Accepted Manuscript

Title: Survey on Current Practices for NeurologicalPrognostication after Cardiac Arrest

Author: Hans Friberg Tobias Cronberg Martin W. DunserJacques Duranteau Janneke Horn Mauro Oddo

PII: S0300-9572(15)00036-2DOI: http://dx.doi.org/doi:10.1016/j.resuscitation.2015.01.018Reference: RESUS 6279

To appear in: Resuscitation

Received date: 28-10-2014Revised date: 12-1-2015Accepted date: 16-1-2015

Please cite this article as: Friberg H, Cronberg T, Dunser MW, Duranteau J, Horn J,Oddo M, Survey on Current Practices for Neurological Prognostication after CardiacArrest, Resuscitation (2015), http://dx.doi.org/10.1016/j.resuscitation.2015.01.018

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

Page 1 of 18

Accep

ted

Man

uscr

ipt

1

Survey on Current Practices for Neurological 1

Prognostication after Cardiac Arrest 2

3

RUNNING TITLE: Prognostication after Cardiac Arrest 4

5

AUTHORS: 6

Hans Friberg1, Tobias Cronberg2, Martin W. Dünser3, Jacques Duranteau4, 7

Janneke Horn5, Mauro Oddo6 8

1 Skåne University Hospital, Dept of Anaesthesiology and Intensive Care, 9

Lund University, Sweden 10

2 Skåne University Hospital, Dept of Neurology, Lund University, Sweden 11

3 Salzburg General Hospital and Paracelsus Private Medical University, 12

Salzburg, Austria 13

4 Dept of Anaesthesia and Intensive Care, Bicêtre Hospital, Hôpitaux 14

universitaires Paris-Sud, Université Paris-Sud, Assistance Publique - 15

Hôpitaux de Paris, 94275 Le Kremlin Bicêtre, France 16

5 Academic Medical Center, Dept of Intensive Care, Amsterdam, The 17

Netherlands 18

6 Department of Intensive Care Medicine, CHUV-Lausanne University, 19

Switzerland 20

Word count: 2494 21

22

23

Corresponding author: 24

Hans Friberg 25

e-mail: Hans.Friberg@skane.se 26

Telephone: 0046-46-177953 27

Fax: 0046-46-17605028

*Manuscript

Page 2 of 18

Accep

ted

Man

uscr

ipt

2

29 ABSTRACT 30

PURPOSE: 31

To investigate current practices and timing of neurological prognostication in comatose 32

cardiac arrest patients. 33

METHODS: 34

An anonymous questionnaire was distributed to the 8000 members of the European 35

Society of Intensive Care Medicine during September and October 2012. The survey had 36

27 questions divided into three categories; background data, clinical data, decision-37

making & consequences. 38

RESULTS: 39

A total of 1025 respondents (13 %) answered the survey with complete forms in more 40

than 90 %. Twenty per cent of respondents practiced outside of Europe. Overall, 22 % 41

answered that they had national recommendations, with the highest percentage in the 42

Netherlands (>80 %). Eighty-nine per cent used induced hypothermia (32-34°C) for 43

comatose cardiac arrest patients, while 11 % did not. Twenty per cent had separate 44

prognostication protocols for hypothermia patients. Seventy-nine per cent recognized 45

that neurological examination alone is not enough to predict outcome and a similar 46

number (76 %) used additional methods. Intermittent electroencephalography (EEG), 47

brain computed tomography (CT) scan and evoked potentials (EP) were considered 48

most useful. Poor prognosis was defined as cerebral performance category (CPC) 3-5 49

(58 %) or CPC 4-5 (39 %) or other (3 %). When prognosis was considered poor, 73 % 50

would actively withdraw intensive care while 20 % would not and 7 % were uncertain. 51

CONCLUSION: 52

National recommendations for neurological prognostication after cardiac arrest are 53

uncommon and only one physician out of five uses a separate protocol for hypothermia 54

treated patients. A neurological examination alone was considered insufficient to predict 55

outcome in comatose patients and most respondents advocated a multimodal approach; 56

