Moderators of the Association between Symptoms of Anxiety and Autism

Background: Anxiety is one of the most common co-occurring diagnoses for autism spectrum disorder (ASD; White et al., 2009). Up to 80% of youth with ASD experience impairing symptoms of anxiety, and 40% of youth meet diagnostic criteria for an anxiety disorder (Simonoff et al., 2008). The literature on the association between anxiety and autism is mixed (Kerns et al., 2012), with different studies revealing positive, negative, and no association. Examining factors that influence the association between symptoms of autism and anxiety may help clarify these conflicting results and better characterize the presentation of anxiety in ASD. Understanding under what contexts autism and anxiety symptoms correlate may help identify individuals at greatest risk for anxiety early in development. However, few studies have investigated factors that influence the association between anxiety and autism symptoms.

Objectives: Examine the association between symptoms of autism and anxiety in a large clinic-referred sample of individuals with ASD, as well as factors that moderate that association.

Methods: Participants included 291 individuals evaluated for ASD and received an ASD diagnosis in a university-based autism diagnostic clinic. Participants ranged in age from 2 to 21 (191 males). Autism symptoms in both the social-communication and restricted/repetitive behavior (RRB) domains were measured using Autism Diagnostic Observation Schedule, domain calibrated severity scores (ADOS-CSS; Gotham, Pickles, & Lord, 2009; Hus, Gotham, & Lord, 2012). Anxiety symptoms were indexed using the Behavior Assessment System for Children (Kamphaus & Reynolds, 2015). Demographic moderators of interest included age, gender and verbal and nonverbal IQ.

Results: Across the full sample, total autism symptoms were negatively correlated with level of anxiety, p=.001. This was driven by a negative correlation between RRBs and anxiety, p<.001, as no correlation emerged between social-communication symptoms and anxiety, p>.05. Furthermore, age, VIQ and NVIQ were all positively correlated with anxiety levels, p’s<.001. Anxiety symptoms did not differ by gender, p>.05. The association between ADOS-CSS and anxiety symptoms was not moderated by age, VIQ, NVIQ, or gender for either domain of symptoms.

Conclusions: This was the first study to examine factors that moderate the association between anxiety and autism symptoms in a large clinic-referred sample of youth with ASD. Individuals who evidenced fewer RRBs during the ADOS were rated as more anxious. It may be that highly anxious individuals are aware of the social evaluation involved in a behavioral assessment such as the ADOS and regulate their behavior in that setting. Additionally, factors such as insistence on sameness may relate to both anxiety and RRBs, indicating a need for differentiating contributors to anxiety versus autism symptoms (Gotham et al., 2013). Our results also indicate that the association between autism and anxiety symptoms is preserved across a range of ages, cognitive levels and gender. Future research is needed to develop measures of anxiety validated for use with autism, and to determine whether moderations exist when using more sensitive measures of anxiety, as the presentation of anxiety may differ depending on a child’s intellectual or language ability, age, gender, or a combination of each.