Quantitative electroencephalographic patterns in normal preterm infants over the first week after birth

https://doi.org/10.1016/j.earlhumdev.2005.07.009Get rights and content

Abstract

Background

Currently available tools to assist clinicians with prediction of neurodevelopmental outcome of preterm infants are inadequate. Modified cotside electroencephalography (EEG) has the ability to produce quantitative electrophysiologic measures. These measures may be useful in future prediction of outcome.

Aim

To determine patterns of change in quantitative EEG measures in preterm infants during their first week after birth.

Design

Observational.

Subjects

Preterm infants born at less than 32 weeks completed gestation surviving to discharge with unremarkable serial ultrasound scans.

Outcome measures

Changes in continuity, amplitude and spectral edge frequency measures of EEGs obtained over the first week after birth.

Results

Results of EEGs performed using a novel EEG device on 63 infants are reported here. Their median (range) gestation was 29 (24–31) weeks and birthweight was 1235 (540–1980) g. Quantitative measures of EEG continuity increased over the first week after birth from 72 (25–99)% to 92 (54–100)% at the 25 μV threshold, and from 39 (10–87)% to 64 (34–75)% at the 50 μV threshold, both p < 0.0001. There was a related 32% increase in median amplitude from 5.8 (2.6–10.6) μV on day 1 to 7.6 (4.3–9.4) μV on day 4, p = 0.005. There was a trend for average spectral edge frequency to fall from 10.7 (9.3–12.9) Hz on day 1 to 9.9 (8.1–12.3) Hz on day 3, p = 0.06. Each gestational tertile showed similar patterns.

Conclusions

There are consistent changes in quantitative neurophysiologic measures over the first week after birth, and particularly measures of continuity over the first 4 days, in normal preterm infants.

Introduction

Although the survival rates for sick preterm neonates have improved, the rate of long-term neurodevelopmental disability in survivors remains stable, resulting in increased numbers of survivors with disability [1]. Early prediction of long-term outcome would facilitate early counselling of parents and potentially allow selection for neuroprotective interventions if these become available. However, cranial ultrasound scan and neurological examination are suboptimal in early prediction [2], [3]. Early cranial ultrasound scans demonstrate intraparenchymal haemorrhage [4], [5] but white matter changes, if detected, may not be evident until later in the neonatal course [6], [7].

Electroencephalography (EEG) may also be used for prognostic purposes in preterm infants, with the possibility of predicting cognitive impairment as well as motor disability. Seizure activity on standard EEG has been associated with increased mortality [8], and positive Rolandic sharp waves occur more frequently in infants who subsequently have abnormal motor development [9]. However, there are some difficulties with the use of EEG for prognosis in preterm infants. Maturational changes have to be considered; continuity of the background EEG activity increases with gestational age, both around the time of delivery [10], [11], [12], [13] and sequentially after birth [13]. There is a decrease in interburst interval over the first 3 days after birth in preterm infants [14]. In addition, standard multi-channel EEGs are also often difficult to obtain in the NICU setting and techniques must be modified to maintain the integrity of the fragile preterm skin. Furthermore, the interpretation of preterm neonatal standard EEG is a specialised area and experts in this field are not widely available.

One modified system widely available in NICUs is the amplitude-integrated EEG (aEEG). It produces a single-channel, time-compressed output that has been used for prognostication in term infants with neonatal encephalopathy [15], [16]. There are also some data available for preterm infants regarding normal patterns, reference values and maturational changes of aEEG recordings [17], [18], [19]. However, there are no data for early, sequential recordings in extremely premature infants. EEGs were recorded over a 5 day period using another system, the Oxford Medilog recorder, to show gestational changes in continuity patterns [20]. However, there were insufficient extremely preterm infants in the cohort to analyse patterns over the first 5 days after birth. An automated system for analysing the continuity and amplitude of EEGs recorded using this system was developed [21] but has not been incorporated into routine NICU care.

We have investigated the utility of a novel cotside EEG device with quantitative neurophysiological analyses, including continuity, amplitude and spectral edge frequency (SEF), to assist with prediction of long-term neurodevelopmental outcome in preterm infants. However, it is essential to characterize the ‘normal’ temporal changes in these measures as a part of future development of predictive algorithms using this EEG modality. The aim of the current paper was to characterize the changes in quantitative EEG over the first week after birth in a cohort of premature infants with no ultrasound evidence of white matter changes or severe intraventricular haemorrhage during their NICU course.

