Food and Behaviour Research

Donate Log In

Choline - A Neglected Nutrient Vital for Healthy Brains - BOOK HERE

The Gluten-Free/Casein-Free Diet: A Double-Blind Challenge Trial in Children with Autism

Hyman SL, Stewart PA, Foley J, Cain U, Peck R, Morris DD, Wang H, Smith T (2016)  J Autism Dev Disord.; 46(1):205-20. doi: 10.1007/s10803-015-2564-9. 205-20. doi: 10.1007/s10803-015-2564-9 

Web URL: Read the abstract and associated research on PubMed here


To obtain information on the safety and efficacy of the gluten-free/casein-free (GFCF) diet, we placed 14 children with autism, age 3-5 years, on the diet for 4-6 weeks and then conducted a double-blind, placebo-controlled challenge study for 12 weeks while continuing the diet, with a 12-week follow-up.

Dietary challenges were delivered via weekly snacks that contained gluten, casein, gluten and casein, or placebo. With nutritional counseling, the diet was safe and well-tolerated. However, dietary challenges did not have statistically significant effects on measures of physiologic functioning, behavior problems, or autism symptoms.

Although these findings must be interpreted with caution because of the small sample size, the study does not provide evidence to support general use of the GFCF diet.


With a sample size of only 14 children in total, it would be quite extraordinary for anyone to expect 'statistically significant' differences between those who were allocated to a gluten-free diet for 12 weeks, and those who were not (unless of course the intervention were to be as extreme as 'the use of a parachute or not' when jumping from a plane...)  

To say that 'the findings must be interpreted with caution because of the small sample size' is thus a rather major understatement.  With so few children included, this study can at best be regarded as a pilot 'feasibility' study only.  

Without suitably qualified help in planning the diet, the removal of all gluten grains can increase the risk of nutrient deficiencies. The authors' comment that 'with nutritional counseling, the diet was safe and well-tolerated' is thus potentially informative and encouraging for planning future studies.

However, those do need to be of sufficient size to be able to detect any effects that might follow from a gluten-free diet. (Ideally, such trials should also be large and well-designed enough to examine subgroups defined by other relevant features - because 'autism' is a purely descriptive diagnosis that conceals substantial variability, and all the existing evidence points to there being no single 'cause' - but rather, a multiplicity of risk factors).

See also: