Electroencephalogram in children with hepatic encephalopathy: towards a new classification?

Diagnosis of hepatic encephalopathy (HE) is challenging in children. The electroencephalogram (EEG) is an easily accessible and pivotal tool but no satisfactory classification is currently available.

We studied inter-observer agreement with Gwet's AC2 coefficient with quadratic weights, for Navelet et al.'s (1990) classification (1990) in EEGs successively performed for chronic or acute liver disease in a reference centre. We analysed reasons for discordance and proposed a new classification. We studied the association between both classification, the neurological impairment and hepatic biological dysfunction.

171 EEGs were analysed. Slowing of the dominant rhythm and loss of reactivity characterized the pathological awake EEGs. Using the Navelet classification, there was a very high inter-rater agreement beyond chance. We have refined it with Children's Hepatic Encephalopathy EEG Recording Scale (CHEERS). Inter-rater agreement for the CHEERS classification in 27 supplementary EEGs was almost perfect beyond chance. Association between each EEG scale and the neurological impairment and hepatic biological dysfunction was significant.

Our proposed classification is reliable in detailing abnormalities when suspecting HE in children.

The next step will be to validate the classification in an external cohort by independent readers and to prove its utility in detecting the first stages of deterioration of hepatic function.