Last winter, Brian Noonan read online that some doctors were prescribing an obscure drug, typically given to cancer patients, for autism. Curious, he looked into it for his son Benjamin, who had just been diagnosed with autism in October.

“We jumped on it,” Noonan said. “It felt right and it made sense.”

The medication was leucovorin, also called folinic acid. It’s a synthetic form of vitamin B9 or folate, which the body needs to make healthy blood cells. During pregnancy, folate is important to reduce the risk of birth defects.

Last Monday, the drug was thrust into the national spotlight by President Donald Trump in a rambling press conference about autism that mainly focused on the president and Health Secretary Robert F. Kennedy’s claim that taking acetaminophen during pregnancy could cause autism. Trump briefly mentioned that an “existing drug” — referring to leucovorin — may help with certain symptoms of autism.

For those in the autism community, leucovorin is not new. Dr. Richard Frye, a behavioral child neurologist in Phoenix has studied leucovorin and autism for two decades. He made clear that it is not a cure for autism and that more research is needed. He does, however, prescribe it to some people with autism.

Noonan’s son Benjamin, who is 4 years old, is one of Frye’s patients.

“He’s higher functioning,” Noonan said. “He’s verbal, but he really struggled to put together sentences.” Since starting on the medicine, the family believes Benjamin’s speech has improved, though he still has difficulties with hyperactivity and impulsiveness.

Still, Noonan added, he’s under no impression that the drug is a miracle pill. Benjamin also attends a behavioral program preparing him for kindergarten, Noonan said, and he plans to enroll him in speech and occupational therapy. “We very, very much understand we’re still taking an experimental medication,” he said.

Other families say the drug conferred larger benefits.

Ben Blomgren, of Minneapolis, said his 11-year-old son, Josh, had been prescribed leucovorin off-label in February after he was at risk of being kicked out of school, even as they tried behavioral modification methods.

“He’s pretty severe,” Blomgren said. “He didn’t have any language skills. He was not toilet trained.”

After starting the medication, Blomgren said Josh’s sleep improved, he’s fully toilet trained and he’s stopped running away from them.

“It wasn’t overnight, but we saw major improvement,” he said.

Yomarie Miranda, of Florida, said she saw improvements in her 7-year-old son Ethan after he started the medication, including following instructions in class.

Ethan was prescribed the medication off label earlier this month, she said.

“He’s now looking at me when I speak with him. He’s talking more than before with complete sentences,” she said.

A highly unusual move

The Food and Drug ministration first approved leucovorin in the 1980s to help reduce the toxic side effects of certain chemotherapy drugs.

In the decades since, researchers have also studied whether it might treat cerebral folate deficiency (CFD), a neurological condition that makes it harder for folate to reach the brain. Some children with autism also have cerebral folate deficiency, which neurologists say can affect speech, mood and behavior. Last Monday, the FDA said it planned to update the drug’s label to include that use.

But the agency’s decision, experts say, rests on just a handful of small studies. And despite Trump’s endorsement, the maker of the brand-name version, GSK, has yet to submit an application to the agency to change the label. In a statement, GSK said it will be submitting the request to expand the approval to include cerebral folate deficiency, though the statement does not mention autism. (Because leucovorin is already an FDA-approved drug, doctors have been able to prescribe it “off label” for other uses, though insurance may not cover it.)

It’s a highly unusual move for the FDA: pushing a drug for chemotherapy side effects as a therapy for autism without a formal submission and limited evidence.

“It’s incredibly irresponsible,” said Dr. Aaron Kesselheim, a professor of medicine at Harvard Medical School. “For 60-plus years, we’ve counted on the FDA to help patients distinguish between products that work and products that don’t work. And here we saw the FDA making an announcement relying on a summary of unclear data and announcing that they had already decided to approve it.”

Alycia Halladay, chief science officer at the nonprofit Autism Science Foundation, was frustrated by the implication that this is a breakthrough in autism research.

“It’s not like scientists have just been staring at their belly buttons for 20 years, not looking at autism treatments,” Halladay said in an interview earlier this week. “They have, but the standards have been very high to get [treatments] approved.”

Andrew Nixon, a spokesperson for the Department of Health and Human Services, which oversees the FDA, said in a statement: “Analysis across 23 publications from 2009-2024 demonstrated the effectiveness for CFD. Overall, 85% of patients experienced some type of clinical benefit including improved speech/communication capabilities.” Nixon’s statement did not mention autism, and he did not respond to follow-up questions.

David Mandell, a psychiatry professor at the University of Pennsylvania, said HHS’ claim that 85% of patients experienced a benefit is a “quite a conceptual leap” because it assumes people with cerebral folate deficiency also have autism.

Cerebral folate deficiency “is an extremely rare event, and it is accompanied by symptoms of epilepsy, really severe neurodevelopmental problems and some of those symptoms can look like autism, but it’s not,” he said.

“I could not think of a single FDA approval that has such weak evidence in the past 20 years,” Mandell added.

Frye, the Phoenix neurologist, estimates up to three-quarters of autistic children have antibodies linked to cerebral folate deficiency, based on his research. Other estimates are much lower: Alycia Halladay, chief science officer at the nonprofit Autism Science Foundation, put the rates at roughly 10% to 30%.

The studies on leucovorin are small but look interesting, said Dr. Scott Gottlieb, a former FDA commissioner.

“It certainly merits follow up,” Gottlieb said.

One of the studies was conducted by Frye: a placebo-controlled trial of more than 40 children with the antibodies, published in Molecular Psychiatry in 2018. It found around two-thirds of kids who got the drug saw improvements in verbal communication after 12 weeks. Others saw no changes. The children all had language impairments, as well as a common type of autism that’s not linked to other neurological diseases.

Mandell, however, said the results shouldn’t be taken as evidence.

“They claim that in certain subgroups of kids in their already small sample, they find these very large effects,” he said, adding that larger-scale trials that establish clear outcomes ahead of time are needed to verify the results.

Kesselheim said it’s important for patients to have access to medications that could have a benefit, especially when there is an unmet medical need. But, he said, there are still a number of unanswered questions.

“There is no accurate testing for this,” he said. “What are low folate levels? What are normal folate levels? All of that stuff should be worked out.”

Frye said he currently uses two methods to assess whether a child has a folate deficiency: a spinal tap — also known as a lumbar puncture — which involves inserting a needle into the lower back to collect fluid from the spinal cord, and a folate receptor antibody test originally developed for pregnant women that isn’t approved by the FDA for diagnosing folate deficiency in children.

Edward Quadros, a research professor at SUNY Downstate Health Sciences University who has worked with Frye studying leucovorin for autism, said parents are already trying the drug, including supplement versions, which aren’t regulated and can be dangerous.

“By making the FDA approve it, and reputable pharmaceutical companies manufacturing it and selling it, at least it gives you quality dosing,” Quadros said.