Do therapeutic diets for autism work? Also called special diets or elimination diets, therapeutic diets for autism are popular, and many families try them. Hands down, correctly implemented and especially with professional guidance, I saw these diets turn in miraculous results over and over again in my pediatric nutrition practice. But what about scientific evidence to back them up? If therapeutic diets for autism work, why hasn’t your pediatrician told you to jump in? Where’s the referral for the right resources?
Therapeutic diets for autism may be popular with parents, but they are not so much with the pharmaceutical industry. Using foods and supplements as therapy generally means using fewer prescription drugs. And since prescription drugs are what your doctor is trained to use, physicians don’t know much about how to guide you on using a therapeutic diet for autism, much less on how to assess what nutrition components your child may need. In my pediatric nutrition practice, it was clear that there is no one-size-fits-all therapeutic diet for autism. The best outcomes were had with individualized assessment and plan, based on each child. This is exactly how any health care is best delivered, and nutrition is no exception. No surprises there.
You may have heard of a study from the University of Rochester that claimed to debunk whether special diets – specifically the gluten and casein free (GF CF) diet – can help children with autism. This is one of those studies that gets legs and is off and running. Well, it got under my skin. Not only was this study on therapeutic diets for autism diagnoses poorly designed (in spite of the researchers calling it “the most controlled to date”) – turns out, I was born in the hospital where it was done! The challenge was on to drill down and find out exactly how the researchers approached the therapeutic-diets-for-autism question.
The study’s flaws are many. In fact, I couldn’t help but notice that it appeared designed to bury or defeat a beneficial effect from using a special diet measure for an autism diagnosis. The take-away from this work was essentially that “special diets don’t help children with autism” – when it should be “therapeutic diets for autism can help many kids, but we can’t know which ones unless they are screened for nutrition and GI problems.” The trouble is, many kids on the autism spectrum never get that screening – bringing to mind the old adage, if you don’t look, the problem isn’t there.
The Rochester study missed the mark. Here’s what went wrong:
It was too small – Only 14 children completed the trial. By contrast, remember Andrew Wakefield, the gastroenterologist skewered in 1998 for reporting that in a series of his pediatric patients autism, his team found a unique type of inflammatory bowel disease? He was roundly criticized for reporting an effect in only 12 children, in the now famously-retracted case series published in the Lancet. But, the scientific process prevails, and we can’t have it both ways: If it’s invalid to report an effect because you saw it in only 12 children, then it’s also invalid to report there can be no effect if you only studied 14 children. Both outcomes deserve deeper investigation. Indeed, since the Wakefield fiasco and the publication of the Rochester study, others have documented findings of gastrointestinal disorders (and many other co-morbidities) in children with autism to the point where this is now widely accepted. The value of using nutrition interventions in GI disorders is well established, so to dismiss the option of therapeutic diets for autism is short sighted.
Children with gastrointestinal symptoms were excluded – Lead author Susan Hyman MD acknowledges this flaw: “The study didn’t include children with significant gastrointestinal disease. It’s possible those children and other specific groups might see a benefit.” That’s a full stop for me. It essentially means that this study actually left out kids with perhaps the most to gain from a therapeutic diet approach. Given the fact that so many children with autism have co-morbid GI conditions, it’s critical to study the effect of nutrition interventions for them.
Was soy allowed? Nobody knows – No mention is made of what protein sources replaced gluten and casein in this study that reviewed the use of GF CF diet. If soy was allowed, this is a sure fail: Just like gluten and casein, soy protein and pea protein concentrates can form opiate-like peptides in the context of weak digestion, which many children on the spectrum have. These peptides can cause bloating, constipation, or painful inflammatory reactions. They can keep expressive language stunted and alter behavior profoundly, effects noted and linked to autism decades ago. Using it as a daily protein replacement (soy milk, soy yogurt, tofu, edamame or other soy foods) will erase benefits of removing gluten and casein.
It was too short – Study duration on GF CF regimen was only “four to six” (which was it?) weeks – too short to notice a significant effect from removing gluten or casein. In my pediatric nutrition practice, I recommend a three month trial at minimum, to judge efficacy of this measure. This is because it can take three to six months for antibodiy reactions to food proteins to drop, once they are removed from the diet. It may also take several weeks for diet-sourced opiate peptides from foods to drop also. Even though the authors cut the trial this short, they did report a positive effect, but it did not reach statistical significance. With a more appropriate length trial, a statistically significant effect may have been seen. Typically, in my clinical practice, benefits are just starting to emerge at about week three or four of a GF CF trial.
Nutrition assessment was partial and weak – A weak pre-trial nutrition assessment was done. It screened for iron status, vitamin D level, and IgE food allergy to wheat or dairy. It left out screening for IgG food sensitivity reactions (which may drive anxiety, inattention, tactile hypersensitivity), gut dysbiosis, other gut health concerns, other mineral or vitamin status, or growth impairments. There was no post-trial nutrition assessment, and there’s no mention of standards like body mass index or growth parameters. This means we have no idea what these kids’ nutrition diagnoses actually were at the start. We also never learn if their nutrition status improved during the trial, if the GF CF diet was even indicated for them, or if any children’s nutrition status weakened during the trial. Many nutrition problems are documented in children with autism, including low body mass index, growth failure, low ferritin, other vitamin and mineral deficiencies, low essential fatty acids and omega 3 fatty acids, bowel infections, and so on. The authors had an opportunity to get a good work up on these kids pre and post trial, but passed on it – meaning that there may have been undiagnosed, untreated nutrition impairments that an elimination diet can’t fix by itself.
Was the food truly “blinded”? – Though the authors made an effort to “blind” the food, it’s a stretch to think that food can indeed be blinded – especially for children with autism. These tend to be kids with extremely picky appetites and a fierce radar for ingredients and textures they either want, or hate. But we’re left in the dark on how successful that was for families in the study.
The study errs in its very premise – It’s hard to effectively test a single diet measure as a treatment for autism because nutrition itself is a symphony, not a solo performance! In other words, foods, individual nutrients, and gut biome work together and affect each other profoundly. A GF CF therapeutic diet for autism may work great if you also correct that child’s iron deficit or gut dysbiosis; it may be just meh if you leave the mineral status and dysbiosis festering. In my experience, kids with autism can vary for their nutrition and GI concerns, and respond best to individualized support based on their work up, not a single diet dogma.
It’s a fail. In short, the study authors picked children with autism who were least likely to respond, tested them too briefly, and permitted confounders like opiate forming peptides, other nutrition or growth impairments, or GI problems to carry on. This design is akin to randomly choosing a group of kids with autism with widely varying and unassessed differences in behavior or functional ability, giving all of them Ability for two weeks, and then, when results are lackluster, declaring all psychiatric medications to be of no value to children with autism.
Therapeutic diets for autism can yield truly miraculous results. Engage your provider team by starting with my E-Book Five Essential Lab Tests For Kids With Autism – to define the workable pieces most likely to help your child, and to learn short cuts to get nutrition assessment tests done correctly, quickly, and at lower cost. Here’s to your kids’ health, wellness, and happiness!