Possible gastrointestinal symptoms in a subset of children with autism

There is a continuing debate as to whether or not gastrointestional (GI) problems may exist in a subset of children with autism. If GI disorders are common in some children with developmental disorders, why could they not also exist in children with autism? It is known that constipation and fecal impaction are common in children with mental retardation [1], Rett syndrome [2], Prader–Willi syndrome, Hunter disease and autism [3,4]. It is also an established fact that a certain percentage of children with autism have varying degrees of mental retardation, thus is it not possible that these children may experience the increased risk of GI symptoms associated with mental retardation? A recent review by Buie et al. in Pediatrics reported that the prevalence of GI symptoms (constipation, diarrhea, bloating, belching, abdominal pain, reflux, vomiting and flatulence) in children with autism spectrum disorders (ASD) range widely from nine to 91% based on 11 studies that average 44% [5]. It is interesting to note that D’Eufemia et al. observed that 43% (nine out of 21) of children with ASD had altered intestinal permeability (leaky gut) compared with 0% of controls (zero out of 40) [6].

Other GI problems common in genetically related mental disorders include: possible functional megacolon in Rett syndrome, where 76% of these patients experience GI symptoms [7]; gallbladder dysfunction (gallstones and cholecystectomy) in Rett syndrome [8]; and gastroparesis in Prader–Willi syndrome [9]. Levy even stated that “Down syndrome is recognized as one of the most common predisposing conditions for a group of serious GI anomalies – tracheo–esophageal fistula, duodenal obstruction with or without pyloric stenosis, annular pancreas, imperforate anus and Hirschsprung’s disease” [10]. He added that “intestinal anomalies can be found in many other genetic disorders, with recent evidence suggesting the presence of GI developmental regulatory genes on chromosome 13q.”

In terms of celiac disease, The North American Society for Pediatric Gastroenterology, Hepatology and Nutrition recognizes in their clinical guidelines that at least three developmentally delayed conditions have higher risk rates than the normal population [11]. They reported that strong evidence exists for an association between Down syndrome and celiac disease, with a prevalence rate of 5–12%. Approximately one third of Down syndrome patients with celiac disease exhibit no GI symptoms. Hilhorst et al. reported a 43-times higher celiac rate in patients with Down syndrome [12]. According to the 1999 Health Care Guidelines of the Down Syndrome Medical Interest Group, an initial screening for celiac disease is recommended at the age of 2 years [13]. Approximately 4–8% of children with Turner syndrome have celiac disease [14–18], and one study revealed an 8% rate in children with Williams syndrome [19]. The North American Society for Pediatric Gastroenterology, Hepatology and Nutrition also recognized increased rates of celiac disease in children with Type 1 diabetes mellitus. The 2009 American Diabetes Association Standards of Medical Care in Diabetes listed the increased rates of celiac disease to be 1–16% versus 0.3–1% in the general population, and recommended celiac disease screening in these patients [20]. Children with cerebral palsy are also at risk as 43% have high serological celiac markers [21].

Hans Asperger himself suggested a possible link between celiac disease and autism more than half a century ago [22]. Why would the very real possibility of GI problems, including gluten problems, existing in a subset of children with autism even be debated by some researchers? The fact that certain GI problems already exist in some children with developmental delay disorders indicates that children with autism might also be susceptible. This is particularly true since approximately 30% of children with autism exhibit some degree of mental retardation, and that celiac disease is observed in approximately 10% of children with Down syndrome, Turner syndrome and Williams syndrome. Finally, GI symptoms in children with ASD range from 9 to 91%, averaging 44% [5]. In light of these findings, does it make logical sense to eliminate the possibility of GI problems and/or celiac disease existing in a subset of children with autism without thorough scientific investigation? Where is the scientific objectivity when weak research results [4] are nationally promoted as “putting to rest the nagging suggestion that there exists a link between autism and gastrointestinal disease” [23]. This comment in Pediatrics was based on a study by Ibrahim et al. who reported no significant differences in prevalence of GI symptoms in children with autism (77.2%) compared with controls (72.2%) [4]. Their retrospective study of Mayo Clinic medical records (121 children with autism between 1976 and 1997) attempted to determine whether autistic children had more visits to the hospital for GI symptoms than nonautistic children. The conclusion that there is no link between GI symptoms and autism was used to suggest that parents do not need to place their children on restrictive diets.

Serious limitations of Ibrahim’s study include, but are not limited to:

• • It was a retrospective study relying entirely upon medical records that tend to be incomplete;

• • This was not a clinical study testing the efficacy of diet in autistic children;

• • It is not clear whether these were children under the broader ASD or the more specific autistic disorder. Specifically, it was not clear if Asperger syndrome and pervasive developmental disorder, not otherwise classified, were included in the autism diagnosis;

• • Subjects were not all formally diagnosed with autism by clinicians, but most frequently as developmental delay (33%), delayed speech and language (42%), attention-deficit/hyperactivity disorder (18%) and mental retardation (16%) [24];

• • Categorization of GI problems ranged from very mild to severe, but were grouped together anyway. For instance, gastroesophageal reflux disease is known to be higher in autistic children, but in this study gastroesophageal reflux disease and vomiting were grouped together, possibly diluting any significant difference;

• • Much variability exists among physicians regarding inquiries of GI symptoms in children with or without autism;

• • It was not possible to include time-to-event data that are important for dietary reactions;

• • They failed to assess or compare the duration, severity and recurrence of the GI symptoms between the autistic subjects and the normal controls;

• • Researchers did not document diets (especially gluten-/casein-free diets), dietary supplement use or medications with possible GI side effects;

• • The misclassification of symptoms and clinical diagnoses may have reduced the overall power of the study;

• • They concluded that the significantly greater prevalence of constipation and food intolerance among children with autism could be behaviorally induced.

Propagating medical conclusions based on a single, retrospective study with the above limitations, in light of the fact that GI problems do exist in subsets of children with developmental delayed disorders (including autism), is not only curious, but may jeopardize the health of a subset of children with autism who may have GI symptoms that need to be treated. Recent recommendations for the evaluation and treatment of common GI problems in children with ASD have been published in Pediatrics[25]. The researchers state that children with ASD can benefit from the evaluation of abdominal pain, chronic constipation and other GI symptoms.

Financial & competing interests disclosure

Amy C Brown is CEO of Natural Remedy Labs, LLC, a Cengage textbook author and current grant recipient of the Broad Medical Research Program of the Broad Foundation. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.