EEG, brain CT and EP were considered most useful. Uncertainty regarding neurological 57

prognostication and decisions on level of care was substantial. 58

59

Page 3 of 18

Accep

ted

Man

uscr

ipt

3

60 INTRODUCTION 61

Cardiac arrest is common and, in spite of improvements in care in recent years, only one 62

of ten patients survives.1, 2 Most patients with return of spontaneous circulation (ROSC) 63

who are admitted to a hospital are in coma and less than 50 % eventually wake up.3, 4 64

Early clinical signs of a poor neurological prognosis, such as the absence of ocular 65

reflexes on admission, cannot in general inform the clinician about survival chances for 66

the individual patient.5, 6 The best prognostic sign after cardiac arrest is that of 67

awakening but due to modern intensive care, including temperature management and 68

coronary intervention of comatose survivors, most patients are pharmacologically 69

affected and cannot be adequately assessed during the first days.7 70

71

Current American recommendations regarding neurological prognostication after 72

cardiac arrest have been widely adopted but are not optimal since they are based on 73

studies performed prior to the era of temperature management and coronary 74

intervention.8 As a result, current practices for assessment of coma after cardiac arrest 75

may differ between and within countries. National recommendations have been 76

published in the Netherlands and in Sweden.9 Updated and evidence-based guidelines 77

are requested and work is ongoing by several societies. Recently, an advisory statement 78

from the European Resuscitation Council (ERC) and the European Society of Emergency 79

Medicine (ESICM) was published.10 Most authors in the field recommend a multimodal 80

approach using several independent methods in addition to a clinical neurological 81

investigation, which remains the foundation.11, 12 Additional methods advocated include 82

computed tomography (CT) scan of the brain, magnetic resonance imaging (MRI), 83

electroencephalography (EEG), evoked potentials (EP) and biomarkers for brain 84

damage.13-15 85

The aim of the present survey was to investigate current practices for neurological 86

prognostication after cardiac arrest among members of the ESICM. In addition, data on 87

the use of induced hypothermia were collected. 88

89

METHODS 90 An anonymous questionnaire was generated by the authors and endorsed by the 91

European Research Committee of the ESICM. A link to the electronic survey was 92

distributed by e-mail through the ESICM membership database, consisting of 93

Page 4 of 18

Accep

ted

Man

uscr

ipt

4

approximately 8000 members, during September and October 2012. All replies were 94

collected in a central database and saved. The survey consisted of 27 questions, 10 on 95

background data, 11 on clinical data and 6 on decision-making and consequences 96

(ESICM survey, Electronic Supplementary Material (ESM)). All survey responses were 97

used for data analysis. Descriptive data is presented per question including missing data. 98

No statistical analysis was performed. No approval from an ethical review board was 99

sought in any country. 100

101

RESULTS 102

A total of 1025 replies (13 %) were collected and analysed. Background data was 103

complete in all surveys, clinical data was complete in 984 (96 %) and decision-making 104

and consequences was complete in 951 (93 %). 105

106

Background data (Q 1-10) 107

Eighty per cent of respondents were from Europe (n=822) with high participation from 108

the United Kingdom (n=156), France (n=71) and Switzerland (n=70) while 20 % 109

practiced in countries outside of Europe (n=203) (Fig. 1). Respondents were mainly 110

attending physicians or consultants (73 %), 11 % were fellows and 16 % were residents. 111

Their main specialty was intensive care (74 %), anaesthesiology (19 %) or internal 112

medicine (3 %); remaining specialties had absolute numbers below ten. Most 113

respondents represented university or university-affiliated hospitals (67 %) with up to 114

20 ICU beds (Fig. 2a). The number of admitted cardiac arrest patients per year differed 115

but 10-30 patients was most common (Fig. 2b). Induced hypothermia (32-34°C) for 116

comatose cardiac arrest patients was used by 89 %, 61 % used hypothermia for a 117

majority of patients independently of initial rhythm or location of arrest, 15 % only for 118

patients with shockable rhythm, 13 % only for patients with OHCA and shockable 119

rhythm while 11 % did not use induced hypothermia (ESM Fig. 3). There were no 120

differences in the use of hypothermia between university and university-affiliated 121

hospitals as compared to other hospitals within a country (data not shown). National 122

recommendations on prognostication in postanoxic coma patients were reported by 22 123

