Article Text

Download PDFPDF

Cooling and seizure burden in term neonates: an observational study
  1. Evonne Low1,
  2. Geraldine B Boylan1,
  3. Sean R Mathieson2,
  4. Deirdre M Murray1,
  5. Irina Korotchikova1,
  6. Nathan J Stevenson1,
  7. Vicki Livingstone1,
  8. Janet M Rennie2
  1. 1Neonatal Brain Research Group, Department of Paediatrics and Child Health, University College Cork, Cork, Ireland
  2. 2Elizabeth Garrett Anderson Institute for Women's Health, University College London Hospitals, London, UK
  1. Correspondence to Geraldine B Boylan, Professor of Neonatal Physiology, Department of Paediatrics & Child Health, Clinical Investigations Unit, Cork University Hospital Wilton, Cork, Ireland; g.boylan{at}


Objective To investigate any possible effect of cooling on seizure burden, the authors quantified the recorded electrographic seizure burden based on multichannel video-EEG recordings in term neonates with hypoxic-ischaemic encephalopathy (HIE) who received cooling and in those who did not.

Study design Retrospective observational study.

Patients Neonates >37 weeks gestation born between 2003 and 2010 in two hospitals.

Methods Off-line analysis of prolonged continuous multichannel video-EEG recordings was performed independently by two experienced encephalographers. Comparison between the recorded electrographic seizure burden in non-cooled and cooled neonates was assessed. Data were treated as non-parametric and expressed as medians with interquartile ranges (IQR).

Results One hundred and seven neonates with HIE underwent prolonged continuous multichannel EEG monitoring. Thirty-seven neonates had electrographic seizures, of whom 31 had EEG recordings that were suitable for the analysis (16 non-cooled and 15 cooled). Compared with non-cooled neonates, multichannel EEG monitoring commenced at an earlier postnatal age in cooled neonates (6 (3–9) vs 15 (5–20) h)and continued for longer (88 (75–101) vs 55 (41–60) h). Despite this increased opportunity to capture seizures in cooled neonates, the recorded electrographic seizure burden in the cooled group was significantly lower than in the non-cooled group (60 (39–224) vs 203 (141–406) min). Further exploratory analysis showed that the recorded electrographic seizure burden was only significantly reduced in cooled neonates with moderate HIE (49 (26–89) vs 162 (97–262) min).

Conclusions A decreased seizure burden was seen in neonates with moderate HIE who received cooling. This finding may explain some of the therapeutic benefits of cooling seen in term neonates with moderate HIE.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Approximately, 50–75% of neonatal seizures at term are attributable to neonatal hypoxic-ischaemic encephalopathy (HIE).1 Neonatal outcome studies have shown that seizures are powerful predictors of death or permanent neurological disability.2 ,3 However, these studies relied almost entirely on the detection of seizures using clinical criteria2 or amplitude-integrated (aEEG).3 It is well known that clinical assessment4 and aEEG5,,7 can miss many seizures and therefore cannot accurately quantify the precise seizure burden in neonates. Accurate identification and quantification of neonatal seizures require continuous multichannel video-EEG monitoring.

What this study adds

  • The seizure burden was less in a group of neonates treated with therapeutic hypothermia compared with a similar group who were not cooled.

  • This is the first study using prolonged continuous multichannel EEG to quantify the seizure burden in non-cooled and cooled term neonates with hypoxic-ischaemic encephalopathy.

What is already known on this topic

  • Cooling has been shown to reduce the combined rate of death and disability at 18 months of age in term neonates with hypoxic-ischaemic encephalopathy.

  • Early, prolonged and continuous multichannel EEG provides accurate identification and quantification of electrographic seizure burden in term neonates.

The evidence of benefit is considered sufficient for the National Institute for Health and Clinical Excellence to endorse the use of cooling for hypoxic perinatal brain injury in the UK.8 A meta-analysis of three trials which enrolled 767 neonates showed that cooling reduced the combined rate of death or disability at 18 months.7 However, the precise mechanism by which cooling achieves neuroprotection in neonates with HIE is unknown. In the biphasic model of neuronal death following hypoxic injury, the cascade of events which occurs in the secondary reperfusion phase may be associated with seizures, an accumulation of cytotoxins and the failure of oxidative cerebral metabolism.9 ,10 Cooling may reduce seizure burden in neonates by affecting some mechanisms during this vital phase of brain injury.

