Predictive Value of Amplitude Integrated EEG (aEEG) after Rescue

Hypothermic Neuroprotection for Hypoxic Ischemic Encephalopathy: A

Meta-analysis

Manigandan Chandrasekaran MD FRCPCH1, 2, Badr Chaban FRCPCH1,

Paolo Montaldo PhD1, Sudhin Thayyil PhD FRCPCH1.

1. Centre for Perinatal Neuroscience, Imperial College London, London, UK

2. Cloudnine Hospital, Chennai, India

Running title: Predictive value of aEEG after therapeutic hypothermia

Funding support: None received

Corresponding author

Dr Manigandan Chandrasekaran MD FRCPCH

Centre for Perinatal Neuroscience

Department of Paediatrics

Imperial College London

Ducane Road

London W12 0HS.

Email: mani@pediatrician.com

Phone: 00919840653244

Fax: 0044 2033131122

1

Abstract

Background:  Amplitude-integrated electro-encephalography (aEEG) is a

useful bedside tool in predicting the neurodevelopmental outcome after

neonatal encephalopathy, however the prognostic accuracy may be altered by

rescue hypothermic neuroprotection.

Objective:  To examine the prognostic accuracy of aEEG for predicting long

term neurodevelopmental outcomes in term newborn infants undergoing

therapeutic hypothermia for neonatal encephalopathy.

Study Design:  We examined all studies (Medline, CINAHL, and the

Cochrane Library; 2000 to 2014) comparing aEEG (6, 24, 48 or 72 hours) in

term encephalopathic babies undergoing therapeutic hypothermia, with

neurodevelopmental outcome at 1 year or more. We extracted individual

patient data from the eligible studies to calculate prognostic indices with exact

confidence intervals (CI). We considered continuous normal voltage as

normal aEEG pattern, and discontinuous normal voltage, burst suppression,

flat trace and persistently low voltage as abnormal, and defined adverse

outcome as death or moderate/severe disability at 1 year.

Results:  We reviewed a total of 70 articles; 17 of which met the inclusion

criteria. 8 studies were excluded, 9 studies (Nn=520) were included in the

meta-analysis. The pooled sensitivity and specificity for an abnormal trace at

6 hours of age to predict adverse outcome were 96% (95%CI 91 – 98%) and

39% (95%CI 32 – 46%). The diagnostic odds ratios of an abnormal trace was

peak at 48 hours [ere 8.2 (95%CI 3.8, 17.7), 39.6 (95%CI 13.1, 119), 66.9

(95%CI 19.7, 227.2)] and 47.99 (95%CI 11.4, 201.7) at 6, 24, 48 and 72 hours

2

respectively.

Conclusions:  A persistantly abnormal aEEG at 48 hours or more is

associated with an adverse neurodevelopmal outcome. The positive

prognostic value of 6-hour aEEG is poor, and good outcome may occur

despite abnormal aEEG. Conversely, a normal 6 hour aEEG has a good

negative predictive value although do not exclude adverse outcomes.

3

Introduction

The bedside assessment of brain injury and accurate prediction of

neurodevelopmental outcome remains the major challenge for clinicians

treating infants with neonatal encephalopathy (NE) in order to counsel the

parents and to allow the most appropriate clinical management. Amplitude-

integrated electroencephalography (aEEG) is a widely used bedside tool to

identify potential candidates for therapeutic hypothermia after a perinatal

asphyxia insult, for identifying seizures, and for guiding limitation of life

sustaining therapy in severely encephalopathic infants. The aEEG acquired

within the first 6 hours of age was considered one of the best predictors of

neurological outcome at 18 months in non-cooled NE infants(1). However,

since the widespread use of cooling therapy, the predictive value of early

aEEG has changed and NE infants have been shown to have a normal

neurological outcome if the aEEG background voltage activity recovers by 48

h (2-4).