% of respondents, while 35 % had local recommendations. The Netherlands stands out 124

with its high awareness of national recommendations (83 %). 125

126

Page 5 of 18

Accep

ted

Man

uscr

ipt

5

Clinical data (Q 11-21) 127

Glasgow coma scale (GCS) alone (25 %) or in combination with brain stem reflexes (66 128

%) was used by most respondents for prognostication. Other, mainly the FOUR-score 129

(Full Outline of UnResponsiveness) and the RLS (Reaction Level Scale) were used by 5 % 130

and no specified score by 4 %. A neurological examination was performed by the 131

intensivist (70 %), the nurse (30 %), and the neurologist (24 %) or by all of them (31 %) 132

(more than one answer possible), and at least daily (97 %), by many respondents twice 133

daily (60 %). Twenty per cent (n=197) of respondents had separate routines for 134

neurological prognostication for patients receiving hypothermia. Ongoing hypothermia 135

treatment alters the results of a neurological assessment according to 48 % of 136

respondents while 37 % answered that hypothermia in general, including the time after 137

the intervention, alters the results of a neurological assessment. The proportions 138

believing that ongoing sedation as opposed to sedation in general alters the results were 139

65 % and 30 % respectively. Seventy-nine per cent answered that a neurological 140

examination alone is not enough to predict outcome in comatose patients and 76 % used 141

additional methods. Fourteen per cent did not use additional methods because of lack of 142

resources and 10 % because they did not believe additional methods were useful. The 143

most commonly used additional methods as well as the three methods considered to be 144

most useful are listed in Table 1. Intermittent EEG, CT scan of the brain and EP were 145

considered most useful. 146

147

Decision-making and consequences (Q 12-27) 148

Fifty-eight per cent of respondents defined a poor outcome as a cerebral performance 149

category (CPC) of 3-5 while 39 % used a CPC of 4-5 and 3 % another definition. The final 150

decision on level of care was done as a collaborative effort between two or more 151

physicians according to 60 % of respondents while 36 % answered that the intensivist 152

alone made the decision, in 3 % the relatives and in 1 % the cardiologist. The most 153

common answer for when a decision was made was that no precise timing was used and 154

that it was based on individual patient data (44 %). Most physicians specifying a time 155

stated that a decision was commonly made at 3 days after the arrest or later and few 156

made a decision earlier than 48 hours (Fig. 4a). Seventy per cent considered prognosis 157

to be uncertain in 5 % or more of patients and 18 % of colleagues considered prognosis 158

to be uncertain in 30 % or more (Fig. 4b). 159

Page 6 of 18

Accep

ted

Man

uscr

ipt

6

When prognosis is considered poor, 73 % would actively withdraw care while 20 % 160

would not stop care and 7 % did not know. More than half of the respondents from 161

Greece (18/29, 62 %), Turkey (6/10, 60 %), Saudi Arabia (10/17, 59 %) and India 162

(21/39, 54 %) would not withdraw care in spite of a poor prognosis statement (ESM 163

Table 2). Access to an ethical council or equivalent for support was available for 54 % of 164

the respondents. 165

166

DISCUSSION 167

The main findings of this survey were that national recommendations for 168

prognostication after cardiac arrest are uncommon and that only one of five physicians 169

uses a separate protocol for hypothermia treated patients. A neurological examination 170

alone was considered insufficient to predict outcome in comatose patients and most 171

respondents used additional methods; EEG, brain CT and evoked potentials were 172

considered most useful. Uncertainty regarding neurological prognostication and 173

decisions on level of care was substantial. 174

A striking finding in the present survey is the prevailing uncertainty regarding 175

assessment of neurological prognosis with almost one of five considering prognosis to 176

be unreliable in 30 % or more at the time of evaluation. The reason for this uncertainty 177

may be a result of the ongoing debate highlighting that clinical findings alone are not 178

