Continuous and Extreme Autistic Trait Ratings Are Associated with Avoidant Restrictive Food Intake Disorder

Background:   From its earliest description, autism spectrum disorder (ASD) has been linked with feeding-related problems. Dietary abnormalities are very common in ASD, including food selectivity, food refusal, and insistence on sameness while eating. Moreover, food selectivity is not limited to childhood, but clearly persists into the adolescent and adult years in ASD. With the adoption of DSM-5, a new eating/feeding disorder was introduced, avoidant restrictive food intake disorder (ARFID). ARFID is characterized by clinically significant difficulties with eating/feeding, including areas of likely phenotypic overlap with ASD, such as food avoidance based on sensory characteristics. In spite of this, no study to date has linked these two conditions. The present study seeks to fill this gap in knowledge. 

Objectives:   Examine continuous and extreme autistic trait ratings within a large group of participants oversampled for issues related to selective eating, including ARFID. 

Methods:   A large group of 1,992 adults (1,445 females) completed online surveys including self-ratings of autistic traits utilizing the original 50-item Autism-Spectrum Quotient (AQ) and eating-related behaviors (e.g., limitations in dietary intake, weight loss connected to limited dietary intake, sensory-related food sensitivities, etc.) validated to assess features of ARFID. The degree to which these eating related problems interfered with broader social and job functioning was also assessed. 

Results:   Individuals with symptoms consistent with ARFID (n=1,390), had higher overall autistic trait ratings (M=21.28, SD=7.38) than those without ARFID (n=602; M=20.16, SD=7.50; t=3.08, p=.002).  More specifically, the ARFID group had elevated autistic traits in the areas of social skills (M=4.42, SD=2.76 vs. M=4.08, SD=2.78), attention switching/flexibility (M=5.35, SD=2.25 vs. M=4.94, SD=2.31), and communication (M=3.48, SD=2.28 vs. M=3.19, SD=2.23) compared to the non-ARFID group (ts>2.44, ps<.02). In turn, participants with extreme autistic trait ratings (total AQ score>31; n=160) were more likely to rate themselves as having ARFID (Χ²=3.45, p=.036, one-tailed), than individuals with subthreshold scores. Finally when comparing how eating-related problems affected broader functioning, participants with both ARFID and elevated autistic traits (n=121) reported greater interference with their jobs than did those with ARFID alone (n= 1,268; Χ²=3.39, p<.05, one-tailed) and greater anxiety and avoidance of social situations involving food (ts>2.25, ps<.025).

Conclusions:   This is the first study to examine phenotypic overlaps of ASD and ARFID.  Perhaps unsurprisingly, there were relatively strong associations between both continuous autistic trait ratings and extreme levels of autistic traits and ARFID.  More specifically, the social-communication and inflexibility components of the ASD phenotype appeared most closely linked with ARFID.  The inflexibility component of the ASD phenotype is particularly striking given conceptual connections drawn between it and selective eating in ASD. Additionally, elevated autistic traits appeared to exacerbate interference with everyday social and occupational functioning among individuals with ARFID. Much more work is needed to understand the shared and unique mechanisms driving eating-related idiosyncrasies that disrupt daily life among individuals with ASD and/or ARFID, which will serve to inform future intervention development.