The aim of this study was to review systematically the published literature and

perform a meta-analysis of the prognostic accuracy of aEEG for predicting

long-term neurodevelopmental outcomes in near term and term newborn

infants with NE undergoing rescue hypothermic neuroprotection.

Methods

We followed the guidelines of the Cochrane Diagnostic Test Accuracy

Working Group for undertaking and reporting the meta-analysis. We included

all the studies that compared the prognostic accuracy of aEEG during the first

4

few days after birth with neurodevelopmental outcome at 1 year of more, after

therapeutic hypothermia (selective head cooling or whole body cooling) for NE

neonatal encephalopathy in babies >36 weeks. If both hypothermic and

normothermic infants were included in a study, we extracted the aEEG data of

the hypothermic infants, separately. We excluded the study if such separate

analysis was not possible or if the aEEG background was not available any

described as in pattern or voltage classification(1, 5). ption of the pattern.

Index test – aEEG

We used the following definition for each tracing: CNV (continuous normal

voltage) is a continuous background activity with voltage of 10 to 25

microvolts; DNV (discontinuous normal voltage) is a discontinuous trace with

voltage >5 microvolts; BS (burst suppression) is discontinuous trace with

periods of low cortical activity (<5 microvolts) with bursts of higher amplitude;

CLV (continuous low voltage) is continuous background pattern of voltage,

around or less than 5 microvolts; FT (flat tracing) is mainly isoelectric tracing

of <5 microvolts; and epileptiform activity is a single or repetitive event with

sustained cortical activity. We considered continuous normal voltage pattern

as normal and discontinuous normal voltage, burst suppression, flat trace and

persistently low voltage as abnormal (6). If the background was described by

voltage method(5), we considered normal trace as normal, moderately

abnormal, severely abnormal trace and burst suppression as abnormal. We

collected the aEEG data at 6 hours, 24 hours, 48 hours and 72 hours after

birth, whenever available.

Reference test

5

All included babies had a detailed neurological examination and

neurodevelopmental score using a standard measurement tool [Griffiths

Mental Developmental Scales, Bayley Scales of Infant Development; BSID,

Alberta infant Motor Scale (7)] at 1 year or more. The presence and type of

cerebral palsy was determined according to Hagberg(8) and severity

classified using the Gross Motor Function Classification System (GMFCS) (9).

We considered death or moderate or severe neurodisability as adverse

outcomes. Moderate-severe neurodisability was defined in the included

studies as any of the following: GMFCS level 3 to 5 (corresponding to

moderate/severe cerebral palsy); hearing impairment that required hearing

aids; blindness; mental development index less than 70 (2, 4, 10, 11) or

Griffiths Mental Developmental Scales DQ less than 85 (12). Hallberg et al

(3) considered severe disability in case of neurological examinations (13) with

overt signs of spasticity and an Alberta infant Motor Scale (7) score below the

5th percentile.

Search strategy

We searched Medline (PubMed interphase January 2005 to September 2014)

and Embase (interphase January 2005 to September 2014) by using both

Medical Subject Heading terms and the following key words:

("infant, newborn"[MeSH Terms] OR ("infant"[All Fields] AND "newborn"[All

Fields]) OR "newborn infant"[All Fields] OR "newborn"[All Fields]) OR

("infant"[MeSH Terms] OR "infant"[All Fields]) OR ("infant, newborn"[MeSH

Terms] OR ("infant"[All Fields] AND "newborn"[All Fields]) OR "newborn

6

infant"[All Fields] OR "neonate"[All Fields]) And Ischemia: ("anoxia"[MeSH

Terms] OR "anoxia"[All Fields] OR "hypoxia"[All Fields]) OR ("ischaemia"[All

Fields] OR "ischemia"[MeSH Terms] OR "ischemia"[All Fields]) OR

encephalopathy[All Fields] OR ("asphyxia"[MeSH Terms] OR "asphyxia"[All

Fields]) And Hypothermia: ("hypothermia, induced"[MeSH Terms] OR

("hypothermia"[All Fields] AND "induced"[All Fields]) OR "induced

hypothermia"[All Fields] OR ("induced"[All Fields] AND "hypothermia"[All

Fields])) OR cool[All Fields] And amplitude integrated EEG

(amplitude-integrated[All Fields] AND ("electroencephalography"[MeSH

Terms] OR "electroencephalography"[All Fields] OR

"electroencephalogram"[All Fields]))

We also searched the Cochrane CENTRAL library, and conference abstracts.