enough for a reliable prognosis when temperature management and concomitant 179

sedation are used.7, 16 Consequently, a majority answered that additional methods were 180

needed; the most widely used were EEG, CT scan and MRI of the brain, all commonly 181

accessible in most hospitals. When the respondents were asked to list what three 182

methods they believed were of greatest help to predict outcome, MRI was surpassed by 183

evoked potentials. Among evoked potentials, somatosensory evoked potentials (SSEP) 184

are best studied after cardiac arrest and are considered more reliable than both EEG and 185

brain CT.17 However, performance of SSEP requires specific skills and experience and 186

interobserver variability is not negligible.18 In addition, SSEP was used to predict 187

outcome in most trials where its reliability was tested, introducing a high risk of so-188

called self-fulfilling prophecy. Nevertheless, bilateral absence of N20 SSEP waves during 189

normothermia is considered a reliable predictor of poor outcome.17, 19 190

Our survey showed that routine EEG and brain CT were the most widely used methods 191

to predict outcome after cardiac arrest which is an interesting finding and in line with a 192

Page 7 of 18

Accep

ted

Man

uscr

ipt

7

previous Norwegian survey.20 Neither EEG nor brain CT is included in the current 193

American guidelines (8) and the evidence to use either for prognostication purposes is 194

limited. Both methods are, however, incorporated in a prognostication algorithm that 195

was recently published.10 A plausible explanation for their general use is their 196

accessibility and their relatively low cost. It was not within the scope of the present 197

survey to analyse how the results from routine EEG and brain CT were used. 198

Generalized suppression to <20 µV, burst-suppression pattern with generalized 199

epileptiform activity, or generalized periodic complexes on a flat background are 200

considered “malignant” EEG-patterns and strongly, but not invariably, associated with a 201

poor outcome.8 In addition, a non-reactive EEG background pattern after rewarming 202

from hypothermia has been found to be a strong predictor of poor outcome.21 Also, a 203

return of a continuous EEG background pattern at any time after cardiac arrest using 204

continuous EEG-monitoring was strongly associated with recovery and a good 205

outcome.22 A major problem with EEG interpretations is the lack of consensus regarding 206

EEG terminology, a problem that may become less prominent by the introduction of a 207

standardized classification for critical care EEG that was recently proposed.23 Next to 208

EEG, a CT scan of the brain was the most commonly used examination to assess 209

prognosis. This is likely related to the extensive use of brain CT to exclude traumatic 210

injury early after cardiac arrest and to the fact that this technique is readily available at 211

all hospitals. Evidence supporting the use of brain CT for prognostication is however 212

limited and restricted by selection bias.24, 25 The widespread use of brain CT for 213

prognostication may therefore be a reason for concern but emphasizes at the same time 214

the need of large prospective studies evaluating commonly used methodologies such as 215

brain CT and routine EEG. 216

217

Biomarkers were the least commonly used predictors in our survey despite their 218

simplicity, independence from sedation and low cost. The published literature on 219

biomarkers has pointed in different directions and study results have in general been 220

inconclusive, which may be explained by small sample sizes, variability in test results 221

between laboratories and a lack of standard.26, 27 222

223

Another striking finding in our survey was the lack of national and/or local 224

recommendations for neurological prognostication after cardiac arrest in many 225

Page 8 of 18

Accep

ted

Man

uscr

ipt

8

countries and hospitals. Only 22 % claimed that they had national recommendations 226

with physicians in the Netherlands standing out with their high awareness. 227

Approximately one of three claimed that they had local recommendations, which is low 228

but at least a great leap forward as compared with a previous investigation.20 After 229

conducting the survey, national recommendations from Sweden have been launched and 230

published.9 New and updated international guidelines are underway, and a joint 231

advisory statement from the ESICM and the ERC has been published.10 Since legal, 232

cultural and religious matters differ between countries, national societies may need to 233

adapt international guidelines into useful national practice guidelines. 234

235

Induced hypothermia was used by almost nine of ten respondents, six of whom used 236

induced hypothermia for most cardiac arrest patients independently of initial rhythm or 237

location of arrest. A minority used separate routines for neurological prognostication in 238

