Characterizing Children with Autism Spectrum Disorder (ASD) Who Respond to a Gluten-Free Casein-Free (GFCF) Diet

Background: The Gluten-Free Casein-Free (GFCF) diet continues to be one of the most commonly used complementary and alternative treatments in children with Autism Spectrum Disorders (ASD). However, the current state of science on the GFCF diet is mixed with some studies showing the diet may improve ASD symptoms in some children, while other studies have found no effects from the diet. Similarly, several healthcare providers and parents of children with ASD adamantly report some children to have noticeable positive effects from the GFCF diet. Given these mixed findings, evidence is mounting to indicate that a subgroup of children with ASD may exist, which could potentially explain the mixed effects of the GFCF diet. Yet, to date, there has been no attempt to identify this subgroup of children with ASD and the characteristics they possess.

Objectives: To identify children with ASD documented as responding to the GFCF diet and characterize them based on clinical presentations, laboratory findings, and medical histories.

Methods:  A retrospective chart review was conducted at a pediatric primary care clinic to examine medical records of children with ASD on a GFCF diet (2005-2015). IRB approval was obtained. Data collected from the medical records were inputted into the statistical software program SPSS. Statistical analysis evaluated the differences between GFCF diet responders and non-responders using a chi-square and Wilcoxon rank sum test.

Results: A total of 33 participants (n=33) were included in this study. The sample included 26 males and seven females. The age of participants ranged from 1.5-16 years (mean=5.07 years, SD=4.04). A total of 1,195 variables were collected from 33 participants. Twenty-two participants (n=22) were identified as responders to the GFCF diet and 11 participants (n=11) were identified as non-responders. Three types of ASD were identified: regressive after 12-months-old (n=13), non-regressive (n=18), and failed to progress after 12-months-old (n=2). A chi-square determined a significant difference existed in GFCF diet response and type of ASD (p=.026), plasma amino acid 3-methylhistidine (p=.013), and plasma amino acid alanine (p=.034). A Wilcoxon rank sum found significant differences between GFCF non-responders and responders in urine octenedioic (p=.006), fecal SIgA (p=.006), urine 3-hydroxyglutaric (p=.007), plasma amino acid glycine (p=.019), plasma amino acid alanine (p=.037), and serum alkaline phosphatase (p=.044).

Conclusions: Findings from this study indicate statistically significant differences in stool, plasma, and urine variables between children with ASD who responded to a GFCF diet and those who did not. These findings lend preliminary support for the clinical accounts of some children with ASD responding more favorably to a GFCF diet than others. It is expected that characterizing known responders to the GFCF diet will provide critical information to assist in clinical decision making, as well as help develop inclusion/exclusion criteria necessary for future clinical trials. Additionally, findings from this study revealed unexpected similarities among laboratory markers of all participants. Findings are expected to not only contribute to what is currently known, but also generate future research on biomarkers, microbiome, and characterization of all children with ASD.