PEDIATRICS Vol. 118 No. 6 December 2006, pp. 2552-2554 (doi:10.1542/peds.2006-2519)
EXPERIENCE & REASON |
Artifacts on Electroencephalograms May Influence the Amplitude-Integrated EEG Classification: A Qualitative Analysis in Neonatal Encephalopathy
a Centre of Perinatal Brain Research, Institute of Women’s Health, University College London, London, United Kingdom
b Department of Paediatrics, Hammersmith Campus, Imperial College London, London, United Kingdom
ABSTRACT
This is a case report and a descriptive study demonstrating that artifacts are common during long-term recording of amplitude-integrated electroencephalograms and may lead to erroneous classification of the amplitude-integrated electroencephalogram trace. Artifacts occurred in 12% of 200 hours of recording time sampled from a representative sample of 20 infants with neonatal encephalopathy. Artifacts derived from electrical or movement interference occurred with similar frequency; both types of artifacts influenced the voltage and width of the amplitude-integrated electroencephalogram band. This is important knowledge especially if amplitude-integrated electroencephalogram is used as a selection tool for neuroprotection intervention studies.
Key Words: amplitude-integrated EEG • neonatal encephalopathy • neuroprotection
Abbreviations: EEG, electroencephalogram • aEEG, amplitude-integrated electroencephalogram • CFM, cerebral function monitor
Amplitude-integrated electroencephalograms (aEEGs) provide a continuous, on-line trend recording of cerebral activity at the bedside and are increasingly used in the NICU.1 aEEG studies have been performed in term infants with neonatal encephalopathy, and prediction of neurodevelopmental outcome is possible as early as 3 to 6 hours after birth.2–4 Indeed, the aEEG is currently used as a screening tool for selection of infants for randomized, controlled trials of mild hypothermia.4,5
The aEEG is derived from the underlying raw EEG, which is amplified and passed through an asymmetrical band-pass filter to minimize artifacts from sweating, muscle activity, and electrocardiographic and other electrical interference. With first-generation aEEGs (cerebral function monitor [CFM], Lectromed, Herts, United Kingdom), the underlying raw EEG could not be inspected. New digital monitors simultaneously display and record EEGs and aEEGs (BRM 2 brain monitor [Brainz Instruments Ltd, Auckland, New Zealand] and Olympic CFM 6000 [Olympic Medical Corporation, Seattle, WA]); direct inspection of the underlying raw EEG and comparison with the aEEG can be performed in real time and retrospectively. As experience with these new digital monitors grows, there is a realization that this dual facility is crucial for the correct interpretation of the aEEG. In particular, we and others have recognized that EEG artifacts occur frequently despite the filtering process and can influence the aEEG recording significantly.
CASE REPORT
A mother presented to the hospital in the second stage of labor. Her fetus was breech, and a buttock was visible on admission. The infant was delivered instrumentally with difficulty and was assigned Apgar scores of 0 and 1 at 5 and 10 minutes, respectively. The cord gas had a pH of 6.8, PCO2 of 14.5 kPa, and a base excess of –16.5 mmol/L. The infant required intubation, chest compressions, intravenous adrenaline and bicarbonate, and a saline bolus. By 13 minutes of age the infant's heart rate was detected, and by 19 minutes a heart rate >100 beats per minute was noted, with occasional gasping. The infant was transferred to the neonatal unit while ventilated and was noted to be encephalopathic. An aEEG recording was performed at 3 hours of age to assess whether the infant fulfilled entry criteria for the Total Body Hypothermia (TOBY) trial. The initial recording (the first 30 minutes) showed a lower and upper margin within normal limits (both >5–10 µV), with no detectable seizure activity (Fig 1) despite a markedly abnormal neurologic examination (floppy, comatose, absent reflexes, no spontaneous movements). On inspecting the raw EEG, a rhythmic regular artifact was seen (Fig 1A). The electrodes were repositioned closer to the vertex, which led to a downward shift of the baseline (Fig 1B). Despite this maneuver, artifacts were still observed on the underlying raw EEG. The infant was not making any obvious spontaneous movements. The artifacts were interpreted as being caused by either electrical interference or muscle activity. After administration of a muscle relaxant (pancuronium), the regular rhythmic artifacts disappeared on the raw EEG trace, and the aEEG background became suppressed, with a background voltage of both upper and lower margins <5 µV (Fig 1C). At this point, the aEEG trace fulfilled criteria for entry into the TOBY trial. The infant developed Sarnat stage II hypoxic ischemic encephalopathy with clinical and electrographical seizure activity at the age of 48 hours.
