Document Type


Publication Date



Maine Medical Center, Medical Education, Maine Medical Center Research Institute, Psychiatry, Pediatrics

MeSH Headings

Child, Humans, Autism Spectrum Disorder, Child, Hospitalized, Sleep, Gastrointestinal Diseases


Background: Sleep and gastrointestinal (GI) problems are prevalent medical comorbidities in children with autism spectrum disorder (ASD). Sleep and GI problems are correlated with one another in both ASD and non-ASD children; however, mechanisms underlying this association are unclear. Anxiety is also highly prevalent in ASD and has been identified as a potential mediator of the sleep-GI association in non-ASD individuals. The role of anxiety in the sleep-GI association in ASD is unclear.

Objectives: To determine if anxiety symptoms statistically mediate associations between sleep problems and GI problems in a sample of psychiatrically hospitalized children with ASD.

Methods: Data from the Autism Inpatient Collection (AIC) were used for this analysis. The AIC is a multi-site study of children with ASD admitted to one of six inpatient units specialized for the psychiatric treatment of children with ASD. Caregivers provided information regarding medical comorbidities and completed the Child Behavior Checklist (CBCL). Autism diagnosis, autism severity, verbal ability, and intellectual functioning were assessed by research-reliable Autism Diagnostic Observation Schedule-2 and the Leiter International Performance Scale-3. Complete data were available in 276 patients. Logistic regressions were used to determine if caregiver-reported sleep problems predicted GI problems independent of age, sex, autism severity, verbal ability, intellectual functioning, and anxiety symptoms. Sobel tests were used to determine if anxiety symptoms statistically mediated associations between sleep and GI problems. A sensitivity analysis was used to examine if results held when the CBCL Anxiety Problem question about nightmares was removed from the total score. Exploratory logistic regressions examined items within the CBCL Anxiety Problem subscale to determine if specific indicators of anxiety were driving analyses.

Results: Prior to adjusting for anxiety symptoms, Logistic regression analyses revealed that children with ASD and a sleep problem were 78% more likely to experience a GI problem (OR=1.78, 95% CI: 1.05-3.01) after controlling for age, sex, autism severity, verbal ability, and intellectual functioning. This association was no longer significant when anxiety symptoms were entered into the model. Every point increase on the CBCL Anxiety Problem t-score was associated with a 4.7% increase in the likelihood that a participant would have a GI problem (OR=1.05, 95% CI: 1.02-1.07). Sobel tests suggested that anxiety symptoms are a statistically significant mediator of the association between sleep and GI problems (Sobel test statistic = 3.14, SE = 0.10, p = 0.002). Anxiety symptoms continued to mediate the association between sleep and GI problems after removing the “nightmare” item from the subscale. Three indicators of anxiety were found to drive the mediation results. These included being 1) nervous, high-strung, or tense; 2) too fearful or anxious; and 3) worrying.

Conclusions: Anxiety symptoms mediated associations between sleep problems and gastrointestinal problems independent of autism severity, intellectual functioning, and verbal ability. Though the current study is cross-sectional, these results suggest that sleep problems may contribute to heightened anxiety symptoms which may in turn contribute to or exacerbate GI disturbances. Sleep disturbance and anxiety may be important intervention targets in children with ASD and GI disturbances.


2020 Costas T. Lambrew Research Retreat