We included only human studies.

Statistical analysis and data synthesis  

Three investigators (MC, BC and PM) independently extracted the individual

patient data from each of the studies into a predefined database. The fourth

reviewer (ST) resolved any inter-reviewer differences. We used Revman

(version 5Version 5.3. Copenhagen: The Nordic Cochrane Centre, The

Cochrane Collaboration, 2014.) for meta-analysis. We then pooled the

individual patient data from all studies to create 2x2 tables, and calculated the

prognostic indices with 95% confidence intervals using exact statistics (14)

(14). For direct comparison of diagnostic utility at various time points, we used

diagnostic odds ratios(15) and area under the curve (AUC). Inter study

heterogeneity was investigated using Chi-square tests and I2 index.

7

To minimize clinical heterogeneity, we excluded all preterm infants (< 36

weeks gestation), babies who did not receive hypothermic neuroprotection,

and infants who did non’t have neurodevelopmental assessment at the age of

12 months or after. We examined the quality of all the included studies using

Quadas 2 method(16); this included patient selection, index test, reference

standard, flow and timing and applicability concerns for each study.

Results

We reviewed 70 abstracts; 17 met the initial inclusion criteria and were

selected for review with full text. Five studies were excluded because they did

not report long-term outcome (17-21). One study was excluded because of

duplication (22), one study was excluded as we were unable to extract data

(23)and another one was excluded because of using only acquisition of sleep

wake cycle rather than full background of aEEG as an index test (24). Thus, a

total of 9 studies were included in the final meta-analysis (2-4, 10-12, 25-27).

(Figure 1)

Individual patient data were available from all the nine included studies (table

1); thus we compared the aEEG data from 520 cooled encephalopathic

babies who had neurodevelopmental outcome assessments aged >12

months. The pooled results of the meta-analysis, indicating the accuracy of

severe aEEG tracings acquired at different hours of age, to predict

moderate/severe disability or death, are summarized in figures 2 - 4.

8

Accuracy of 6-hour aEEG (prior to the initiation of therapeutic hypothermia)

Although aEEG had a good sensitivity 96% (CI 95%, 91-98) at 6 hours of age,

the specificity (39%; CI 95%, 32-46) was poor (Figure 2). The diagnostic

odds ratio (8.2; CI 95%, 3.8 – 17.7) for predicting adverse outcomes was

modest (Figure 4). A severely abnormal aEEG trace at 6 hours of age was

associated with an adverse outcome in 111 (94%) out of 117 infants. A good

outcome was seen in 216 cooled babies, out of which only 84 (39%) had a

normal background at 6 hours of age.

The predictive value of aEEG between 24 to 72 hours (i.e. during

therapeutic hypothermia)

The predictive values of an abnormal trace during hypothermia at 24 (n= 175

babies), 48 (n= 130 babies) and 72 (n= 125 babies) hours of age are shown in

Figure 3. The diagnostic odds ratios were 39.6 (95%CI 13.1, 119), 66.9

(95%CI 19.7, 227.2) and 47.99 (95%CI 11.4, 201.7) at 24, 48 and 72 hours

respectively (Figure 4). The positive predictive value of a persistently

abnormal aEEG increased from 66% at 24 hours, to 85% at 48 hours and

89% at 72 hours, for predicting adverse outcomes (supplementary

informationweb extra table). We checked the accuracy of this predictive

ability; the accuracy was best at 48 hours (AUC 0.90 SE0.06) [supplementary

9

information]web extra figure 2]. Only 3 out of 27(11%) babies had a normal

outcome despite a persistently abnormal aEEG at 72 hours.

