Epidemiological and genomic studies suggest a significant effect of comorbidity of intellectual disability towards estimates of autism prevalence. S. Girirajan1, J. A. Rosenfeld2, A. Polyak1 1) Biochem & Molecular Biol, Pennsylvania State University, University Park, PA. 16802; 2) Signature Genomics Laboratories, PerkinElmer, Inc., Spokane, WA 99207.

   Current estimates of autism prevalence fail to take into account the effect of comorbidity of related neurodevelopmental phenotypes. We analyzed 11 years (2000-2010) of longitudinal data on approximately 6 million children per year from special education enrollment and 5,894 children referred for clinical microarray testing due to autistic features. We measured changes in autism prevalence and frequency, age-specific rates, and copy number variation (CNV) burden of comorbid features in autism. We found a 331% increase in the prevalence of autism from 2000-2010 based on special education enrollment. This prevalence increase corresponded to a concomitant decrease in the prevalence of intellectual disability (ID). The combined prevalence of ID and autism showed a 15% increase from 2000-2010 (22-fold less than autism prevalence alone). The decrease in ID prevalence equaled 61.5% of the increase in autism prevalence suggesting recategorization of diagnosis of ID to autism. Further, the frequency of phenotypes comorbid with autism was influenced by ascertainment, CNV burden, and gender. Comorbidity rates were higher for children in the clinical testing population compared to recent epidemiological estimates (40.3% vs 55.5%, p<0.0001, OR=1.84) and even higher among those with genomic disorders (64% vs 55.5%, p<0.0001, OR=2.51). The frequency of comorbid features varied across autism related genomic disorders such as 16p11.2 deletion (50%), 16p11.2 duplication (100%), 15q11.2 deletion (77%), and 1q21.1 duplication (56%). In the clinical testing population females showed a higher frequency of comorbid features (60% vs 46%, p=8.210-5, OR=1.27) and were more likely to manifest epilepsy comorbid with autism compared to males (p=0.02, OR=1.63). Among 1,588 autistic individuals carrying rare CNVs, females showed a higher large CNV burden for autism comorbid with ID (p=0.006, OR=2.14). Additional evidence suggests related genetic factors for varying comorbid features: individuals with epilepsy and autism were more likely to have ID compared to those with autism only (p=7.7210-13, OR=1.97), and similar CNV burdens were observed for both autistic individuals with epilepsy and those with ID (p=0.44). These observations support a global neurodevelopmental model where causal genes are shared between autism and related neurodevelopmental phenotypes, and phenotypes can be modified by genetic background to affect the age of onset, severity and variability of the disorder.

You may contact the first author (during and after the meeting) at