What They Will Try to Sell You

Within weeks of a diagnosis, sometimes within days, somebody will offer you a cure. A relative with a WhatsApp forward. A clinic with a chamber. A practitioner with a urine test that proves your child is full of heavy metals. A diet, a drip, an injection flown in from abroad.

They will not always be dishonest. Many of them believe it. But you are about to be asked for money — sometimes a very great deal of it — at the exact moment you are least able to think clearly, and that is the situation this page is for.

We are not going to tell you everything on this list is nonsense, because that would not be true. One or two are genuinely unresolved. The point is to tell you which is which, and what to ask.

How this happens: the secretin story

In 1998 a gastrointestinal hormone called secretin was proposed as an effective treatment for autism, on the strength of anecdote. Within a couple of years it was one of the most sought-after treatments in the world; families spent enormous sums and travelled long distances for infusions.

A Cochrane review eventually assessed it. Sixteen studies met the inclusion criteria. There was no benefit.

Nobody in that story was necessarily lying. A hormone was suggested, desperate families tried it, some children improved for the reasons children improve — development, attention, expectation, chance — and the story spread faster than anyone could test it. That is the mechanism, and it is running right now with a different set of treatments.

The one rule that does most of the work

Ask what the comparison group got.

Autistic children develop. Any child given a new treatment is also, at that moment, being given extra attention, closer observation, and a parent full of hope. That is why "he started talking after we began X" is not evidence, however true it is that he started talking.

The only design that separates the treatment from everything else surrounding it is a controlled trial, where a similar group of children got a convincing placebo and nobody knew who had which. When someone offers you a treatment, one question does most of the work: has this been compared against placebo, and what happened?

Chelation — do not

This is the clearest case on the page, so we will be blunt.

Chelation uses drugs that bind metals so they are passed in urine. The claim is that autism is caused by accumulated heavy metals and that removing them improves symptoms.

A Cochrane review searched for randomised trials comparing chelating agents against placebo in autism. It found one eligible study. Nine others were excluded because they were not randomised, or were withdrawn before enrolment. An overview of seventeen Cochrane reviews of autism interventions listed chelating agents among those for which no benefits were found.

Now the part that matters most, from a commentary in the Journal of Medical Toxicology. It estimates that around half a million people with autism are chelated annually in the United States alone, the overwhelming majority for indications that are not medically accepted — and it describes exactly how families are convinced:

A urine sample is taken after a dose of a chelating agent, and the result is compared against reference ranges for non-chelated urine. Chelation is designed to pull metals into urine. So the test reports high metals, the parent is shown a number that looks alarming, and treatment follows.

The paper's own summary of the risk is that unnecessary chelation is expensive, can cause significant acute adverse effects, and may have long-term consequences.

If someone shows you a post-provocation urine metals result, you now know what you are looking at. Ask what the reference range was measured on.

Hyperbaric oxygen — genuinely unresolved, and expensive

Here we have to be more careful, because a flat dismissal would misrepresent the literature.

Against it: the overview of seventeen Cochrane reviews lists hyperbaric oxygen therapy among the interventions for which no benefits were found. A review in Canadian Family Physician concluded that case series and randomised trials show no evidence of benefit, and that the single randomised trial reporting effectiveness had not been replicated.

For it: a 2025 systematic review and meta-analysis pooled 17 studies with 890 patients and did find significant effects — core autism symptoms SMD −0.66 (95% CI −1.04 to −0.28), communication −0.88, cognitive awareness −0.93, behaviour −0.80.

Before that changes your mind, read the authors' own caveat, which we quote rather than summarise: the review has "limitations such as poor quality and high heterogeneity of the included study", and they call for rigorously designed, high-quality studies to confirm the efficacy and establish protocols. It included quasi-experimental studies alongside randomised ones — meaning some of that pooled effect comes from designs that cannot separate treatment from time and attention.

Our honest position: unresolved. Not proven useless, not remotely proven useful. Which means it is not a thing to sell a piece of family land for — and in practice, that is what the decision usually is.

Stem cells and cord blood — far earlier than the marketing suggests

There is now a real randomised trial, and it is much smaller than what is being advertised to you.

A 2026 single-centre, double-blind, placebo-controlled trial gave 34 children aged 3 to 8 — 17 treated, 17 placebo — four infusions of umbilical cord blood mononuclear cells alongside their existing therapy, and followed them for 13 weeks. The treated group showed a greater reduction in total social responsiveness score (−6.03, 95% CI −11.93 to −0.14) and in social cognition (−9.95). No major transplantation-related adverse events occurred.

Look at that first confidence interval. Its upper bound is −0.14 — it only just clears zero. That is a genuine but fragile result, in 34 children, at one centre, over three months, with the children also receiving their normal rehabilitation.

It is a reason for more research. It is not a reason to fly abroad and pay for infusions, and it is certainly not what the commercial clinics are basing their claims on.

Gluten-free and casein-free diets — the reviews disagree

You will be told this works by people who are entirely sincere. Here is the whole picture, including the part that supports them.

• A systematic review in the European Journal of Nutrition found six randomised trials with 214 participants. With few exceptions there were no statistically significant differences in core symptoms. Its conclusion: little evidence that a GFCF diet is beneficial.
• A systematic review in Pediatrics covering 19 randomised trials found the evidence on gluten/casein-free diets insufficient to draw conclusions, and found omega-3 supplementation did not affect challenging behaviours.
• A review of 18 studies found 10 reporting a positive association and 8 finding none, and concluded there is little scientific evidence to support dietary therapies.
• A narrative review published in the Journal of the Pakistan Medical Association — the one most likely to be quoted to you locally — concluded the diet was safe with therapeutic benefits. It included 7 of 80 studies screened, and states in its own conclusion that the trials are sparse, small, and show indication of bias.