patients receiving hypothermia, although almost all believed that either hypothermia or 239

concurrent sedation or both alter the results of a neurological assessment. As long as 240

clinicians acknowledge that ongoing or lingering sedation will influence the neurological 241

examination and that sufficient time is allowed to pass, common prognostication 242

protocols for patients treated with or without hypothermia may not constitute a 243

problem. Importantly, very few made a decision on level of care earlier than 48 hours 244

after arrest while a majority made a decision 3 days or later and almost as many 245

answered that there was no precise timing, rather that timing was based on the 246

performance of the individual patient. 247

248

A poor outcome was defined as CPC 3-5 by a majority of respondents but many believed 249

that CPC 4-5 was a better definition. The trend during the last decade has been that 250

more studies use CPC 3-5 as a poor outcome measure while CPC 4-5 was more 251

commonly used before.28 An interesting finding was that two of ten would not withdraw 252

care when a poor prognosis statement had been made, which probably is a consequence 253

of legal, cultural and religious traditions among individuals and in different countries. 254

Respondents who declared that they would not withdraw care under these 255

circumstances more often practiced in countries in southern Europe, the Middle East, 256

India and the USA. It would be of great interest to systematically study the reliability of 257

prognostic indicators and the natural course of comatose survivors after cardiac arrest 258

Page 9 of 18

Accep

ted

Man

uscr

ipt

9

in these countries. Most prognostication studies are hampered by the impact of the so-259

called self-fulfilling prophecy and there are few systematic studies on the long-term 260

recovery of patients with prolonged coma after cardiac arrest.29 Finally, an ethical 261

council or equivalent for support in decision-making was available for approximately 262

half of the respondents in our survey but whether or not this support was used was not 263

investigated. 264

265

LIMITATIONS 266

A group of ICU physicians, all members of the ESICM, were asked to complete the survey. 267

Approximately one of eight answered the survey and the results may not represent the 268

common view in Europe or abroad since physicians with a special interest in 269

neurological prognostication after cardiac arrest may have been over-represented. In 270

addition, our results represent above all the view of European colleagues. The survey 271

was anonymous and we have no knowledge of whether some respondents were from 272

the same institution. No systematic pre-testing or validation of the survey was 273

performed. 274

275

CONCLUSIONS 276

National recommendations are uncommon and only one of five physicians uses a 277

separate protocol for hypothermia treated patients. Neurological examination alone was 278

considered insufficient to predict outcome in comatose patients by the majority and 279

most respondents advocated a multimodal approach using several additional methods; 280

EEG, brain CT and SSEP were considered most useful. Uncertainty regarding 281

neurological prognostication and decisions on level of care was substantial. 282

283

284 285 286

ACKNOWLEDGEMENT: 287

No specific funding. 288

This project was endorsed by the European Research Committee of the European 289

Society of Intensive Care Medicine. 290

291

292

Page 10 of 18

Accep

ted

Man

uscr

ipt

10

293

CONFLICT OF INTEREST STATEMENT: 294

Hans Friberg has received lecture fees from Bard Medical and Natus Inc. 295

All other authors declare no conflict of interest.296

Page 11 of 18

Accep

ted

Man

uscr

ipt

11

References: 297

298

1. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of-299 hospital cardiac arrest: a systematic review and meta-analysis. Circ Cardiovasc 300 Qual Outcomes. 2010;3:63-81. 301

2. Stromsoe A, Svensson L, Axelsson AB, et al. Improved outcome in Sweden after 302 out-of-hospital cardiac arrest and possible association with improvements in 303 every link in the chain of survival. Eur Heart J. 2014. 304

3. Dragancea I, Rundgren M, Englund E, Friberg H, Cronberg T. The influence of 305 induced hypothermia and delayed prognostication on the mode of death after 306 cardiac arrest. Resuscitation. 2013;84:337-42. 307

4. Nielsen N, Wetterslev J, Cronberg T, et al. Targeted temperature management at 308 33 degrees C versus 36 degrees C after cardiac arrest. N Engl J Med. 309 2013;369:2197-206. 310