Heterogeneity and methodological quality of included studies

Overall, the risks of bias in the index test and, flow and timing were low (Table

2). In patient selection, the study by Hallberg et al(3) had a high risk of bias,

as they did not use standard patient selection criteria. They also used a non-

standard method of outcome assessment and measured only motor outcome.

The studies by Ancora et al(26) and Gucuyener et al(10) were limited by small

sample size. In Gucuyener study, had some of the infants had their outcome

was assessed at beforelow 12 months of age(10). However, majority of the

included studies had large sample size.

Five studies (3, 4, 10, 11, 25) used total body cooling and three studies (12,

26, 27) used selective head cooling with mild systemic hypothermia. One

study used both whole body cooling and selective head cooling(2). Regarding

the aEEG monitors, five studies used single channel Olympic monitors

(Olympic CFM 6000 Monitor or CFM 5330; Olympic Medical, Seattle, WA,

USA), one study used both Olympic and Brainz (BRM; Natus, Seattle, WA)

monitors, the others used multi – channel Nervus monitor (Viasys, Nicole

Biomedical, Madison, WI, USA), single channel Lectromed (Lectromed,

Letchworth, United Kingdom) and multi channel Moberg monitor (Moberg

Research, Inc, Ambler, PA). Despite the difference in devices, the aEEG

background was reported in structured manner in all the included studies,

using pattern classification (1) in 7 studies and voltage classification(5) in 2

10

studies. This standard classification minimized the bias. Overall, none of the

included studies had high risk for bias (Table 2).

Discussion

In this review, we systematically evaluated the predictive value of an

abnormal trace, acquired at 6 hours, 24 hours, 48 hours and 72 hours of age,

among infants who had therapeutic hypothermia for moderate or severe NE.

The data highlight two clinically important observations; (a) A persistently

severely abnormal aEEG background at 48 hours of age or beyond predicts

the adverse outcome (PPV 85% and DOR 67 at 48h). (b) At 6 hours of age,

the aEEG background in hypothermia treated infants has a good sensitivity

96% (95%CI 89 – 97%) but unfortunately, the specificity 39% (95% CI 32 –

45%) is quite poor.

A persistently abnormal aEEG between 6 and 24 hours of age has been

associated with adverse outcomes in the pre-cooling era (1, 2, 5). Our data

suggest that therapeutic hypothermia has shifted this time window, and

adverse outcome were seen only if a severely abnormal aEEG was persistent

for at least 48 hours. This shift in the prognostic accuracy could be explained,

at least partly by the neuroprotective effects of therapeutic hypothermia.

Clinicians often use persistent aEEG background abnormalities to re-direct life

sustaining therapy and to counselling parents about long-term adverse

11

neurological outcome. This alteration in the prognostic accuracy of aEEG with

therapeutic hypothermia needs to be considered in such decision-making.

Abnormal six-hour aEEG was a mandatory inclusion criterion for some cooling

trials (23, 27, 28), although the incremental benefits of aEEG over a

structured neurological examination remains unclear. Our data suggest that

the positive predictive value of an abnormal 6-hour aEEG is poor in cooled

babies, and is lower than previous studies in the precooling era (29, 30).

Again, this observation could be explained by the beneficial effect of

therapeutic hypothermia. This finding might have important implications if

aEEG is used as an inclusion criterion in future clinical trials of adjunct cooling

therapies. Our findings are in agreement with the study of Sarkar et al. (19)

who showed that an abnormal 6-hour aEEG does not always correlate with

early death or abnormal MRI brain scan. Although, our study underlines that

6-hour aEEG has a good sensitivity and negative predictive value there is still

a seven percent of babies with a poor outcome despite a normal aEEG.