What we would actually say. No serious harm has been reported from the diet itself. But it is expensive here, it is a great deal of work, and in a child who is already a narrow eater it can shrink an already inadequate diet — see helping a picky eater and feeding difficulties. If you want to try it, do it the way a trial would: set an end date, decide in advance what improvement you are looking for, have someone who is not you rate the child, and stop if it is not there.

Supplements — mostly no, a few "with caution"

A systematic review of experimental drug and supplement trials in autism identified 115 published randomised controlled trials across 133 compounds. The headline: for the vast majority — including cannabidiol, vasopressin and probiotics — there is insufficient evidence of efficacy and safety. The compounds with the strongest theoretical backing (arbaclofen, balovaptan, bumetanide) all failed to reach their primary endpoints.

The same review named a small subset — N-acetylcysteine, folinic acid, l-carnitine, coenzyme Q10, sulforaphane and metformin — as possibly considerable for clinical use with due caution, because there is promising evidence. That is a doctor's decision with monitoring, not a shelf purchase.

Separately, the Cochrane overview found no benefit for omega-3, vitamin B6 with magnesium, sound therapies or secretin.

So what does have evidence?

This is the part that should be reassuring, and it comes from the same overview of seventeen Cochrane reviews that dismissed the treatments above. The interventions where benefits were found:

• Early intensive behavioural intervention
• Parent-mediated early intervention
• Social skills groups
• Music therapy
• Theory of Mind approaches, and acupuncture

And now the sentence that keeps us honest: that same review rated the quality of evidence for all of them as very low to low. We are describing our own field, and we are not going to apply a stricter standard to a hyperbaric chamber than to ourselves.

The difference is not that our evidence is overwhelming. It is that these approaches are cheap or free, carry no physical risk, teach skills your child keeps, and do not require you to sell anything. When the evidence for everything is weak, that asymmetry is the whole argument.

Five questions before you pay

1. "Has this been compared against a placebo, and what happened?" Not "is there research" — there is always research. What did the controlled trials find?
2. "What does it cost in total, including follow-up?" Get the whole number in writing, not the per-session one.
3. "What are the risks and what would make you stop?" Anyone who says there are no risks at all has told you something important about themselves.
4. "How will we know in eight weeks whether it worked?" Agree the measure in advance, in writing, and have someone who is not you assess it.
5. "Would you put that claim in writing?" Very little survives this question.

Two red flags that are close to definitive

• The word "cure". Autism is a developmental difference, not an infection. Anyone promising to remove it is either mistaken or selling.
• A test you had never heard of that proves the diagnosis their treatment addresses. The post-chelation urine metals result is the classic, but the pattern is general: an unusual test, an alarming number, and a treatment conveniently to hand.

Where we fit

We have no financial interest in talking you out of a hyperbaric chamber, and we have said above, in writing, that the evidence for our own field is rated low quality. We would rather you knew that than trusted us more than the evidence deserves.

What we will do is look at what your child can actually do now and what comes next, and be straight about what is likely to help. If you are being pushed towards something expensive and you want a second opinion from someone who is not selling it, bring the leaflet and we will read it with you. Our page on what causes autism deals with the beliefs underneath many of these offers, just diagnosed is written for the first fortnight, and our autism page covers what support actually looks like.

A consultation is Rs 1,500 and lasts up to 50 minutes, we work in Urdu or English, and you do not need a referral.

For a treatment that does have strong short-term trial evidence, and what its limits are, see ADHD medication.

Sources

• Lyra L, et al. What do Cochrane systematic reviews say about interventions for autism spectrum disorders? Sao Paulo Medical Journal. PubMed 28538871
• James S, Stevenson SW, Silove N, Williams K. Chelation for autism spectrum disorder (ASD). Cochrane Database of Systematic Reviews. PubMed 26106752
• Brent J. Commentary on the abuse of metal chelation therapy in patients with autism spectrum disorders. Journal of Medical Toxicology. PubMed 24113859
• Tu Y, et al. The effectiveness of hyperbaric oxygen therapy in children and adolescents with autism spectrum disorders: a systematic review and meta-analysis. Progress in Neuro-Psychopharmacology & Biological Psychiatry. PubMed 39826608
• Sakulchit T, Ladish C, Goldman RD. Hyperbaric oxygen therapy for children with autism spectrum disorder. Canadian Family Physician. PubMed 28615394
• Zhang X, et al. Efficacy and safety of allogeneic umbilical cord blood mononuclear cells therapy in children with autism spectrum disorder and immune dysregulation: a single-center, double-blinded, randomized, placebo-controlled trial. Stem Cell Research & Therapy. PubMed 42321872
• Piwowarczyk A, et al. Gluten- and casein-free diet and autism spectrum disorders in children: a systematic review. European Journal of Nutrition. PubMed 28612113
• Sathe N, Andrews JC, McPheeters ML, Warren ZE. Nutritional and dietary interventions for autism spectrum disorder: a systematic review. Pediatrics. PubMed 28562286
• Monteiro MA, et al. Autism spectrum disorder: a systematic review about nutritional interventions. Revista Paulista de Pediatria. PubMed 32187297
• Akhter S, et al. A narrative review on manifestations of gluten free casein free diet in autism and autism spectrum disorders. JPMA. PubMed 36660995
• Persico AM, et al. The pediatric psychopharmacology of autism spectrum disorder: a systematic review — Part II: the future. Progress in Neuro-Psychopharmacology & Biological Psychiatry. PubMed 39490514
• Williams K, Wray JA, Wheeler DM. Intravenous secretin for autism spectrum disorders (ASD). Cochrane Database of Systematic Reviews. PubMed 22513913