We aimed to determine whether cooling was associated with a reduction in seizure burden in HIE neonates by quantifying seizure burden using continuous multichannel video-EEG monitoring.

Patient and methods

Non-cooled neonates were enrolled between June 2003 to September 2006 and January 2009 to March 2010 from Cork University Maternity Hospital (CUMH), Ireland. Cooled neonates were enrolled between January 2009 and September 2010 from CUMH and University College London Hospitals (UCLH), UK. Neonates >37 weeks gestation with HIE were enrolled for EEG monitoring if they fulfilled ≥2 of the following criteria: Apgar score <6 at 5 min, a continued need for resuscitation after birth, clinical evidence of encephalopathy or seizures within 24 h of birth. At both hospitals, every neonate was assigned a clinical grade of encephalopathy using the modified Sarnat score at 24 h of age.11This study was conducted with the approval from the Clinical Research Ethics Committees of the Cork Teaching hospitals, Ireland, and the National Health Service in the UK, via the Integrated Research Application Service. Written, informed consent was obtained from at least one parent of each neonate who participated in this study.

Neonates were cooled according to the entry criteria and guidelines set by the UK Total Body Hypothermia for Neonatal Encephalopathy (TOBY) cooling registry (from the UK TOBY Cooling Register Clinician's Handbook, section 2.1, Either the Tecotherm TS med 200 (Tec-Com, Halle, Germany) or the CritiCool MTRE machine (Charter Kontron, Milton Keynes, UK) was used. Neonates were cooled to a rectal temperature of 33–34°C for 72 h (unless contraindicated) and were slowly rewarmed. Within both hospitals, treatment was based on clinical observation and EEG findings. Both groups had continuous EEG monitoring but the neonatologists did not interpret the EEG recordings. On our monitoring system, the aEEG and the multichannel EEG were simultaneously recorded and many of our neonatologists would have used the aEEG as an aid to clinical decision-making. All clinical seizures were treated. The aEEG was used to confirm clinically suspected seizures. If they were concerned about any abnormal clinical behaviours or aEEG patterns, the encephalographers would be asked to interpret the multichannel EEG at a later stage. Immediate reporting of the multichannel EEG was not available, so that aEEG and clinical suspicion were the mainstays of seizure diagnosis. Phenobarbitone was the first-line anticonvulsant administered to a maximum dose of 40 mg/kg intravenously. Second-line anticonvulsants were administered if clinical and/or electrographic seizures recurred following phenobarbitone administration. In both hospitals, second-line anticonvulsant was either intravenous phenytoin or midazolam. Although standardized protocols for the use of anticonvulsants were similar in both hospitals, the choice of second-line anticonvulsant administration was at the discretion of the attending clinician.12 ,13 The timing and dose of each anticonvulsant as well as morphine administered were recorded in all neonates.

Throughout the study, EEG recording methods were identical at both hospitals. A Nicolet monitor (CareFusion NeuroCare, Wisconsin, USA) was used to record multichannel video-EEG, using the 10-20 system of electrode placement modified for neonates.14 EEG monitoring was commenced as soon as possible after birth and continued for at least 20 h of artefact-free EEG. Scalp electrodes were placed at F3, F4, C3, C4, T3, T4, O1, O2 and Cz locations to record the EEG activity from the frontal, central, temporal and occipital areas. Parietal electrodes (P3 and P4) were also used wherever possible. Impedances of below 5 kΩ were maintained. The entire EEG recording from each neonate was independently reviewed by two experienced encephalographers (GBB and SRM). Cases of disagreement were resolved by consensus. An electrographic seizure was defined as a sudden and evolving repetitive stereotyped waveform with a definite start, middle and end, lasting for at least 10 s15 on at least one EEG channel. Status epilepticus was defined as continuous16 or accumulative17 electrographic seizure activity lasting ≥50% of each 1 h period. The recorded seizure burden was defined as the total duration of recorded electrographic seizures in minutes. Seizure number was counted as the number of seizure events recorded on the EEG. Mean seizure duration was calculated for all recorded electrographic seizures in each neonate.