5. Earnest MP, Breckinridge JC, Yarnell PR, Oliva PB. Quality of survival after out-of-311 hospital cardiac arrest: predictive value of early neurologic evaluation. 312 Neurology. 1979;29:56-60. 313

6. Okada K, Ohde S, Otani N, et al. Prediction protocol for neurological outcome for 314 survivors of out-of-hospital cardiac arrest treated with targeted temperature 315 management. Resuscitation. 2012;83:734-9. 316

7. Samaniego EA, Mlynash M, Caulfield AF, Eyngorn I, Wijman CA. Sedation 317 confounds outcome prediction in cardiac arrest survivors treated with 318 hypothermia. Neurocrit Care. 2011;15:113-9. 319

8. Wijdicks EF, Hijdra A, Young GB, Bassetti CL, Wiebe S, Quality Standards 320 Subcommittee of the American Academy of N. Practice parameter: prediction of 321 outcome in comatose survivors after cardiopulmonary resuscitation (an 322 evidence-based review): report of the Quality Standards Subcommittee of the 323 American Academy of Neurology. Neurology. 2006;67:203-10. 324

9. Cronberg T, Brizzi M, Liedholm LJ, et al. Neurological prognostication after 325 cardiac arrest--recommendations from the Swedish Resuscitation Council. 326 Resuscitation. 2013;84:867-72. 327

10. Sandroni C, Cariou A, Cavallaro F, et al. Prognostication in comatose survivors of 328 cardiac arrest: An advisory statement from the European Resuscitation Council 329 and the European Society of Intensive Care Medicine. Resuscitation. 2014:1779-330 89. 331

11. Cronberg T, Rundgren M, Westhall E, et al. Neuron-specific enolase correlates 332 with other prognostic markers after cardiac arrest. Neurology. 2011;77:623-30. 333

12. Oddo M, Rossetti AO. Early multimodal outcome prediction after cardiac arrest in 334 patients treated with hypothermia. Crit Care Med. 2014;42:1340-7. 335

13. Friberg H, Cronberg T. Prognostication after cardiac arrest. Best Pract Res Clin 336 Anaesthesiol. 2013;27:359-72. 337

14. Oddo M, Rossetti AO. Predicting neurological outcome after cardiac arrest. Curr 338 Opin Crit Care. 2011;17:254-9. 339

15. Taccone F, Cronberg T, Friberg H, et al. How to assess prognosis after cardiac 340 arrest and therapeutic hypothermia. Crit Care. 2014;18:202. 341

16. Friberg H, Rundgren M, Westhall E, Nielsen N, Cronberg T. Continuous evaluation 342 of neurological prognosis after cardiac arrest. Acta Anaesthesiol Scand. 343 2013;57:6-15. 344

Page 12 of 18

Accep

ted

Man

uscr

ipt

12

17. Bouwes A, Binnekade JM, Kuiper MA, et al. Prognosis of coma after therapeutic 345 hypothermia: a prospective cohort study. Ann Neurol. 2012;71:206-12. 346

18. Pfeifer R, Weitzel S, Gunther A, et al. Investigation of the inter-observer 347 variability effect on the prognostic value of somatosensory evoked potentials of 348 the median nerve (SSEP) in cardiac arrest survivors using an SSEP classification. 349 Resuscitation. 2013;84:1375-81. 350

19. Kamps MJ, Horn J, Oddo M, et al. Prognostication of neurologic outcome in cardiac 351 arrest patients after mild therapeutic hypothermia: a meta-analysis of the current 352 literature. Intensive Care Med. 2013;39:1671-82. 353

20. Busch M, Soreide E. Prognostication after out-of-hospital cardiac arrest, a clinical 354 survey. Scand J Trauma Resusc Emerg Med. 2008;16:9. 355

21. Rossetti AO, Carrera E, Oddo M. Early EEG correlates of neuronal injury after 356 brain anoxia. Neurology. 2012;78:796-802. 357

22. Rundgren M, Westhall E, Cronberg T, Rosen I, Friberg H. Continuous amplitude-358 integrated electroencephalogram predicts outcome in hypothermia-treated 359 cardiac arrest patients. Crit Care Med. 2010;38:1838-44. 360