Besides, caution needs to be exercised in withholding cooling therapy based

on a normal 6-hour aEEG alone.

We analyzed the studies, which reported the aEEG background based on

voltage classification, separately. The accuracy of an abnormal aEEG at 6

hours to predict adverse outcome (web extra figure 1) did not differ from the

one earlier mentioned based on pattern classification. The pooled sensitivity

was 87% (95% CI 80 – 92%) and specificity was 36% (95% CI 29 – 44%).

However, the heterogeneity among the studies was very high (I2 > 80%). This

12

was difficult to explain; nevertheless, it could be due to the fact that aEEG

background was used as an inclusion criterion for some of the included

studies.

Three previous systematic reviews have examined the prognostic accuracy of

aEEG in neonatal encephalopathy; however, our study is the only meta-

analysis focussing specifically on cooled encephalopathic infants at multiple

time points. Spitzmiller et al, 2007 reported high prognostic accuracy

(sensitivity (91%) and specificity (88%)) of aEEG in neonatal encephalopathy,

but their work included only normothermic babies (31). van Laerhoven et al.,

2013, also reported high prognostic accuracy of aEEG (sensitivity and

specificity >90%), but again this review focussed on normothermic infants.

(32). Finally, Awal et al., 2015 (33) reported high prognostic accuracy of

aEEG in neonatal encephalopathy, but they included both cooled and

normothermic infants, and included studies with short-term outcome (3

months). Furthermore, the authors combined EEG and aEEG data, and did

not specifically examine the prognostic accuracy of aEEG.

There are some limitations to our review. Although, we analysed studies

based on the description of the aEEG background (pattern recognition or

voltage) separately, we projected the data from the studies by pattern

recognition only as there was significant heterogeneity among the studies

based on voltage classification when we combined the data (I2 > 80%).

Perhaps, some may argue that pattern recognition is more subjective, is

considered more sensitive than voltage measurements in predicting outcome

13

(29, 34). Secondly, the TOBY(4) and Coolcap studies(27) which were

included in the analysis, required the presence of an abnormal aEEG for

eligibility. However, this may not have influenced the e positive predictive

value could not to be affected.. Finally, we have not examined other facets of

an aEEG like sleep-wake cycles, time to recovery, seizures and use of

sedatives and its effects on outcome. The data available on such factors were

limited. Moreover, the quality of the included studies may have been affected

by the other factors like varied cooling equipment, different method of cooling

between the studies, small sample size in couple of studies and different

centres ‘policy about redirecting care. However, the quadas assessment did

not reveal any major risk of bias in the included studies. Besides, the mortality

rate in the different studies is in keeping with that reported in literature (up to

30%) (35) which makes unlikely that withdrawal of intensive support might

have affected our results.

Conclusion

Therapeutic hypothermia alters the predictive value of an aEEG background

at all-time points during induction and maintenance phase of hypothermia.

The predictive ability of an aEEG background at 6 hours of age during

induction of TH is sub-optimal, and there may be little benefit in using an

abnormal aEEG as an inclusion criteria in future clinical trials of adjunct

neuroprotective therapies. Persistently abnormal trace at 48 hours of age or

more was associated with adverse long-term outcome, and may be useful as

an adjunct bedside investigation for clinical decision making about limitation of

life sustaining therapy.

14

Funding – None

Competing financial interest – None

15

List of abbreviations:

aEEG – Amplitude integrated Electroencephalography

EEG – Electroencephalography

CNV – Continuous Normal Voltage

DNV – Discontinuous Normal Voltage

BS – Burst Suppression

FT – Flat Trace

CLV – Continuous Low Voltage

BSID – Bayley Scales of Infant Development

NE – Neonatal Encephalopathy

MeSH – Medical Subject Heading

16

References:

1. Hellstrom-Westas L, Rosen I, Svenningsen NW. Predictive value of early

continuous amplitude integrated EEG recordings on outcome after severe birth

asphyxia in full term infants. Archives of disease in childhood Fetal and neonatal

edition. 1995 Jan;72(1):F34-8. PubMed PMID: 7743282. Pubmed Central PMCID:

2528413.