Statistical analysis

Inter-rater agreement between the two encephalographers was assessed using a Cohen's κ statistic. Continuous variables were described using medians and interquartile ranges (IQR) and categorical variables using frequencies. For comparisons between the two groups (non-cooled and cooled), the Mann–Whitney test was used for continuous variables and the χ2 test or Fisher's exact test (in the case of small expected counts) was used for categorical variables. All statistical analyses were performed using PASW Statistics 17.0. All tests were two-sided and a p value <0.05 was considered to be statistically significant.


During the study, 107 neonates were diagnosed with HIE (figure 1). The clinical Sarnat grade for HIE was assigned as mild in 43, moderate in 34 and severe in 30 neonates. Among the 64 neonates with moderate or severe HIE, electrographic seizures were recorded in 37 neonates. Of these, six neonates were excluded from the study analysis. Four neonates with moderate HIE were excluded: two cooled neonates had secondary events shortly after EEG was commenced (one with cardiopulmonary arrest and the other with pulmonary haemorrhage), one cooled and one non-cooled neonate had less than 20 h of artefact-free EEG. Two neonates with severe HIE were excluded: one cooled neonate with a subsequent principal diagnosis of mitochondrial respiratory chain disease and one non-cooled neonate with less than 20 h of artefact-free EEG. The remaining 31 neonates formed our study group (16 non-cooled and 15 cooled).Tables 13 summarize the clinical characteristics of neonates in both groups.

Figure 1

Flow diagram of study selection.

Table 1

Clinical characteristics of the neonates included in the study

Table 2

Individual characteristics of non-cooled neonates with hypoxia-ischaemic encephalopathy

Table 3

Individual characteristics of cooled neonates with hypoxia-ischaemic encephalopathy enrolled for this study

Eight of 16 non-cooled neonates and none of the cooled neonates received at least one dose of phenobarbitone before EEG monitoring commenced. However, there was no significant difference in the number of anticonvulsants received between the two groups (non-cooled: 2 (1-3) vs cooled: 1 (1-2); p=0.274) and in the total administered dose of first-line anticonvulsant (non-cooled: 30 (20-40) vs cooled: 20 (20-20) mg/kg; p=0.203). There was also no significant difference in the ages at which the first-line anticonvulsant (non-cooled: 12 (9-19) vs cooled: 14 (10-24) h; p=0.504) and the second-line anticonvulsant were administered (non-cooled: 28 (24-31) vs cooled: 26 (19-38) h; p=0.556). All cooled neonates received morphine compared with eight non-cooled neonates (p=0.002 from Fisher's exact test).

Cooling commenced at the median (IQR) age of 5 (2-6) h (table 3). In six of seven cooled neonates with severe HIE, cooling was commenced within 6 h of age. However, following decisions to withdraw life-sustaining support in five of these seven neonates, the duration of cooling and EEG monitoring were shorter. The recorded seizure burden in these neonates was higher than in cooled neonates with moderate HIE. In three of eight cooled neonates with moderate HIE, passive cooling commenced earlier during the transport to UCLH, but the recorded age at which active cooling commenced was after 6 h. Despite this, all eight neonates with moderate HIE received cooling for at least 72 h.

The inter-rater agreement for seizure identification was consistent with a high level of agreement (κ=0.872). In eight non-cooled and one cooled neonate, seizures were ongoing when EEG recording commenced. The postnatal age of first recorded electrographic seizure was similar in both groups (non-cooled: 18 (12-22) vs cooled: 13 (11–22) h; p=0.252). The median recorded seizure burden was significantly less in the cooled than in the non-cooled group (cooled: 60 (39-224) vs non-cooled 203 (141-406) min; p=0.027) (table 4). Between the cooled and the non-cooled group, there was no difference in the number of seizure events, mean seizure duration or the presence of status epilepticus (p=0.105, 0.192 and 0.095, respectively). An exploratory subgroup analysis was performed to assess the influence of cooling on neonates with different severity of encephalopathy in non-cooled and cooled groups. Cooling had a significant reduction of recorded seizure burden in neonates with moderate HIE (non-cooled: 162 (97–262) vs cooled: 49 (26–89) min; p=0.020) while no such difference was seen in neonates with severe HIE (non-cooled: 223 (172–720) vs cooled: 224 (60–289) min; p=0.558).

