We conducted a retrospective comparative cohort study to determine the phenotypic and real-world management differences in children with epilepsy and co-occurring autism as compared to those without autism.
Clinical variables, EEG, brain MRI, genetic results, medical and non-medical treatment were compared between 156 children with both epilepsy and autism, 156 randomly selected and 156 demographically matched children with epilepsy only. Logistic regression analyses were conducted to determine predictors of drug-resistant epilepsy (DRE).
As compared to the’matched’ cohort, more patients with autism had generalized motor seizures although not statistically significant after Benjamini-Hochberg correction (54.5%, vs 42.3%, p = .0314); they had a lower rate of electroclinical syndromes (12.8%, vs 30.1%, p = .0002). There were more incidental MRI findings but less positive MRI findings to explain their epilepsy in children with autism (26.3%, vs 13.8% and 14.3%, vs 34.2%, respectively; p = .0003). In addition, LEV, LTG, and VPA were the most common ASMs prescribed to children with autism, as opposed to LEV, OXC, and LTG in children without autism. No difference in the major EEG abnormalities was observed. Although the rates of DRE were similar (24.8%, vs 26.6%, p = .7203), we identified two clinical and five electrographic correlates with DRE in children with both epilepsy and autism and a final prediction modeling of DRE that included EEG ictal findings, focal onset seizures, generalized motor seizures, abnormal EEG background, age of epilepsy onset, and history of SE, which were distinct from those in children without autism.
Our study indicates that detailed seizure history and EEG findings are the most important evaluation and prediction tools for the development of DRE in children with epilepsy and co-occurring autism. Further studies of epilepsy in specific autism subgroups based on their etiology and clinical severity are warranted.

Copyright © 2023 Elsevier Inc. All rights reserved.