The largest study to date of leucovorin’s effectiveness for treating autism traits has been retracted because of data inconsistencies and statistical issues, according to a notice posted last week by the European Journal of Pediatrics.
The study included 77 autistic children and is one of only five randomized clinical trials that have tested leucovorin, also known as folinic acid, in autistic people.
“The retraction of this paper removes a significant portion of the already weak evidence supporting the value of folinic acid as a treatment for autism,” Thomas Challman, a pediatrician at Geisinger College of Health Sciences who specializes in neurodevelopmental conditions, wrote in an email to The Transmitter. “Until we have acceptable evidence of safety and effectiveness, folinic acid use as a treatment for autism is not appropriate outside of a well-designed clinical trial.”
The retracted study, published in September 2024, claimed that 24 weeks of daily treatment with oral folinic acid reduced symptom severity in children with autism as compared with a placebo. But some of the numbers in the data tables did not add up correctly, Challman and Scott Myers, a pediatrician and professor of developmental medicine at Geisinger College of Health Sciences, found when they looked closely at the data. They posted their concerns on PubPeer in September 2025 and contacted the journal.
In response to these and other PubPeer comments, the journal reached out to the authors, who identified several errors in their data, according to the retraction notice. A review by the journal then “confirmed several of the concerns raised with the data and statistical analysis and was unable to replicate the results reported in the article from the dataset provided,” the notice stated. “The [e]ditor therefore no longer has confidence in the validity of the results and conclusions reported in this article.”
Two of the study’s six authors agree with the retraction, and the other four did not respond, according to the journal’s notice. “There were some unintentional statistical analysis errors,” study author Prateek Kumar Panda, a pediatric neurologist at All India Institute of Medical Sciences, wrote in an email to The Transmitter.
L
Folate receptor autoantibodies, similar to those found in people with cerebral folate deficiency, may be more prevalent in autism and may play a significant role in the neuropathogenesis of the condition, says Shafali Jeste, chair of pediatrics at the University of California, Los Angeles. However, “neither of those statements has been actually proven, so we’re already working under a premise that is not evidence based.”
The four other studies that have looked at the use of folinic acid in autism have received many criticisms, including that they were not blinded properly and had small sample sizes, says Dorothy Bishop, emeritus professor of developmental neuropsychology at the University of Oxford. “The quality of the research is uniformly poor.”
As for the retracted study, “the statistics were all over the place,” Bishop says.
In addition, the clinical relevance of the study’s outcome measure, the Childhood Autism Rating Scale (CARS) score, is unclear, according to a correspondence published in response to the article in May 2025. CARS, Jeste says, is meant to help clinicians confirm a diagnosis based on a list of symptoms, “not as a quantitative outcome.”
Despite the retraction, the journal offered the authors an opportunity to address the concerns and submit a revised version of their article, which the authors agreed to do. “We are preparing the revised manuscript after rectifying those [concerns] and will submit in the same journal shortly,” Panda wrote in his email to The Transmitter.
Still, given the lack of robust scientific evidence, other researchers are not expecting studies using leucovorin in people with autism to pan out, Jeste says. “I think the reason this specific drug has been of such interest is because of the publicity around it, not because we as scientists believe that this is the drug that we should be testing.”
