Autism is a devastating neurological disorder with marked impairment in social skills, behavior, cognitive functions and verbal communication. Frequent co morbidities include seizure disorders, mental retardation, mood and sleep disorders. In Arkansas, the prevalence of autism is 1 in 145 children with higher incidence in boys than girls (odds 1.78 to 1). Recent studies suggest that one third of children in the autistic spectrum disorders (ASD) will develop seizures with onset prior to age 5 years (infancy) or after age 10 years (adolescence). Most children with seizures and ASD are likely to have an abnormal EEG reading. Moreover, in children with ASD without seizures, approximately 15-20% has an epileptiform EEG. About 68% of children with ASD with episodes of sudden or unprovoked crying, holding of the hands to the ears, unusual blinking patterns and brief staring spells had epileptiform activity in simultaneous EEG recording. These clinical findings suggest that subclinical seizures may be misdiagnosed in this population. Since the overt seizures are not usually evident until many years after the onset of autistic features, an early detection of subclinical seizures may improve the neurodevelopment of autistic children.
Our specific aim was designed to rigorously characterize this subgroup of autistic children in terms of specific behaviors and subclinical seizures and to determine the feasibility of early intervention strategies to avoid the progression into overt seizure activity.
Twenty children with Autistic Disorder or PDD-NOS using DSM-IV criteria, Childhood Autism Rating Scales (CARS) and the ADOS (Autism Diagnostic Observatory Scale) were recruited. Ten children had a history of subclinical seizures whereas the other ten were not diagnosed with subclinical seizure activity, thus this latter group was considered the control group. The population consisted of girls and boys with ages ranging between 18 months and 6 years. Informed consent was obtained and each child was subjected up to 2 hours of EEG recording including one hour of sleep and behavior testing using Vineland Adaptive Scales.
All children had computerized digital EEG recording utilizing 20 scalp electrodes with the standard 10-20 electrode placement and a channel of EKG recording in total 24 leads with montage reformatting capabilities. Interical epileptiform abnormalities specifically focal or multifocal with frontal, occipital or centro-parietal spikes were identified as abnormal. All EEGs were read by the PI, and two additional readers to reduce bias.
5 children had abnormal EEGs and all were identified with the initial symptoms of subclinical seizures. Logistic regression was used to determine the relevance of different factors to determine the likelihood of detecting changes in EEG readings. Few dimensions were moderately significant to determine the outcome of EEG readings, namely adaptive score (OR=0.75, p<0.1), personal dimension (OR=1.8, p<0.19), and previously diagnosed subclinical seizures (OR=20.9, p< 0.1).
We were unable to confirm the utility of Vineland Behavior Scales to determine EEG changes in autistic children, thereby limiting the use of this scale as a marker for subclinical seizures for this population.
See more of: Medical Co-Morbid Conditions
See more of: Symptoms, Diagnosis & Phenotype