Despite the widespread popularity of complementary, alternative, or integrative medicine (CAIM) for the treatment of autism spectrum disorder (ASD), there is minimal evidence that they are worthwhile.

A large umbrella review of 19 common alternative therapies for autism showed that even those therapies showing some benefit lacked strong evidence of safety and/or efficacy.

“The key takeaway is that the evidence supporting the effects of CAIM in autism remains highly uncertain. This uncertainty is particularly critical when it comes to safety as most interventions have not been adequately evaluated for potential risks or adverse effects,” study investigator, Corentin J. Gosling, PhD, a lecturer in neuropsychiatry in the DysCo Laboratory, Paris Nanterre University in Nanterre, France, told Medscape Medical News. 

The study was published online on August 28 in Nature Human Behaviour.

Widespread Use

Despite limited and inconsistent evidence of efficacy, CAIM use among individuals with ASD is widespread, with 90% reporting they have tried it at least once, the researchers noted.

A 2017 systematic review revealed prevalence of any CAIM use among people with ASD ranged from 28% to 95%, with a median prevalence of 54%. A generally positive public attitude toward the safety and efficacy of CAIM treatments may explain this widespread use, the authors noted.

Personal beliefs, including doubts about the efficacy of conventional therapies, may also contribute to CAIM’s popularity, the investigators noted.

Furthermore, previous reviews often focused on specific interventions and lacked comprehensive quality assessments, leaving clinicians and families with conflicting information.

At the same time, ASD treatment guidelines for children and teenagers from the National Institute for Health and Care Excellence in the UK expressly state what not to use. For example: “Do not use omega‑3 fatty acids to manage sleep problems in autistic children and young people. Do not use the following interventions to manage autism in any context in children and young people: secretin, chelation, or hyperbaric oxygen therapy.”

Wide-Ranging Research

To address these knowledge gaps, the researchers conducted an umbrella review of meta-analyses covering 19 common CAIM interventions to evaluate both the safety and efficacy to create an accessible online resource to help clinicians, individuals with ASD, and their families to make informed decisions.

“Rather than looking at individual trials, we reviewed all the available meta-analyses, which are a compilation of many trials. This allowed us to evaluate the full body of evidence across different treatments,” Gosling said in a release.

For the review, the researchers searched five databases to identify systematic reviews with meta-analyses published through December 2023. They assessed 248 meta-analyses encompassing 200 clinical trials and more than 10,000 participants.

After removing overlapping meta-analyses, the primary analysis of core ASD symptoms included 138 studies. Of these, 18% focused on very young children, 62% on school-aged children, 12% on adolescents, and 8% on adults.

Of all of the CAIM interventions, oxytocin was the one with the strongest evidence base, as determined by the highest Grading of Recommendations Assessment, Development, and Evaluation rating. Even so the evidence for core symptom efficacy was “negligible and not statistically significant,” the researchers noted.

The investigators added that a small effect size was observed with oxytocin in adults for restricted/repetitive behaviors and there were no safety concerns.

Low-Quality Evidence

Gosling explained that oxytocin has been studied in autism because of its role in social bonding and social communication in people and because of documented lower levels of oxytocin children with ASD. “Some researchers hypothesized that it might help support aspects of social interaction for autistic people,” he said.

The researchers also evaluated secondary outcomes, including sleep quality and duration, quality of life, anxiety, and language skills. Some interventions showed large effect sizes with statistical significance, but the evidence was low quality.

Gosling said CAIM research in ASD is often based on a limited number of studies, many of which have methodological shortcomings. “As a result, even when some studies suggest potential effects, the overall certainty of evidence remains low.”

“We see this as an important finding in itself, given how widely these interventions are used,” a dded Gosling, who is also a visiting professor in child and adolescent psychiatry at the Robert Debré Hospital AP-HP in Paris, France, and at the Centre for Innovation in Mental Health, School of Psychology, University of Southampton in Southampton, England.

“Our hope is that [this research] highlights the urgent need for more rigorous, higher-quality research in this field,” Gosling said.

In addition, the researchers created the open-access Evidence-Based Interventions for Autism: Clinical Trial online database to highlight these and future findings. The goal is to make it easier for individuals with autism, their families, and healthcare providers to assess the evidence behind each alternative treatment. The investigators plan to update this resource annually for the next 5 years and as warranted to support informed decision-making.

Caution Warranted

Commenting on the research for Medscape Medical News, Douglas Vanderbilt, MD, MS, MBA, division chief of developmental-behavioral pediatrics at Children’s Hospital Los Angeles, Los Angeles, said ASD is not one genetic disorder and presents across a “huge phenotypic range.” 

Therefore, the umbrella review lacked some precision in inclusion criteria for those with ASD.

“Things that work for one child may not work for another no matter how well-designed a study is,” Vanderbilt said.

The low level of evidence for efficacy and lack of safety parameters indicates the need for well-designed randomized, controlled trials, he added.

“Due to the lack of evidence, providers should be cautious in recommending any of these CAIMs, especially in children,” he said.

Also commenting for Medscape Medical News, Peter J. Chung, MD, developmental-behavioral pediatrician and associate clinical professor of pediatrics and medical director at The Center for Autism & Neurodevelopmental Disorders at University of California, Irvine, California, described the research as “an excellent analysis of a very broad topic that synthesized a very large amount of data.”

“Research in autism spectrum disorders is notably challenging due to the significant heterogeneity of the condition, so it is important for clinicians and parents to understand the many facets to consider when exploring CAIM,” said Chung.

The accompanying online resource clearly summarizes the available literature with an option to drill down on the specific studies that factor into the summary score, he added.

“There is certainly much more work to do in the study of CAIM efficacy and risk,” Chung added and suggested the authors include leucovorin, cannabinoids, and neurobio feedback in future updates.

This study was independently supported. Gosling, Chung, and Vanderbilt reported no relevant financial relationships.