Emerging therapy

Probiotics, prebiotics and the gut–brain axis

Lactobacillus, Bifidobacterium, synbiotics, fibre supplements, faecal microbiota transfer

Genuinely interesting science with a few solid, specific uses in children — and a large gap between those uses and what is sold to families of autistic and neurologically disabled children.

Where it stands, condition by condition

The same therapy can be well supported for one problem and completely untested for another. This is the single most common place families are misled.

What it is

Probiotics are live micro-organisms given in the hope of a health benefit. Prebiotics are fibres that feed the bacteria already present. Synbiotics combine the two. Faecal microbiota transfer moves a whole donor community into the recipient.

The single most important thing to understand is that “probiotic” is not one thing. Effects are strain-specific, and often dose-specific. Evidence for Lactobacillus rhamnosus GG in one situation tells you nothing about a different strain in a different situation — yet products are marketed as though benefit belonged to the category.

Most products sold to families are food supplements. They do not have to prove benefit, and independent analyses regularly find fewer live organisms than the label claims, or different species altogether.

The gut–brain axis is real. The gut has its own nervous system, communicates with the brain through the vagus nerve, immune signalling and bacterial metabolites, and gut bacteria are different in some neurological conditions. None of that is in dispute. What is in dispute is whether swallowing a capsule changes a child's development.

How it is meant to work

Proposed routes: producing short-chain fatty acids that affect immune and brain signalling, strengthening the gut barrier so fewer bacterial products enter circulation, altering vagal signalling, and shifting inflammation.

The honest difficulty is direction. Autistic children do, on average, have different gut bacteria — but they also, on average, eat very differently. A restricted diet changes the microbiome; the microbiome does not obviously change the diet. Much of the published association can be explained that way, and disentangling the two is exactly what the field has not yet done.

A second difficulty: swallowed organisms largely pass through. Lasting colonisation is uncommon, which is why effects stop when the product stops.

What has actually been tested

• Where the evidence is good, it is narrow. Specific strains for antibiotic-associated diarrhoea, and probiotics for necrotising enterocolitis in very preterm infants, are the two areas where randomised evidence is substantial. Both are short-term, well-defined and have nothing to do with development.
• Where it is weakest is exactly where it is sold hardest. For core autistic features, individual randomised trials are small and generally null on their primary outcome — including a recent pilot in young autistic children where neither microbial composition nor diversity shifted detectably, and the only positive signal came from a secondary analysis in 23 children.
• The 2025 meta-analysis that is often quoted as proof pooled 15 randomised trials across autism and ADHD and found a statistically significant but small effect of about −0.12. It also pooled probiotics with prebiotics, dietary change and faecal transfer, which are not the same intervention.
• On faecal microbiota transfer, the striking autism results that circulate came from an open-label study of a small number of children with no control group, followed up by the same team. That is a reason to run a randomised trial, not a reason to treat a child. Those trials are running now.
• Gut symptoms are a different and more defensible target. Constipation, reflux and discomfort are common in autistic children and in children with cerebral palsy, they affect sleep, behaviour and school, and they deserve treating in their own right — whether or not the treatment is a probiotic.

What we still do not know

• Which strain, which dose, for how long, for which child — almost none of this is settled, even where the general category has evidence.
• Whether any change in the microbiome is a cause of a child's difficulties or a consequence of their diet, medication and activity.
• Whether faecal transfer has a real effect in autism, which the running randomised trials should begin to answer.
• What is actually in a given supplement, since live organism counts at the point of sale frequently do not match the label.

Risks and costs

• In a healthy child, ordinary probiotics are generally well tolerated; bloating and wind are the usual complaints.
• The serious exception is the fragile child. In very preterm infants, immunosuppressed children, and children with a central line or short bowel, live organisms can cause bloodstream infection. In 2023 the FDA warned hospitals after a preterm infant died of sepsis caused by the Bifidobacterium in the probiotic product they had been given — confirmed by genetic matching — and a product recall followed. European neonatal groups have since argued the benefit still outweighs the risk with good-quality products, but the point stands: this is a medicine-level decision in a fragile baby, not a supplement choice.
• Faecal transfer carries transmission risk and is a procedure, not a product.
• Supplement quality: contamination has caused serious infection, and in one case death in a preterm infant.
• Cost and distraction, as everywhere — a monthly purchase that feels productive while the harder work goes undone.

Questions to ask before you agree

1. Which exact strain and what dose, and which study used that strain for this problem?
2. Are we treating a gut symptom I can describe, or are we treating development?
3. Does my child have any reason to avoid live organisms — prematurity, a central line, immune suppression, short bowel?
4. What would tell us it is working, and when do we stop if it is not?

More in this section