Defining the Characteristics of Distinct Neurological Diseases in Patients with Autism Spectrum Disorders and Epilepsy

Background: Autism spectrum disorder (ASD) is associated with an increased incidence of epilepsy over the general population. There is considerable variability in the reported incidence; ranging from 5 to 40%. Co-morbid epilepsy contributes negatively to the cognitive, adaptive, behavioral and emotional outcomes for individuals with autism. Early identification of those at particular risk can help target counseling, surveillance measures, and treatment. Various studies have reported risk factors for epilepsy in ASD, some of which are well established including lower IQ, underlying neurologic disorder, language disorders, and sex. 
Objectives: To define distinguishing phenotypes of patients with diseases causing autism and epilepsy in a series of 129 patients with ASD and suspected seizures referred to Vanderbilt University Hospital for electroencephalography from January 2004 through February 2009.
Methods: We analyzed the medical records and 147 EEGs including routine (114) and long term studies (37) to determine characteristics of the ASD patients with epilepsy vs ASD patients without epilepsy. We looked at these variables in both groups: gender, regression, head circumference, rates of epileptiform EEG, and genetic studies.
Results: There was a preexisting diagnosis of epilepsy in 31% of all patients and a history of developmental regression in 17%. Previously diagnosed epilepsy was more likely in patients with a history of developmental regression (p=0.0085). Among patients without a prior history of epilepsy, a history of regression significantly increased the risk of having epileptiform discharges on a screening EEG (p=0.014). When patients with regression were subdivided by age, we found significantly more epileptiform EEGs when the study was performed at age 10 years or older (p=0.0015). Further, the rate of increasing head circumference between < 10 years of age vs > 10 years was significantly less (1.5 cm vs 3.7 cm, p<0.01) in ASD patients with epilepsy as compared to ASD patients without epilepsy. The presence of single gene disorders, copy number variants (gains and deletions), and Trisomy 21 does not explain this difference in ASD patients with epilepsy as opposed to ASD patients without epilepsy.
Conclusions: Given the high frequency of epilepsy in this patient population, a high index of clinical suspicion should be maintained for subtle symptoms of seizures in ASD. There is inconclusive evidence to recommend screening EEGs in asymptomatic ASD patients without spells. We propose that screening EEGs should be considered in patients with ASD over 10 years of age with a history of developmental regression and a history of spells suspect for seizures. Further these data suggest ASD patients with epilepsy have a distinct developmental trajectory of head growth compared to ASD alone. A prospective study of ASD patients with and without epilepsy would be helpful to validate this data.