2. Thoresen M, Hellstrom-Westas L, Liu X, de Vries LS. Effect of hypothermia

on amplitude-integrated electroencephalogram in infants with asphyxia. Pediatrics.

2010 Jul;126(1):e131-9. PubMed PMID: 20566612.

3. Hallberg B, Grossmann K, Bartocci M, Blennow M. The prognostic value of

early aEEG in asphyxiated infants undergoing systemic hypothermia treatment. Acta

paediatrica. 2010 Apr;99(4):531-6. PubMed PMID: 20050830.

4. Azzopardi D, group Ts. Predictive value of the amplitude integrated EEG in

infants with hypoxic ischaemic encephalopathy: data from a randomised trial of

therapeutic hypothermia. Archives of disease in childhood Fetal and neonatal edition.

2014 Jan;99(1):F80-2. PubMed PMID: 23800393. Pubmed Central PMCID: 3888630.

5. al Naqeeb N, Edwards AD, Cowan FM, Azzopardi D. Assessment of neonatal

encephalopathy by amplitude-integrated electroencephalography. Pediatrics. 1999

Jun;103(6 Pt 1):1263-71. PubMed PMID: 10353940.

6. de Vries LS, Hellstrom-Westas L. Role of cerebral function monitoring in the

newborn. Archives of disease in childhood Fetal and neonatal edition. 2005

May;90(3):F201-7. PubMed PMID: 15846008. Pubmed Central PMCID: 1721888.

7. Snyder P, Eason JM, Philibert D, Ridgway A, McCaughey T. Concurrent

validity and reliability of the Alberta Infant Motor Scale in infants at dual risk for

17

motor delays. Physical & occupational therapy in pediatrics. 2008;28(3):267-82.

PubMed PMID: 19064460.

8. Hagberg B, Hagberg G, Olow I. The changing panorama of cerebral palsy in

Sweden. IV. Epidemiological trends 1959-78. Acta paediatrica Scandinavica. 1984

Jul;73(4):433-40. PubMed PMID: 6464731.

9. Gorter JW, Ketelaar M, Rosenbaum P, Helders PJ, Palisano R. Use of the

GMFCS in infants with CP: the need for reclassification at age 2 years or older.

Developmental medicine and child neurology. 2009 Jan;51(1):46-52. PubMed PMID:

19018834.

10. Gucuyener K, Beken S, Ergenekon E, Soysal S, Hirfanoglu I, Turan O, et al.

Use of amplitude-integrated electroencephalography (aEEG) and near infrared

spectroscopy findings in neonates with asphyxia during selective head cooling. Brain

& development. 2012 Apr;34(4):280-6. PubMed PMID: 21741190.

11. Cseko AJ, Bango M, Lakatos P, Kardasi J, Pusztai L, Szabo M. Accuracy of

amplitude-integrated electroencephalography in the prediction of neurodevelopmental

outcome in asphyxiated infants receiving hypothermia treatment. Acta paediatrica.

2013 Jul;102(7):707-11. PubMed PMID: 23586497.

12. Lemmers PM, Zwanenburg RJ, Benders MJ, de Vries LS, Groenendaal F, van

Bel F, et al. Cerebral oxygenation and brain activity after perinatal asphyxia: does

hypothermia change their prognostic value? Pediatric research. 2013 Aug;74(2):180-

5. PubMed PMID: 23728382.

13. Amiel-Tison C. Neuromotor status. In: Taeusch HW, Yogman MW, editors.

Follow-up Management of the High-Risk Infant. Boston MA: Little, Brown and

Company; 1987. pp. 115–126.C. A-T. Textbook of neuromotor status. Boston: Little,

Brown & Company 1987.