|
3 hours of age. The arrows define the points at which each underlying raw EEG was inspected. A, High-frequency artifacts can be seen on the raw EEG. The aEEG baseline is shifted upward >5 µV. B, The electrodes were moved closer to the vertex. The underlying raw EEG continued to show artifacts of lower amplitude, and there was a slight downward shift of the aEEG baseline. C, After administration of pancuronium (medication marker), the underlying raw EEG showed fewer artifacts and was discontinuous. The aEEG background was classified as severely abnormal.DISCUSSION
This case demonstrates clearly that artifacts may lead to erroneous classification of the aEEG trace. To assess the incidence of artifacts on the underlying raw single-channel EEG and how they can influence the aEEG classification, we randomly selected 200 hours of aEEG recordings from a representative sample of 20 infants with neonatal encephalopathy and retrospectively inspected the underlying raw single-channel EEG. We observed that artifacts occurred in 12% of the recording time sampled (Fig 2); 55% of the artifacts were derived from electrical interference and 45% from movement interference, both of which could influence the voltage and width of the aEEG band.
|
The currently used aEEG classifications2,4 rely on the aEEG band voltage and background activity; thus, artifacts may result in an erroneous classification. This is important, especially if an aEEG is used as a selection tool for neuroprotection intervention studies, because, for example, elevation of the voltage band of activity from the baseline of <5 to >5 µV may suggest a more favorable prognosis and exclusion from the trial. The possibility of inspecting the raw EEG in the new digital aEEG systems facilitates the detection of artifacts and improves the accuracy of interpretation of the aEEG. However, although some artifacts may be easily recognized (eg, the regular artifact caused by electrocardiographic interference), in many cases the recognition of an underlying artifact requires an ability to interpret the EEG. Currently, there is much discussion on how to promote brain monitoring by aEEG in clinical practice, and the first International Conference on Neonatal Brain Monitoring on aEEG took place recently (www.cme.hsc.usf.edu/coph/brn). Companies marketing the new-generation CFMs usually provide brief training material, references to publications, and support brief training courses. However, more detailed training and supervision by skilled personnel are required if the intention is to perform long-term monitoring during intensive care and identify seizures.
It is essential to scrutinize the underlying raw EEG for artifacts at the commencement of a recording or when there is a sudden change in appearance of the trace. Skilled personnel should review the aEEG and EEG if the aEEG is to be used to direct clinical care and as a selection tool for neuroprotective studies.
FOOTNOTES
Accepted Aug 30, 2006.
Address correspondence to Nicola J. Robertson, PhD, FRCPCH, University College London, Department of Obstetrics and Gynaecology, 86–96 Chenies Mews, London WC1E 6HX, United Kingdom. E-mail: n.robertson{at}ucl.ac.uk
The authors have indicated they have no financial relationships relevant to this article to disclose.
REFERENCES
- de Vries LS, Hellstrom-Westas L. Role of cerebral function monitoring in the newborn. Arch Dis Child Fetal Neonatal Ed. 2005;90 :F201—F207
- 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. Arch Dis Child Fetal Neonatal Ed. 1995;72 :F34—F38
- 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.
Arch Dis Child Fetal Neonatal Ed. 1999;81
:F19
–F23
[Abstract/Free Full Text] - al Naqeeb N, Edwards AD, Cowan FM, Azzopardi D. Assessment of neonatal encephalopathy by amplitude-integrated electroencephalography.
Pediatrics. 1999;103
:1263
–1271
[Abstract/Free Full Text] - Gluckman PD, Wyatt JS, Azzopardi D, et al. Selective head cooling with mild systemic hypothermia after neonatal encephalopathy: multicentre randomised trial. Lancet. 2005;365 :663 –670[Web of Science][Medline]
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Suk, A. N. Krauss, M. Engel, and J. M. Perlman Amplitude-Integrated Electroencephalography in the NICU: Frequent Artifacts in Premature Infants May Limit Its Utility as a Monitoring Device Pediatrics, February 1, 2009; 123(2): e328 - e332. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
