Paper·2018·Clinical trial
LandmarkFolinic acid improves verbal communication in children with autism and language impairment: a randomized double-blind placebo-controlled trial
In brief
- This was the first trial of leucovorin in autism in which neither families nor researchers knew who was getting the medicine. It followed 48 children for 12 weeks.
- Children who took leucovorin improved more in verbal communication than children who took a placebo.
- Leucovorin is a prescription medicine, not an “autism pill”. It needs the right patient, dose and formulation. Talk to your child’s doctor.
- A randomized, double-blind, placebo-controlled trial. 48 children aged 3 to 14: 23 assigned leucovorin (2 mg/kg per day, maximum 50 mg) and 25 assigned placebo, for 12 weeks.
- Primary outcome: verbal communication. ADTF’s statement reports significant improvement with leucovorin compared with placebo.
- Number needed to treat of 1.8 in children positive for folate receptor autoantibodies. One trial; 43 of 48 children finished. Not an “autism pill”: patient, dose and formulation matter.
Written from the published paper and the foundation’s statement of 1 October 2025.
What did the study ask?
is a prescription form of folate, a B vitamin. Doctors have used it for more than 80 years.
An earlier study by Dr. Frye’s team had found that children with autism improved while taking it. But that study had no group, and everyone knew who was being treated. In a study like that, hope and expectation can look like improvement.
So this trial asked a stricter question. Do children with autism and language impairment improve more in verbal communication on leucovorin than on a placebo? It was the first trial of leucovorin in autism to be run this way.
Open-label work with a waitlist control, published in Molecular Psychiatry in 2013, had reported a positive response to leucovorin (folinic acid) in about two-thirds of treated children with autism spectrum disorder (ASD).
This trial tested the same treatment under randomized, double-blind, placebo-controlled conditions, in children with ASD and language impairment. It was the first trial of leucovorin in ASD with that design. The primary outcome measure was verbal communication.
What did the researchers do?
The trial enrolled 48 children with autism and language impairment. They were 3 to 14 years old.
Each child was assigned at random to one of two groups. In one group, 23 children were given leucovorin. In the other, 25 children were given a placebo. Both groups took it every day for 12 weeks.
The trial was : neither the families nor the researchers knew which child was in which group. The main thing the researchers measured was verbal communication.
Children with ASD and language impairment were randomized to leucovorin or placebo for 12 weeks. Families and investigators were blind to the allocation. The primary outcome measure was verbal communication.
| Detail | Leucovorin | Placebo |
|---|---|---|
| Children assigned | 23 | 25 |
| Age | 3 to 14 years | 3 to 14 years |
| Dose | 2 mg/kg per day, maximum 50 mg | Placebo |
| Length | 12 weeks | 12 weeks |
Of the 48 children, 43 finished and 5 dropped out. The statement does not say how the 5 were divided between the arms.
23 assigned leucovorin
25 assigned a placebo
What did they find?
Over the 12 weeks, verbal communication improved more in the children who took leucovorin than in the children who took the placebo.
The foundation’s statement also reports improvement, compared with placebo, in other symptoms. It names repetitive behavior, social withdrawal and hyperactivity.
Some children with autism have that block folate from reaching the brain. A blood test can find them. According to the statement, children who tested positive were more likely to respond to leucovorin.
The statement gives response rates for this trial. We keep them in the science view, beside what they can and cannot show.
On the primary outcome, the statement reports “significant improvements in language” with leucovorin compared with placebo. It reports improvement in other symptoms as well, including stereotyped behavior, social withdrawal and hyperactivity.
| Measure | In the statement |
|---|---|
| Responded to treatment, “regardless of FRAA status” | 65% |
| Responded “if they were FRAA-positive” | 77% |
| Response rate in the placebo group | Not given |
| Number needed to treat, FRAA-positive children | 1.8, for “significant improvement in language over 12 weeks” |
| Effect size | Described as “large”. No value given. |
Read the two response rates with care. Without the placebo group’s rate beside them, they do not show the size of the effect. A number needed to treat does take the placebo group into account, but this one applies to the FRAA-positive subgroup, and the statement does not give that subgroup’s size.
These are the results of one 12-week trial in 48 children. For the scale used to measure verbal communication, confidence intervals and adverse events, go to the paper (opens the journal site).
What does this mean for families?What does this mean in practice?
Leucovorin is not an “autism pill”
It is a prescription medicine. For it to work, a doctor has to choose the right patient, the right dose and the right formulation.
This trial is a reason to ask your child’s doctor a question. It is not a reason to start a treatment on your own.
The foundation cautions parents against swapping leucovorin for a supplement. Folinic acid sold over the counter is usually measured in micrograms. Leucovorin is prescribed in milligrams, and one milligram is 1,000 micrograms. Folic acid is a different substance, and it is not a substitute.
If you think this could apply to your child, talk to your child’s doctor. You can print a one-page sheet to bring to the appointment.
The statement says prescribers need to know which patients are most likely to benefit, how to prescribe, and which formulations suit children with ASD. It lists points that matter in practice:
- Folate receptor autoantibodies can be detected with a blood test. The statement calls them a more accessible biomarker than cerebrospinal fluid folate, which needs a lumbar puncture.
- Some antiepileptic drugs alter folate metabolism. Medications need to be coordinated and interactions monitored. The same caution applies to methotrexate and other antifolates.
- Over-the-counter folinic acid is dosed in micrograms and is not regulated as a medicine. Folic acid is metabolized differently and is not a substitute.
- Generic formulations differ in their inactive ingredients, and some are poorly tolerated. A pharmacy can switch brands without telling the family.
What can this study not tell us?
This is one trial. It included 48 children and lasted 12 weeks, and 43 of the children finished it.
A trial of this size and length cannot tell us what happens after 12 weeks. It cannot tell us which child will respond, or what dose is best for each child.
The foundation says it supports larger and longer studies of leucovorin, and that it is running them.
- One trial, with 48 children assigned and 43 finishing.
- 12 weeks of treatment. A trial of that length cannot show whether a response lasts.
- The FRAA-positive findings rest on a subgroup, and the statement does not give its size.
- The statement calls the effect size large and gives no value or interval for it.
The statement says ADTF supports, and is conducting, larger and longer studies of leucovorin. Among its stated goals are to identify which children are most likely to benefit, and to determine the best dose for each patient.
What has happened since?
The statement notes that Autism Speaks rated this trial the second most important autism research finding of its year.
Since then, independent research groups in 4 other countries have run their own blinded, placebo-controlled trials. The statement says they have largely replicated the result.
Counting this one, that makes 5 blinded trials and 293 children.
The statement lists the randomized, blinded, placebo-controlled trials that followed, each from an independent group: Renard et al., Biochimie, 2020; Batebi et al., Child Psychiatry & Human Development, 2021, with risperidone in both arms; Panda et al., European Journal of Pediatrics, 2024, over 24 weeks; and Zhang et al., Nutrients, 2025.
It describes them as having “largely replicated and validated” the result of this trial. With this one, that is 5 blinded, controlled trials and 293 children. The statement also notes that Autism Speaks rated this trial the second most important autism research finding of its year.
2018
12 weeks
The first · Dr. FryeFrye et al. Molecular Psychiatry
2020
12 weeks
Renard et al. Biochimie
2021
10 weeks
Batebi et al. Child Psychiatry & Human Development
2024
24 weeks
Panda et al. European Journal of Pediatrics
2025
12 weeks
Zhang et al. Nutrients
Assigned leucovorin, 155Assigned placebo, 138