Q&A with author of new study on autism and gluten/casein-free diet
October 9, 2018
In 2015, we reported on the findings of the first carefully controlled study of how a gluten-free and casein-free (GFCF) diet might benefit children who have autism. This small study found no improvement in behavioral or bowel symptoms.
Neurodevelopmental pediatrician Susan Hyman led the first controlled study to look for benefits of the gluten/casein-free diet in children who have autism.We followed up with study senior author Susan Hyman with a broad range of your questions. Dr. Hyman is a neurodevelopmental behavioral pediatrician at the University of Rochester Medical Center. The center is one of 14 sites in the Autism Speaks Autism Treatment Network.
How did your study differ from past looks at the GFCF diet and autism?
Dr. Hyman: In testing any intervention, it’s important that the expectations of the participants don't influence the outcome. 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.
The term “double blind" means that neither the observer nor the participants knows whether the intervention is the active or control situation.
In our study, the parents and researchers were the observers and the children were the participants. The active situation, or “challenge,” consisted of food containing gluten and/or casein. The control situation was look-alike, taste-alike foods free of both of these two food proteins. In this way, we prevented expectations from influencing the results.
In addition, we designed the study to minimize the likelihood that other factors would influence response to diet. This included:
- Controlling for differences in behavioral therapies. We felt it was important that all the children in our study were receiving intensive, high-quality behavioral intervention (applied behavioral analysis). This enabled us to ensure that any changes could be attributed to the diet alone.
- Not including children with celiac disease or food allergies. Either of these conditions can predispose children to react to foods containing wheat or milk. When we were evaluating children for our study, we used blood tests to look for these conditions. In doing so, we identified two children with celiac disease. Though we didn’t enroll them in the study, we did refer them for GI evaluation and treatment that may include a gluten-free diet.
Might some subsets of children with autism benefit from a GFCF diet?
Dr. Hyman: As just mentioned with celiac disease, exposure to the gluten in wheat can cause physical and/or behavioral symptoms. Allergies to wheat or milk can produce discomfort as well. It’s important that children with autism who have GI symptoms be evaluated for these and other possible underlying medical issues.
It’s also possible that children with lactose intolerance – the inability to digest milk sugar, not milk protein – will have irritability and GI symptoms when they consume milk.
How narrowly or broadly did you look for the potential benefits of the GFCF diet?
Dr. Hyman: We cast a wide net. Before and after the challenges [foods containing casein or gluten], we systematically recorded information on both physiology and behavior.
The physiologic functions included growth, bowel frequency and consistency [diarrhea or constipation], and sleep.
The behaviors included both those specific to autism – like social play, communication and repetitive behaviors – and those that commonly co-occur with autism but are not unique to it. The latter include hyperactivity and disrupted sleep. We recorded this information for 24 hours after each snack to look for any delayed response.
How did you ensure that kids weren’t inadvertently being exposed to gluten or casein during the study?
Dr. Hyman: We had a registered dietitian teach each family how to maintain the diet. The dietician called them each week to go over their 24-hour food diaries. This enabled us to minimize the chance of accidental ingestions of gluten or casein and to identify these occurrences when they happened. We learned that it’s challenging to identify food ingredients, even for very educated and careful families!
In addition, we made sure that all the children were on the GFCF diet for at least four weeks before we started challenging them with gluten or casein. We allowed for a two-week period for parents to gradually introduce the diet. But most families jumped in with both feet from the start. So most of the children were on the GFCF diet for six weeks before we started the challenges. This initial period allowed us to make sure we were seeing true baseline behaviors of children on a GFCF diet.
Then, the children were randomized to weekly snacks containing either gluten or casein, both proteins, or neither. The snacks were developed in the Clinical Research Center at the University of Rochester and pilot tested to make certain that the taste and texture of the products were indistinguishable. We also customized the snacks to the food preferences of each child. A dietitian brought the study snack to the child and weighed any leftovers.
We also recorded behavioral response to what we called "natural challenges." These were the "oops" moments when a child inadvertently ate gluten- or casein-containing foods.
AS: In the Wall Street Journal news story on your study, a participating parent is quoted as saying there were times when she was “sure” their child was getting the gluten/casein snack because his behavior would worsen. But it turned out he wasn’t getting gluten or casein at these times. Were there multiple instances of this?
Dr. Hyman: Yes. I recall one family that requested that we not repeat a specific challenge snack because they thought it was disrupting their child’s sleep. That turned out to be the placebo version of the snack.
AS: Why are larger studies needed?
Dr. Hyman: A large study might identify specific groups of children who respond to dietary interventions.
Are there other diet and autism studies you’d like to see researchers pursue?
Dr. Hyman: We set out to examine only the effect of wheat and milk proteins. More recently, there’s been interest in the potential effect of the microbiome [digestive bacteria] on intestinal health and perhaps behavior in people with autism. We know that altering the diet can alter a person’s gut flora, and this may affect behavior and/or GI symptoms. We don’t yet have evidence that this happens. There may also be other aspects of diet change that can affect behaviors in subsets of children who have autism.
How difficult will it be to do larger studies given the expense and difficulty of double-blinding and preparing the look-alike, taste-alike foods?
Dr. Hyman: Very! The design of future studies needs to include great care in eliminating potential confounding influences and events. There is also the difficulty of recruiting more families, as this was very time intensive for all our participants.
Should future studies look at kids with autism and GI disorders?
Dr. Hyman: We excluded children who tested positive for celiac antibodies and those whose blood testing suggested allergy to wheat or milk. In children with autism and celiac disease, it is possible that elimination of gluten will result in better sleep, more typical stool patterns and less irritability. Also since celiac disease is common in the general population, we might expect overlap to occur in children with autism and GI symptoms. In fact, evaluation for celiac disease should be part of the GI evaluation for these children if they have consistent symptoms. Similarly, lactose intolerance is common and should be considered with GI distress.
What did you find out about ensuring that children get adequate nutrition on a GFCF diet?
Dr. Hyman: Our study suggests that when changing the diet of children with autism, it’s important to give careful attention to nutritional sufficiency. Although all the children in our study consumed enough calories for growth, some needed additional vitamins or iron to keep their micronutrient intake in the recommended range. And most of these families needed counseling from the dietitian about healthy food choices.
The good news is that in another study, one enrolling children from five Autism Speaks Autism Treatment Network (ATN) sites, we found that children on GFCF diets fare no worse than other children with autism so long as they take supplemental calcium and vitamin D.
AS: You’ve mentioned that you encourage families to work with a nutritionist if they are going to try a GFCF diet. For families who can’t afford to do so, what recommendations can you offer?
Dr. Hyman: The GFCF diet removes all dairy products. So caregivers need to make sure that these children age-appropriate amounts of calcium and vitamin D from other sources. As I mentioned, our ATN study on diet and nutrition found no difference in nutrition between children on a regular diet and those on a GFCF diet. However, we found that most of the children in that study were not getting enough calcium or vitamin D regardless of whether they were on the GFCF diet or not. Even those who were getting supplemental calcium and vitamin D were not getting enough.
In terms of affording the services of a dietician, many children with autism under age 3 are able to see a registered dietitian through their Early Intervention Program. Also a family’s pediatrician can provide nutritional guidance and refer to a pediatric registered dietitian if needed.
For more on autism and GFCF diets, also see: