Elsevier

The Journal of Pediatrics

Volume 153, Issue 3, September 2008, Pages 369-374
The Journal of Pediatrics

Original article
Assessment of Neonatal Electroencephalography (EEG) Background by Conventional and Two Amplitude-Integrated EEG Classification Systems

These data were presented in abstract form at the American Epilepsy Society annual meeting, Philadelphia, PA, November 30-December 4, 2007.
https://doi.org/10.1016/j.jpeds.2008.03.004Get rights and content

Objective

To determine the agreement among conventional electroencephalography (CEEG) terminology background classification and a simple and an advanced amplitude-integrated EEG (aEEG) system, and to evaluate whether aEEG interpreter experience or electrographic seizures affect this agreement.

Study design

CEEG background was classified by traditional interpretive criteria for 144 neonatal recordings, from which a single channel was converted to aEEGs. These aEEGs were independently interpreted by neonatologists according to the simple and advanced classification systems.

Results

Interreader agreement was better with the simple aEEG system compared with the advanced aEEG system (multirater κ, 0.66 vs 0.44). Fair-to-moderate agreement was found between both of the aEEG classification systems and CEEG (simple: κ, 0.34 to 0.45; advanced: κ, 0.36 to 0.45). Agreement did not vary significantly based on the aEEG interpreter experience or the presence of seizures.

Conclusions

Neonatologists found better agreement using the simple aEEG system regardless of their expertise or the presence of seizures. This finding has implications for patient selection in future multicenter neonatal neuroprotection studies.

Section snippets

Methods

The Children's Hospital of Philadelphia's Institutional Review Board approved the design of this study. The study's initial goals were to determine the sensitivity of aEEG in recognizing electrographic seizures present on CEEG, the gold standard for seizure detection,9, 15 and to evaluate the agreement in background assessments among aEEG readers using the 2 different classification systems. A sample size of 125 CEEGs was determined to provide a confidence interval of ± 5% regarding the true

Results

A total of 144 neonatal CEEGs from 140 term or near-term neonates (conceptional age, 34 to 50 weeks) were analyzed. The CEEG duration ranged from 23 to 145 minutes. Details of these CEEGs have been described previously.9, 15 Among these CEEGs, 125 records contained seizures, and 19 did not. In those with seizures, the mean seizure rate was 7 seizures per hour (range, 0.5 to 21 seizures per hour). The CEEGs were read simultaneously by 2 of the authors (R.S. and R.C.), and the interictal

Discussion

CEEG has been used for decades to objectively assess the neurologic status of critically ill neonates and is predictive of high risk for subsequent electrographic seizures.7 CEEG has excellent prognostic value1, 2, 3, 4, 5, 6 and is considered the gold standard for electrographic seizure detection and quantification; however, its limited availability and the technical challenges of long-term EEG monitoring have led to the development of complementary types of cerebral function monitoring. aEEG

References (20)

There are more references available in the full text version of this article.

Cited by (65)

  • CARFS<sup>7</sup>: A guide and proforma for reading a preterm neonate's EEG

    2022, Neurophysiologie Clinique
    Citation Excerpt :

    One of the problems in this period is deciding what is normal – as preterm birth itself is not a normal event. There are excellent papers describing in detail the occurrence, maturation and evolution of different patterns (normal and abnormal) from the foetal to preterm to term babies [41,18,86,1,63–65,69,33,83,52,49,48,22–24,16,57,93,6,3,4,82]. Many of these studies illustrate the usefulness of EEG in predicting neurodevelopmental outcome in the term and preterm baby [14,64,23,33,48–51,18–22,16].

  • Quantitative approach to early neonatal EEG visual analysis in hypoxic-ischemic encephalopathy severity: Bridging the gap between eyes and machine

    2021, Neurophysiologie Clinique
    Citation Excerpt :

    The lowest value of Maximal AI was not significantly different between group 1 (29 μV) and 2 (26.5 μV) but decreased significantly in group 3 EEGs (7 μV). These results are consistent with the values found in previous studies [11,23,42,43] and in particular in aEEG classifications of HIE severity proposed by al Naqeeb in 1999 and Hellström-Westas in 2006 [19,32] (Table A.3). Experimental studies have shown that severe cerebral anoxia was associated with an initial suppression of the EEG amplitude during the primary phase of energy failure and during the latent phase despite normalisation of energy mechanism [3,17,21].

View all citing articles on Scopus

The authors have no conflicts of interest to declare.

View full text