Table 4

Characteristics of seizure burden in non-cooled and cooled groups

Eleven cooled neonates had EEG monitoring after cooling was discontinued. Electrographic seizures were observed in 4 of 15 cooled neonates when cooling was discontinued (table 3). Two of the four cases had shorter duration of cooling when a decision was made to withdraw life-sustaining support (case L8 cooled for 19 h, case L7 cooled for 23 h). In the remaining two cases (cases C22 and L5), electrographic seizures were observed following discontinuation of cooling despite the fact that cooling started at 6 and 9 h, respectively after birth and continued for 72 h.


We have shown that term neonates with moderate HIE treated with whole-body cooling have a significantly lower electrographic seizure burden when compared with non-cooled neonates. This is the first study to quantify and compare the recorded seizure burden between non-cooled and cooled neonates using early and prolonged continuous multichannel video-EEG.

Previously published neonatal hypothermia trials could not accurately measure seizure burden as their protocols did not include multichannel EEG monitoring. These studies used clinical18 and/or aEEG monitoring7 ,19 for seizure recognition. The Neonatal Research Network Whole-Body Hypothermia Trial relied on the clinical recognition of seizures and when the authors adjusted their study for cooling and severity of encephalopathy, cooling did not appear to have any impact on the frequency of clinical seizures and outcome.18 However, clinical estimation of seizure burden is notoriously unreliable, with the majority of neonatal seizures being subclinical.4 ,20 The TOBY trial used aEEG as a recruitment and monitoring tool during cooling in some participating neonatal institutions.21 At recruitment, clinical seizures and seizures detected by aEEG were present in 67% (74/110) and 29% (33/115) of neonates, respectively and seizures were considered as a complication during cooling, with a decreasing incidence from days 1 to 4 (75–23%). Clinical recognition of seizures and the aEEG is known to underestimate the true seizure burden. The aEEG cannot detect short seizures, seizures which do not generalize and low voltage seizures.6

In fetal sheep, cooling was associated with a marked reduction in the amplitude of seizures and epileptiform activities in the first 6 h after a complete umbilical cord occlusion.22 The duration of individual electrographic seizures was reduced in the cooled compared to the non-cooled asphyxiated piglets.23 Hypothermia to 30 or 33°C has been shown to completely inhibit the release of glutamate in a rat model of cerebral ischaemia.24 Other effects of cooling such as reduced cytotoxic oedema by reducing amino acid release25 and inhibition of free oxygen radicals26 may have an impact on the reduction in seizure burden.

The results of exploratory analysis showed that the recorded seizure burden was only significantly reduced in cooled neonates with moderate HIE. Possibly, this is related to the higher recorded seizure burden in five of seven cooled neonates with severe HIE who had shorter durations of cooling and EEG monitoring following decisions to withdraw life-sustaining support. Interestingly, the analysis of three neonatal hypothermia trials has revealed that the primary outcome of death and disability at 18 months was significantly reduced by cooling neonates with moderate but not severe HIE.7 However, Simbruner et al. has shown that cooling was strongly neuroprotective even in severe HIE.19 Therefore, it is important to emphasize that further data are required to clarify whether cooling is appropriate for severe HIE, before clinical decisions are made to abort cooling neonates with severe HIE.