Section snippets

Subjects

Infants born before 32 weeks gestation and admitted to National Women's Hospital, Auckland, New Zealand, between December 2002 and February 2004 were recruited to this study. Parents of eligible infants were approached either before delivery, or as soon as possible after delivery, about the study. Written informed parental consent was obtained. This study was approved by the Auckland Ethics Committee.

Protocol

EEGs were recorded four times during the first week after birth. Cranial ultrasounds were

Results

Between December 2002 and February 2004, 153 infants of less than 32 weeks gestation were admitted. Parents of 98 infants (64%) were approached regarding the study. Informed parental consent was obtained for 79 infants (81% of those approached), and declined for 19 infants (19%). The parents of 55 babies were not approached regarding the study: 2 infants were unable to be stabilised and died early; 5 parents/guardians were unavailable for consent; for 4 infants there were language barriers to

Discussion

This study is the first to examine changes in quantitative electrophysiological measures in preterm infants using a novel cotside EEG monitor over the first week after birth. Our results show that there are significant changes in quantitative measures over this period in a group of preterm infants selected on the basis of apparently normal cranial ultrasound scans.

Over the first week after birth there was a 63% increase in the continuity (at the 50 μV threshold) of the EEG traces. There was

Acknowledgments

We would like to thank the parents and infants who participated in this study: without them nothing would have been achieved. We would also like to acknowledge the nursing staff of National Women's NICU who made this study possible. We also acknowledge Dr. Michael Navakatikyan's contribution to the development of the continuity analysis algorithm. CRW was supported by a University of Auckland Senior Health Research Scholarship.

References (33)

  • M.I. Levene et al.

    Extension of neonatal intraventricular haemorrhage

    Arch Dis Childild

    (1984)
  • M. Wheater et al.

    Perinatal infection is an important risk factor for cerebral palsy in very-low-birthweight infants.[comment]

    Dev Med Child Neurol

    (2000)
  • M.S. Scher et al.

    Electrographic seizures in preterm and full-term neonates: clinical correlates, associated brain lesions, and risk for neurologic sequelae

    Pediatrics

    (1993)
  • M. Hayakawa et al.

    Background electroencephalographic (EEG) activities of very preterm infants born at less than 27 weeks gestation: a study on the degree of continuity

    Arch Dis Childild Fetal Neonatal Ed

    (2001)
  • M. van de Bor et al.

    Electrical brain activity in preterm infants at risk for intracranial hemorrhage

    Acta Paediatr

    (1994)
  • S. Victor et al.

    Spectral analysis of electroencephalography in premature newborn infants: normal ranges

    Pediatr Res

    (2005)
  • Cited by (74)

    • Probabilistic graphical model identifies clusters of EEG patterns in recordings from neonates

      2019, Clinical Neurophysiology
      Citation Excerpt :

      Victor et al. (Victor et al., 2005) used automated spectral analyses of EEG frequencies to quantitatively describe the EEG of preterm infants just after birth. West et al. (West et al., 2006) analyzed changes in continuity, amplitude, and spectral edge frequency by using automated spectral analysis. In the work of Tort et al. (Tort et al., 2010) scientists quantified the degree of phase-amplitude couplings between high frequency oscillation amplitudes with the phase shift of low frequencies.

    • Detecting bursts in the EEG of very and extremely premature infants using a multi-feature approach

      2017, Medical Engineering and Physics
      Citation Excerpt :

      Existing methods for detecting bursts in preterm EEG rely on either amplitude or frequency characteristics, or combinations of both [2,6,8,10–19]. Many of these methods, however, were not designed as stand-alone detection methods and have not been assessed with the gold standard, the EEG expert’s visual interpretation of the EEG [2,8,10,11,13,16]. For those methods with performance validation metrics, the more promising methods employ frequency-weighted energy measures, which multiply amplitude and frequency to estimate energy [6,17–19].

    • Electroencephalography in the Preterm and Term Infant

      2017, Fetal and Neonatal Physiology, 2-Volume Set
    View all citing articles on Scopus
    View full text