Joshua B Ewen1,2,3, Alison R Marvin4, Kiely Law4,5, Paul H Lipkin4,5. 1. Department of Neurology and Developmental Medicine, Kennedy Krieger Institute, Baltimore, Maryland. 2. Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, Maryland. 3. Department of Psychological and Brain Sciences, Johns Hopkins University, Baltimore, Maryland. 4. Department of Medical Informatics, Interactive Autism Network at Kennedy Krieger, Baltimore, Maryland. 5. Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, Maryland.
Abstract
Epilepsy is known to occur in a higher-than-expected proportion of individuals with autism spectrum disorders (ASDs). Prior studies of this heterogeneous disorder have suggested that intelligence quotient (IQ) may drive this relationship. Because intellectual disability (ID) is, independently of ASD, a risk factor for epilepsy, current literature calls into question the long-understood unique relationship between ASD and epilepsy. Second, data have been unclear about whether developmental regression in ASD is associated with epilepsy. Using two cohorts from an online research registry, totaling 6,975 children with ASD, we examined the independent role of four ASD severity measures in driving the relationship with epilepsy: ID, language impairment, core ASD symptom severity, and motor dysfunction, controlling for two known relevant factors: age and sex. We also examined whether developmental regression and epilepsy have an independent statistical link. All four ASD severity factors showed independent statistical associations with epilepsy in one cohort, and three in the other. ID showed the largest relative risk (RR) in both cohorts. Effect sizes were modest. Regression similarly showed an independent statistical association with epilepsy, but with small effect size. Similar to previous work, ID showed the greatest contribution to RR for epilepsy among children with ASD. However, other ASD severity markers showed statistical associations, demonstrating that the ASD-epilepsy association is not reducible to the effect of ID. Inconsistencies in the literature may be due to underpowered studies, yet moving forward with larger-n studies, clinical significance and scientific relevance may be dictated by effect size and not merely statistical significance. Autism Res 2019, 12: 1251-1259.
Epilepsy is known to occur in a higher-than-expected proportion of individuals with autism spectrum disorders (ASDs). Prior studies of this heterogeneous disorder have suggested that intelligence quotient (IQ) may drive this relationship. Because intellectual disability (ID) is, independently of ASD, a risk factor for epilepsy, current literature calls into question the long-understood unique relationship between ASD and epilepsy. Second, data have been unclear about whether developmental regression in ASD is associated with epilepsy. Using two cohorts from an online research registry, totaling 6,975 children with ASD, we examined the independent role of four ASD severity measures in driving the relationship with epilepsy: ID, language impairment, core ASD symptom severity, and motor dysfunction, controlling for two known relevant factors: age and sex. We also examined whether developmental regression and epilepsy have an independent statistical link. All four ASD severity factors showed independent statistical associations with epilepsy in one cohort, and three in the other. ID showed the largest relative risk (RR) in both cohorts. Effect sizes were modest. Regression similarly showed an independent statistical association with epilepsy, but with small effect size. Similar to previous work, ID showed the greatest contribution to RR for epilepsy among children with ASD. However, other ASD severity markers showed statistical associations, demonstrating that the ASD-epilepsy association is not reducible to the effect of ID. Inconsistencies in the literature may be due to underpowered studies, yet moving forward with larger-n studies, clinical significance and scientific relevance may be dictated by effect size and not merely statistical significance. Autism Res 2019, 12: 1251-1259.