Addendum to “gluten-free and casein-free diets in treatment of autism spectrum disorders: A systematic review”

Abstract

Shortly after the publication of our literature review on gluten-free and casein-free (GFCF) diets in the treatment of autism spectrum disorders (ASD; Mulloy et al., 2010), Whiteley et al. (2010) published the results of experimental evaluation of a GFCF diet. To update our previous literature review, we herein provide an analysis of the Whiteley et al.’s study. The study was well designed, but also contains substantial weaknesses that suggest the need for caution in interpreting the findings reported by Whiteley and colleagues. In conclusion, we maintain our previous position that GFCF diets are not effective in treatment of core symptoms of ASD and should only be implemented in the event a child with ASD experiences acute behavioral changes, seemingly associated with changes in diet, and/or medical professionals confirm through testing the child has allergies or food intolerances to gluten and/or casein.

Introduction

Shortly after Research in Autism Spectrum Disorders published our literature review on gluten-free and casein-free (GFCF) diets in the treatment of autism spectrum disorders (ASD; Mulloy et al., 2010), a study by Whiteley et al. (2010) involving the experimental evaluation of a GFCF diet was published in Nutritional Neuroscience. Whiteley et al. concluded their data demonstrated that the GFCF diet had a positive effect on the developmental outcomes of some children with ASD. In their study, they implemented a strong research design that commendably included a combination of elements not present in previous studies of the GFCF diet (i.e., random assignment, long diet periods, and a blinded measure). Such elements are required to produce conclusive evidence (Schlosser, 2009, Simeonsson and Bailey, 1991, Smith, 1981). However, the study also presents a number of weaknesses (methodological and data concerns) that recommend caution in interpreting the results, limit the possibility of drawing definite conclusions, and do not currently allow one to reverse the general views expressed in our review.

From our review of the literature on GFCF diets in treatment of ASD, we noted that there were some studies that reported that the diet is efficacious and some that reported it is ineffective. All studies contained weaknesses that compromised their findings’ validity, although studies that reported that the diet was efficacious contained the most severe and numerous flaws. In our review, we concluded that the literature does not support use of GFCF diets in treatment of ASD in light of the methodological problems with these studies, including problems related to internal validity, measurement problems, construct validity, and plausible alternative explanations for study outcomes.

The Whiteley et al.’s (2010) study admirably protected against threats to internal validity, measurement problems, and construct validity with its strong research design. Even so, weaknesses regarding Bonferroni adjustments for statistical tests, systematic bias in the pattern of participant attrition, discrepant outcomes across comparable sub-domains of measures, and potential sources of measurement bias suggest the findings and conclusions of Whiteley and colleagues should be interpreted with caution.