This study has a retrospective design which may have led to some bias. Both cohorts were selected as they were encephalopathic and at high risk of developing seizures. Both groups had continuous multichannel EEG monitoring, and the standard protocol for monitoring was the same at both time points and in both hospitals. We did anticipate a potential bias relating to the choice of anticonvulsants used, as administration was at the discretion of different attending clinicians in both hospitals at that point in time. To date, there is still no consensus on a standard protocol for the use of anticonvulsants among neonatologists.12 ,13 The recorded seizure burden remained higher in non-cooled neonates, despite the fact that they received anticonvulsants earlier. However, there was no significant difference between the groups in the number, dose and age when the first- and second-line anticonvulsants were administered. One of the strengths of this study is that the same reviewers analyzed the EEG recordings of both cohorts using a standardized grading system. EEGs were also recorded in both groups on the same equipment. As we were using a historical cohort, over time some increase in the ability of the EEG annotators to recognize seizures through increasing use of early continuous EEG may have occurred. However, this should have increased the seizure burden in the more recent cooled group. This bias only serves to strengthen our findings.

Phenobarbitone remains the most commonly used first-line anticonvulsant in neonatal units worldwide.12 ,13 It has been shown to augment the neuroprotective effect of hypothermia.27 However, phenobarbitone has been shown to be ineffective in controlling electrographic seizures in non-cooled neonates;28 ,29 this reduced efficacy has been linked to the altered neuronal chloride transport in the developing brain.30 In our study, plasma phenobarbitone levels were not routinely measured. It is known that the half-life of phenobarbitone is significantly increased in cooled neonates and plasma drug levels will accumulate due to reduced hepatic metabolism during hypothermia.31 ,32 Sedative and anaesthetic medications have been shown to facilitate the therapeutic effects of cooling.23 All cooled neonates and half of non-cooled neonates in our study received morphine. However, morphine does not possess anticonvulsive properties and therefore cannot explain the measured difference in the recorded seizure burden between the two groups.

There was no significant difference in Apgar scores between both groups, although the pH was significantly lower in the cooled group. This may reflect more severe disease in the cooled group. However, we have previously shown that neither the condition at birth nor the degree of metabolic acidosis reliably predicts electrographic neonatal seizures33 and therefore we do not think that this had any influence on the seizure burden in the cooled group. All non-cooled neonates would have qualified for cooling if it was available at the time of recruitment. The time of onset and the duration of EEG recording between non-cooled and cooled groups were significantly different. Several non-cooled neonates were already experiencing seizures when EEG recording commenced; the recorded seizure burden in this group may have been underestimated. Despite this, and the fact that there was a longer EEG recording time which increased the possibility of capturing more seizures in the cooled group, the overall recorded seizure burden was still lower in the cooled group.

Two recent hypothermia studies have not quantified the recorded seizure burden and a control cohort was not made available for comparison.34 ,35 In another study, seizures occurred less frequently in the cooled group but this was not significantly different with a control cohort.36 It would no longer be ethical to randomize HIE neonates to normothermia. For this reason, although our study sample size is small, our non-cooled HIE cohort with prolonged continuous EEG monitoring is unlikely to be replicated.

In summary, we found that cooling was associated with a decreased electrographic seizure burden in neonates with HIE. A reduced seizure burden may lead to a reduction in neuronal damage, and may help explain the observed improvement in long-term neurodevelopmental outcome in cooled neonates with moderate HIE. Further studies using prolonged continuous multichannel EEG monitoring are undoubtedly indicated.


The authors specially thank the medical and nursing staff from the neonatal intensive care units in CUMH, UCLH and the parents who gave permission for their babies to be studied.



  • Correction notice This article has been corrected since it was published Online First. The word electroencephalographers has been corrected to encephalographers and “prefix cases” has been deleted from the list of abbreviations in table 3.

  • Funding This study was funded by a translational award from the Wellcome Trust UK (85249/z/08/z). This work was also partly undertaken at the University London Hospitals/University College London who received a proportion of funding from the Department of Health's National Institute for Health Research and Biomedical Research Centers funding scheme. The views expressed in this publication are those of the authors and not necessarily those of the Department of Health in the UK.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval from the Clinical Research Ethics Committees of the Cork Teaching hospitals, Ireland and the National Health Service in the UK, via the Integrated Research Application Service.

  • Provenance and peer review Not commissioned; externally peer reviewed.