Gluten, Dairy and Autism
People say lots of stuff. But what do the studies say?
In the last 20 years, advocates have promoted the gluten-free diet as a treatment for a host of ailments besides celiac disease [1]. Dairy-free food products have followed a similar path that extends beyond those who are lactose intolerant.
Somewhere along the line, autism got added to the list of things that could purportedly be treated by abstaining from wheat and/or dairy. By one recent estimate, one in five kids with autism has tried what is known as the gluten-free casein-free (GFCF) diet — a diet that is free of the gluten proteins in wheat, and the casein proteins in dairy [2].
Testimonials abound, but not in the absence of a handful of studies. What follows is a look at the design and outcomes of the most recent and rigorous study on this question. That study was by Susan Hyman and colleagues of the University of Rochester Medical Center, published in 2016 [3].
Dr. Hyman is a neurodevelopmental pediatrician who specializes in autism, particularly as it relates to diet and nutrition. In this capacity, she has over three decades of experience.
Hyman and colleagues looked at 14 children with autism between the ages of 3 to 5 years old, who were not known to have gluten or dairy intolerances, and were enrolled in an applied behavior analysis (ABA) therapy program. After at least 4 weeks on the GFCF diet, the participants were given randomized weekly food samples that contained gluten, casein, gluten and casein, or a placebo over a 12 week period. Several markers of health and behavior were evaluated throughout the process. The figure below illustrates how this was carried out over a 30 week period.
This study was double blind, meaning that the evaluators and the people administering the food challenges did not know which of the four food challenges the participants were receiving each week.
In a 2018 interview with Autism Speaks, Susan Hyman emphasized the importance of using a double blind approach in her study. “In testing any intervention, it’s important that the expectations of the participants don’t influence the outcome,” Hyman explained. “One problem with prior studies examining dietary intervention in autism is that they were based on the observations of those who knew what they were feeding their children.”
Hyman’s study also controlled for the effects of other therapies that occurred during the study by only including children who were enrolled in similar ABA programs. And the study involved dieticians who could work with the caregivers and ensure that the GFCF diet was being followed.
Weaknesses included a small sample size of only 14 participants, and a short trial period of 30 weeks. Such are the tradeoffs that often occur with more rigorous studies. Another weakness is that the results are not applicable to the fraction of autism cases that are known to have dairy intolerance or gluten sensitivity.
Among all 14 participants, the GFCF diet had no measured effect on behavior, or physiological markers such as sleeping and pooping. The authors were cautious in not drawing any overarching conclusions due to the small sample size. In essence, they found evidence that the diet could be safe, but not that it could be generally effective for treating autism: “our study does not provide evidence to support general use of the GFCF diet.”
A more recent literature review published by Benjamin Lerner and colleagues from Columbia University came to a similar conclusion regarding the efficacy of the gluten-free diet [1].
The Hyman study also addresses four other previous studies that investigated the GFCF diet with autism, and the strengths, weaknesses and outcomes of those studies. In a nutshell, the Knivsberg 2002 study and Whiteley 2010 studies found some signs of improvement with the GFCF diet, but were mostly reliant on reports from the children’s caregivers who were not blinded [4,5]. The Elder 2006 and Johnson 2011 studies saw no improvement with the GFCF diet, but included smaller sample sizes over shorter periods of time [6,7].
My above illustration is too simple to capture the finer details of the Hyman study design and outcomes. If you can, I suggest that you read the whole thing here, which is unfortunately behind a paywall [3]. For more information about how this study was carried out, and how the results should be interpreted, I also suggest reading the full Autism Speaks interview with Hyman that I referenced earlier. It’s quite informative.
See also my previous post here on survey data for various other diets and their effectiveness at reducing autistic behavior — according to parents.
Literature References:
- Lerner, 2019: https://link.springer.com/article/10.1007%2Fs10620-019-05663-x
- Rubenstein, 2018: https://onlinelibrary.wiley.com/doi/full/10.1002/aur.1896
- Hyman, 2016: https://link.springer.com/article/10.1007%2Fs10803-015-2564-9
- Knivsberg, 2002: https://www.tandfonline.com/doi/abs/10.1080/10284150290028945
- Whiteley, 2010: https://www.tandfonline.com/doi/abs/10.1179/147683010X12611460763922
- Elder, 2006: https://link.springer.com/article/10.1007%2Fs10803-006-0079-0
- Johnson, 2011: https://link.springer.com/article/10.1007/s10882-010-9217-x