23. Hirsch LJ, LaRoche SM, Gaspard N, et al. American Clinical Neurophysiology 361 Society's Standardized Critical Care EEG Terminology: 2012 version. J Clin 362 Neurophysiol. 2013;30:1-27. 363

24. Inamasu J, Miyatake S, Suzuki M, et al. Early CT signs in out-of-hospital cardiac 364 arrest survivors: Temporal profile and prognostic significance. Resuscitation. 365 2010;81:534-8. 366

25. Kim SH, Choi SP, Park KN, Youn CS, Oh SH, Choi SM. Early brain computed 367 tomography findings are associated with outcome in patients treated with 368 therapeutic hypothermia after out-of-hospital cardiac arrest. Scand J Trauma 369 Resusc Emerg Med. 2013;21:57. 370

26. Mlynash M, Buckwalter MS, Okada A, et al. Serum neuron-specific enolase levels 371 from the same patients differ between laboratories: assessment of a prospective 372 post-cardiac arrest cohort. Neurocrit Care. 2013;19:161-6. 373

27. Rundgren M, Cronberg T, Friberg H, Isaksson A. Serum neuron specific enolase - 374 impact of storage and measuring method. BMC Res Notes. 2014;7:726. 375

28. Sandroni C, Cavallaro F, Callaway CW, et al. Predictors of poor neurological 376 outcome in adult comatose survivors of cardiac arrest: a systematic review and 377 meta-analysis. Part 1: patients not treated with therapeutic hypothermia. 378 Resuscitation. 2013;84:1310-23. 379

29. Howell K, Grill E, Klein AM, Straube A, Bender A. Rehabilitation outcome of 380 anoxic-ischaemic encephalopathy survivors with prolonged disorders of 381 consciousness. Resuscitation. 2013;84:1409-15. 382

383

Page 13 of 18

Accep

ted

Man

uscr

ipt

Method A B

EEG, intermittent 617 (63%) 488 (50%)

Neuroimaging (CT-scan) 576 (58%) 413 (42%)

Evoked potentials 351 (36%) 397 (40%)

Neuroimaging (MRI) 390 (40%) 235 (24%)

Biomarkers, NSE 187 (19%) 219 (22%)

EEG, continuous 113 (11%) 177 (18%)

Biomarkers, S-100B 48 (5%) 62 (6%)

Other 42 (4%) 54 (5%)

Table 1. Respondents (n=984) were asked the following questions: A: “Which tools do you use in routine clinical practice (more than one answer possible)?” B: “Which three tools do you consider most useful to assess prognosis after cardiac arrest and induced hypothermia? (please mark three tools/options)”

Table

Page 14 of 18

Accep

ted

Man

uscr

ipt

Figure 1 - pie chart1

Page 15 of 18

Accep

ted

Man

uscr

ipt

Figure 1 - histogram

Page 16 of 18

Accep

ted

Man

uscr

ipt

Figure 2

Page 17 of 18

Accep

ted

Man

uscr

ipt

Figure 4

Page 18 of 18

Accep

ted

Man

uscr

ipt

Figure captions:

Fig. 1 All 1025 respondents of the survey divided per continent; 80 % were from Europe

(n=822) with notably high participation from the United Kingdom (n=171), France

(n=71) and Switzerland (n=70). Of the remaining 203 respondents, 10 % practiced in

Asia, 5 % in the Middle East, 3 % in South America and 2 % in North America

Fig. 2 All 1025 respondents were asked about a. the number of beds in their intensive

care unit (ICU) and b. the approximate number of treated cardiac arrest patients per

year in their ICU, with or without hypothermia.

Fig. 4 The answers by the respondents (n=951) to the following questions are

presented: a: ”In patients remaining comatose after cardiac arrest, when do you in

general make a decision on prognosis and level of care at your institution?” Percentages

indicate the proportion of respondents. b: ”In your opinion, in what percentage of

patients remaining comatose after cardiac arrest do you consider that prognosis may

still be uncertain at the time specified above (A)?”

Figure captures - Fig 1, 2, 4