Troubles spécifiques du langage oral et électroencéphalogramme : quelles indications ? Quelle interprétation ? À propos d’une cohorte de 24 enfants

Abstract

Electroencephalographic recording (electroencephalogram [EEG]) is frequent in specific language impairment (SLI), whereas the relations between epileptiform activity (EA) and language disorders remain uncertain and the therapeutic approach undetermined. The aim of this prospective study was to clarify EEG indications and interpretation in SLI. We present a prospective study of cognitive (speech-language measures, psychological assessments) and electroencephalographic data on 24 children (20 males, 4 females; mean age: 4 years 5 months; range: 3 years to 4 years 8 months) with a diagnosis of SLI, defined as a pathologic score on at least 2 speech-language measures and IQ performance of at least 80 points, within epileptic seizures. All participants had an EEG after partial deprivation of sleep at night. When nonsporadic EA was found, 24-h EEG was performed. Antiepileptic treatment was prescribed depending on the frequency of discharges and the SLI profile. The follow-up lasted 2 years. All patients reached stage II sleep during their EEG. Seven children had abnormal electroencephalography results, including 5 children with EA. Two patients with mixed SLI prevailing on expression presented a left centrotemporal spike focus on EEG becoming subcontinuous during sleep. In the first case, the language progressed without antiepileptic treatment. The 2nd case was treated with ethosuximide; the EEG normalized on subsequent recordings, but the language disorder remained severe. The lexical and syntactic understanding and syntactic production scores were not different for children presented EA (5 cases) or without (19 cases) (Wilcoxon's test). Finally, the progression of the various linguistic skills was similar whether or not the children had EA (p<0.1). Abnormal electroencephalographic activity is more frequent in SLI than in normal children. It can be seen in all types of SLI but preferentially in the mixed forms. The longitudinal systematic evaluation of all the children with or without EA has never been reported in the literature, which confirms that there is no parallel between EEG progression and language development, contrary to Landau-Kleffner's syndrome. Systematic EEG recording is not recommended in SLI. It is indicated in cases of fluctuation or stagnation of language development or if there is a family history of cognitive disorders or epilepsy. Antiepileptic treatment is justified in cases with fluctuation or stagnation in language development after 6 months of progression associated with frequent EA. However, when the EA is particularly diffuse and dense, an additional moderate aggravating effect on language development cannot be eliminated. The method most likely to clarify this question is to compare a child's baseline phase followed by a treatment phase, with the same double follow-up.