Crit Care 2011;17:254–9. 374
308 reduced after 2–4 days the impact on the EEG is likely small. In 16. Sandroni C, Cavallaro F, Callaway CW, et al. Predictors of poor neurological out- 375
come in adult comatose survivors of cardiac arrest: a systematic review and 376
309 addition in the multivariate models we adjusted for treatment with
meta-analysis. Part 2: patients treated with therapeutic hypothermia. Resusci- 377
310 therapeutic hypothermia, and during this treatment patients were tation 2013;84:1324–38. 378
311 sedated. Some immortal time bias is present in the current study 17. Sandroni C, Cavallaro F, Callaway CW, et al. Predictors of poor neurological 379
outcome in adult comatose survivors of cardiac arrest: a systematic review 380
312 as patients are examined with an EEG from 0 to 9 days (median 3)
and meta-analysis. Part 1: patients not treated with therapeutic hypothermia. 381
313 after OHCA. Physicians may have withdrawn intensive care partly Resuscitation 2013;84:1310–23. 382
314 based on the EEG-patterns, however the prognostication protocol 18. Bouwes A, Binnekade JM, Verbaan BW, et al. Predictive value of neurological 383

315 was multimodal in all hospitals and not solely based on the EEG- examination for early cortical responses to somatosensory evoked potentials in 384
patients with postanoxic coma. J Neurol 2012;259:537–41. 385
316 patterns. Different examiners in each hospital analysed the EEGs 19. Wijdicks EF, Hijdra A, Young GB, Bassetti CL, Wiebe S. Practice parameter: pre- 386
317 according to a standardized classification, and the findings of this diction of outcome in comatose survivors after cardiopulmonary resuscitation 387

318 study thus represent the basis for general decision-making in every (an evidence-based review): report of the Quality Standards Subcommittee of 388
the American Academy of Neurology. Neurology 2006;67:203–10. 389
319 day clinical practice. 20. Rundgren M, Westhall E, Cronberg T, Rosen I, Friberg H. Continuous amplitude- 390
320 In conclusion, our observations suggest that EEG is a useful integrated electroencephalogram predicts outcome in hypothermia-treated 391
321 adjunct in prognostication of patients with OHCA and that EEG pat- cardiac arrest patients. Crit Care Med 2010;38:1838–44. 392
21. Hirsch LJ, LaRoche SM, Gaspard N, et al. American Clinical Neurophysiology 393
322 terns associated with a poor outcome as well as patterns associated Society’s standardized critical care EEG terminology: 2012 version. J Clin Neu- 394
323 with a good chance of a favourable outcome can be identified. The rophysiol 2013;30:1–27. 395
324 current study suggests that Rhythmic Delta Activity (RDA) is asso- 22. Soholm H, Wachtell K, Nielsen SL, et al. Tertiary centres have improved sur- 396
vival compared to other hospitals in the Copenhagen area after out-of-hospital 397
325 ciated with a better prognosis, whereas Periodic Discharges (PD),
cardiac arrest. Resuscitation 2013;84:162–7. 398
326 suppressed voltage and burst-suppression patterns are associated 23. Deakin CD, Nolan JP, Soar J, et al. European Resuscitation Council Guidelines 399
327 with a very poor prognosis. Prospective blinded studies should for Resuscitation 2010 Section 4. Adult advanced life support. Resuscitation 400
2011;81:1305–52. 401
328 assess the diagnostic accuracy of EEG for prognostication.
24. Nolan JP, Deakin CD, Soar J, Bottiger BW, Smith G, European Resuscitation Coun- 402
cil. European Resuscitation Council guidelines for resuscitation 2005. Section 4. 403
329 Funding Adult advanced life support. Resuscitation 2005;67:S39–86. 404
25. Cummins RO, Chamberlain DA, Abramson NS, et al. Recommended guidelines 405
for uniform reporting of data from out-of-hospital cardiac arrest: the Utstein 406
Q4
330 This work was supported by The European Union (EU) InterReg Style. Task Force of the American Heart Association, the European Resuscita- 407
Q5 IV A program.
331 tion Council, the Heart and Stroke Foundation of Canada, and the Australian 408
Resuscitation Council. Ann Emerg Med 1991;20:861–74. 409
26. Jacobs I, Nadkarni V, Bahr J, et al. Cardiac arrest and cardiopulmonary resus- 410
332 Conflict of interest statement citation outcome reports: update and simplification of the Utstein templates 411
for resuscitation registries: a statement for healthcare professionals from a task 412