18

14. Leemis LM Trivedi KS. A comparison of approximate interval estimators for

the Bernoulli parameter, The American Statistician 1996;50(1):63-68.

15. Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PM. The diagnostic odds

ratio: a single indicator of test performance. Journal of clinical epidemiology. 2003

Nov;56(11):1129-35. PubMed PMID: 14615004.

16. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et

al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy

studies. Annals of internal medicine. 2011 Oct 18;155(8):529-36. PubMed PMID:

22007046.

17. Padden B, Scheer I, Brotschi B, Wohlrab G, Latal B, Bernet V. Does

amplitude-integrated electroencephalogram background pattern correlate with

cerebral injury in neonates with hypoxic-ischaemic encephalopathy? Journal of

paediatrics and child health. 2015 Feb;51(2):180-5. PubMed PMID: 25123158.

18. Shellhaas RA, Thelen BJ, Bapuraj JR, Burns JW, Swenson AW, Christensen

MK, et al. Limited short-term prognostic utility of cerebral NIRS during neonatal

therapeutic hypothermia. Neurology. 2013 Jul 16;81(3):249-55. PubMed PMID:

23771483. Pubmed Central PMCID: 3770165.

19. Sarkar S, Barks JD, Donn SM. Should amplitude-integrated

electroencephalography be used to identify infants suitable for hypothermic

neuroprotection? Journal of perinatology : official journal of the California Perinatal

Association. 2008 Feb;28(2):117-22. PubMed PMID: 18004390.

20. Shah DK, Wusthoff CJ, Clarke P, Wyatt JS, Ramaiah SM, Dias RJ, et al.

Electrographic seizures are associated with brain injury in newborns undergoing

therapeutic hypothermia. Archives of disease in childhood Fetal and neonatal edition.

2014 May;99(3):F219-24. PubMed PMID: 24443407.

19

21. Massaro AN, Tsuchida T, Kadom N, El-Dib M, Glass P, Baumgart S, et al.

aEEG evolution during therapeutic hypothermia and prediction of NICU outcome in

encephalopathic neonates. Neonatology. 2012;102(3):197-202. PubMed PMID:

22796967.

22. Wyatt JS, Gluckman PD, Liu PY, Azzopardi D, Ballard R, Edwards AD, et al.

Determinants of outcomes after head cooling for neonatal encephalopathy. Pediatrics.

2007 May;119(5):912-21. PubMed PMID: 17473091.

23. Simbruner G, Mittal RA, Rohlmann F, Muche R, neo.n EnTP. Systemic

hypothermia after neonatal encephalopathy: outcomes of neo.nEURO.network RCT.

Pediatrics. 2010 Oct;126(4):e771-8. PubMed PMID: 20855387.

24. Takenouchi T, Rubens EO, Yap VL, Ross G, Engel M, Perlman JM. Delayed

onset of sleep-wake cycling with favorable outcome in hypothermic-treated neonates

with encephalopathy. The Journal of pediatrics. 2011 Aug;159(2):232-7. PubMed

PMID: 21353680.

25. Shankaran S, Pappas A, McDonald SA, Laptook AR, Bara R, Ehrenkranz RA,

et al. Predictive value of an early amplitude integrated electroencephalogram and

neurologic examination. Pediatrics. 2011 Jul;128(1):e112-20. PubMed PMID:

21669899. Pubmed Central PMCID: 3124102.

26. Ancora G, Maranella E, Grandi S, Sbravati F, Coccolini E, Savini S, et al.

Early predictors of short term neurodevelopmental outcome in asphyxiated cooled

infants. A combined brain amplitude integrated electroencephalography and near

infrared spectroscopy study. Brain & development. 2013 Jan;35(1):26-31. PubMed

PMID: 22082686.

