- The largest trial, 180 children at Duke University, found the infusions safe but no improvement in social communication, autism symptoms or vocabulary.
- A 2026 trial in 34 children selected by immune markers found a small effect that its own authors call preliminary.
- No regulator has approved a cell therapy for autism. Developmental, behavioural and educational support remains the evidence-based care.
Families of autistic children get offered a great many things. This page is about what the controlled trials actually found — including the times they found nothing.
One thing should be said before any of it. Autism is a neurodevelopmental condition, not a disease waiting for a cure. The support with the strongest evidence behind it remains developmental, behavioural and educational, and nothing on this page replaces it.
What the trials found
The largest and best-designed study to date was run at Duke University. It was a phase II randomised, double-blind, placebo-controlled trial in 180 children aged two to seven, who received a single intravenous infusion of their own cord blood, donor cord blood, or placebo, and were assessed six months later (Dawson et al., 2020).
The infusions were safe and well tolerated. But across the whole group there was no improvement in the main outcome — social communication — and none in autism symptoms or vocabulary either. The authors state it plainly: a single cord blood infusion was not associated with improved socialisation or reduced autism symptoms.
An earlier crossover trial had pointed the same way. Twenty-nine children received their own cord blood and placebo in turn, and while there were trends toward improvement, particularly in socialisation, there were no statistically significant differences for any endpoint (Chez et al., 2018).
So two controlled trials, one of them large, looked for the effect and did not find it. That is the honest starting point, and it is the part most often left out.
The signal still being studied
A 2026 trial took a different approach. Instead of enrolling any child with autism, it enrolled 34 children who had autism and measurable immune dysregulation, and gave four infusions of donor cord blood mononuclear cells alongside the rehabilitation they were already receiving (Zhang et al., 2026).
This time the treated group did show a greater reduction in the main social responsiveness score (difference −6.03, 95% CI −11.93 to −0.14, p = .045) and a clearer change in social cognition (−9.95, p = .002), with no major transplant-related adverse events.
What makes it interesting is the selection: it chose patients by a biological marker rather than treating everyone and hoping.
What keeps it from being a basis for treatment is everything else. Thirty-four children, one centre, thirteen weeks of follow-up, and a confidence interval on the main result that very nearly touches zero. The authors call their own findings preliminary, and they are right to.
One more detail worth knowing. In the Duke trial, a subgroup of children without intellectual disability did show improvement in communication skills — but the overall measure of clinical improvement in that subgroup was not statistically significant. A promising thread is not a result.
Before your family considers it
What is consistent across all of this is safety. The infusions were well tolerated in every trial. But safe and effective are different claims, and the gap between them is exactly where most of the marketing lives.
As things stand, no regulator has approved a cell therapy for autism, and the published trials do not support offering one as treatment. Families are frequently told otherwise, often at considerable cost and sometimes far from home.
If someone offers this to your family, three questions are enough: which trial supports it, in which children specifically, and what happened to that trial’s primary outcome? The answers will tell you most of what you need to know.
- Keep developmental, behavioural and educational support at the centre: it has the strongest evidence.
- If a clinic offers cell therapy, ask which trial supports it, in which children, and what happened to its primary outcome.
- Treat a safety record as a safety record. It is not evidence that a treatment works.
Two placebo-controlled trials, one of them with 180 children, found the infusions safe but did not find an improvement in social communication or autism symptoms. A small 2026 trial in children selected by immune markers reported a modest effect that its authors call preliminary. No regulator has approved a cell therapy for autism.
- A phase II randomized clinical trial of the safety and efficacy of intravenous umbilical cord blood infusion for treatment of children with autism spectrum disorder · The Journal of Pediatrics, 2020
- Safety and observations from a placebo-controlled, crossover study to assess use of autologous umbilical cord blood stem cells to improve symptoms in children with autism · Stem Cells Translational Medicine, 2018
- Efficacy and safety of allogeneic umbilical cord blood mononuclear cells therapy in children with autism spectrum disorder and immune dysregulation · Stem Cell Research & Therapy, 2026
Questions
Does stem cell therapy improve autism?
The largest controlled trial, with 180 children, found no improvement in social communication, autism symptoms or vocabulary after a single cord blood infusion. A smaller crossover trial found no significant differences either.
Is it safe?
The infusions were well tolerated in every published trial. Safety is not the same as benefit, and the trials did not show a benefit across the children studied.
Is any cell therapy approved for autism?
No. No regulator has approved a cell therapy for autism, and the published trials do not support offering one as treatment.
What should I ask a clinic that offers it?
Which trial supports it, in which children specifically, and what happened to that trial’s primary outcome.