333 No conflicts of interest to declare. force of the International Liaison Committee on Resuscitation (American Heart 413
Association, European Resuscitation Council, Australian Resuscitation Council, 414
New Zealand Resuscitation Council, Heart and Stroke Foundation of Canada, 415
334 References InterAmerican Heart Foundation, Resuscitation Councils of Southern Africa). 416
Circulation 2004;110:3385–97. 417

335 1. Wissenberg M, Lippert FK, Folke F, et al. Association of national initiatives 27. Langhelle A, Nolan J, Herlitz J, et al. Recommended guidelines for reviewing, 418

336 to improve cardiac arrest management with rates of bystander intervention reporting, and conducting research on post-resuscitation care: the Utstein style. 419

337 and patient survival after out-of-hospital cardiac arrest. J Am Med Assoc Resuscitation 2005;66:271–83. 420

338 2013;310:1377–84. 28. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying 421

339 2. Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypother- prognostic comorbidity in longitudinal studies: development and validation. J 422

340 mia to improve the neurologic outcome after cardiac arrest. N Engl J Med Chronic Dis 1987;40:373–83. 423

341 2002;346:549–56. 29. Schmidt M, Jacobsen JB, Lash TL, Botker HE, Sorensen HT. 25 year trends in first 424

342 3. Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivors time hospitalisation for acute myocardial infarction, subsequent short and long 425

343 of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med term mortality, and the prognostic impact of sex and comorbidity: a Danish 426

344 2002;346:557–63. nationwide cohort study. Br Med J 2012;344:e356. 427

345 4. Rossetti AO, Oddo M, Logroscino G, Kaplan PW. Prognostication after cardiac 30. Mateen FJ, Josephs KA, Trenerry MR, et al. Long-term cognitive outcomes fol- 428

346 arrest and hypothermia: a prospective study. Ann Neurol 2010;67:301–7. lowing out-of-hospital cardiac arrest: a population-based study. Neurology 429

347 5. Nielsen N, Wetterslev J, Cronberg T, et al. Targeted temperature manage- 2011;77:1438–45. 430

348 ment at 33 degrees C versus 36 degrees C after cardiac arrest. N Engl J Med 31. Levy DE, Caronna JJ, Singer BH, Lapinski RH, Frydman H, Plum F. Predicting 431

349 2013;369:2197–206. outcome from hypoxic-ischemic coma. J Am Med Assoc 1985;253:1420–6. 432

350 6. Perman SM, Kirkpatrick JN, Reitsma AM, et al. Timing of neuroprognostication 32. Bro-Jeppesen J, Kjaergaard J, Horsted TI, et al. The impact of therapeutic 433

351 in postcardiac arrest therapeutic hypothermia. Crit Care Med 2012;40:719–24. hypothermia on neurological function and quality of life after cardiac arrest. 434

352 7. Dragancea I, Rundgren M, Englund E, Friberg H, Cronberg T. The influence of Resuscitation 2009;80:171–6. 435

353 induced hypothermia and delayed prognostication on the mode of death after 33. Gold B, Puertas L, Davis SP, et al. Awakening after cardiac arrest and post 436

354 cardiac arrest. Resuscitation 2013;84:337–42. resuscitation hypothermia: are we pulling the plug too early? Resuscitation 437

355 8. Cronberg T, Brizzi M, Liedholm LJ, et al. Neurological prognostication after 2014;85:211–4. 438

356 cardiac arrest—recommendations from the Swedish Resuscitation Council. 34. Grossestreuer AV, Abella BS, Leary M, et al. Time to awakening and neurologic 439

357 Resuscitation 2013;84:867–72. outcome in therapeutic hypothermia-treated cardiac arrest patients. Resuscita- 440

358 9. Prognostication of comatose patients after cardiac arrest; 2013. Available from: tion 2013;84:1741–6. 441

359 http://www.cardio.dk/rapporter/andre-selskaber. 35. Milani P, Malissin I, Tran-Dinh YR, et al. Prognostic EEG patterns in patients 442

360 10. Cocchi MN, Lucas JM, Salciccioli J, et al. The role of cranial computed tomo- resuscitated from cardiac arrest with particular focus on Generalized Periodic 443

361 graphy in the immediate post-cardiac arrest period. Intern Emerg Med 2010;5: Epileptiform Discharges (GPEDs). Neurophysiol Clin 2014;44:153–64. 444

533–8.

Please cite this article in press as: Søholm H, et al. Prognostic value of electroencephalography (EEG) after out-
of-hospital cardiac arrest in successfully resuscitated patients used in daily clinical practice. Resuscitation (2014),
http://dx.doi.org/10.1016/j.resuscitation.2014.08.031