27. Gluckman PD, Wyatt JS, Azzopardi D, Ballard R, Edwards AD, Ferriero DM,

et al. Selective head cooling with mild systemic hypothermia after neonatal

20

encephalopathy: multicentre randomised trial. Lancet. 2005 Feb 19-

25;365(9460):663-70. PubMed PMID: 15721471.

28. Azzopardi DV, Strohm B, Edwards AD, Dyet L, Halliday HL, Juszczak E, et

al. Moderate hypothermia to treat perinatal asphyxial encephalopathy. The New

England journal of medicine. 2009 Oct 1;361(14):1349-58. PubMed PMID:

19797281.

29. Shany E, Goldstein E, Khvatskin S, Friger MD, Heiman N, Goldstein M, et al.

Predictive value of amplitude-integrated electroencephalography pattern and voltage

in asphyxiated term infants. Pediatric neurology. 2006 Nov;35(5):335-42. PubMed

PMID: 17074604.

30. Toet MC, Hellstrom-Westas L, Groenendaal F, Eken P, de Vries LS.

Amplitude integrated EEG 3 and 6 hours after birth in full term neonates with

hypoxic-ischaemic encephalopathy. Archives of disease in childhood Fetal and

neonatal edition. 1999 Jul;81(1):F19-23. PubMed PMID: 10375357. Pubmed Central

PMCID: 1720950.

31. Spitzmiller RE, Phillips T, Meinzen-Derr J, Hoath SB. Amplitude-integrated

EEG is useful in predicting neurodevelopmental outcome in full-term infants with

hypoxic-ischemic encephalopathy: a meta-analysis. Journal of child neurology. 2007

Sep;22(9):1069-78. PubMed PMID: 17890403.

32. van Laerhoven H, de Haan TR, Offringa M, Post B, van der Lee JH.

Prognostic tests in term neonates with hypoxic-ischemic encephalopathy: a systematic

review. Pediatrics. 2013 Jan;131(1):88-98. PubMed PMID: 23248219.

33. Awal MA, Lai MM, Azemi G, Boashash B, Colditz PB. EEG background

features that predict outcome in term neonates with hypoxic ischaemic

encephalopathy: A structured review. Clinical neurophysiology : official journal of

21

the International Federation of Clinical Neurophysiology. 2015 May 31. PubMed

PMID: 26105684.

34. van Rooij LG, Toet MC, Osredkar D, van Huffelen AC, Groenendaal F, de

Vries LS. Recovery of amplitude integrated electroencephalographic background

patterns within 24 hours of perinatal asphyxia. Archives of disease in childhood Fetal

and neonatal edition. 2005 May;90(3):F245-51. PubMed PMID: 15846017. Pubmed

Central PMCID: 1721875.

35. Jacobs SE, Berg M, Hunt R, Tarnow-Mordi WO, Inder TE, Davis PG. Cooling

for newborns with hypoxic ischaemic encephalopathy. The Cochrane database of

systematic reviews. 2013;1:Cd003311. PubMed PMID: 23440789. Epub 2013/02/27.

eng.

22

Figure Legends :

Figure 1: Flowchart indicating search and selection of studies.

Figure 2: Forest plot showing the sensitivity and specificity of aEEG at 6 hours

and 24 hours of age in infants with NE including individual data from all

included studies.

Figure 3: Forest plot showing the sensitivity and specificity of aEEG at 48

hours and 72 hours of age in infants with NE including individual data from all

included studies.

Figure 4: Forest plot showing the diagnostic odds ratio of aEEG at 6, 24, 48

and 72 hours of age in infants with NE including individual data from all

included studies.

Web extraSupplemental figure 1: Forest plot showing the sensitivity and

specificity of aEEG at 6 hours in infants with NE including individual data from

all included studies based on voltage classification

Web extraSupplemental figure 2: The prognostic accuracy of aEEG from all

studies based on Pattern classification at 6, 24, 28, 72 hours of age. AUC –

Area Under the Curve.

23

24