Children with autism do not appear to leak damaging proteins from their intestines, a study into the so-called "leaky gut" theory has suggested.
The "leaky gut" theory has led to advice to restrict dairy consumption
It has been claimed autistic children cannot fully digest proteins found in many foods - and that the resulting peptides escape and affect the brain.
But UK researchers found children with autism did not have more peptides in their urine than a control group.
They have published their findings in the Archives of Diseases in Childhood.
The "leaky gut" theory is based in part on the idea that vaccines such as MMR - given to immunise against measles, mumps and rubella - damage the wall of the intestines.
This causes the digestive problems which lead to the production of peptides, the theory goes.
To try to counter the effects of this, some parents of autistic children then reduce the amount of proteins such as gluten - found in wheat, oats, rye and barley - and casein - found in dairy products, such as milk, cheese and yogurt - in their child's diet.
Looking for a cure
But a team from Great Ormond Street Hospital, Guy's and St Thomas' Hospital and the University of Edinburgh have found no evidence of a higher level of peptides in the urine of autistic children.
They looked at 65 boys with autism and 158 without.
"It is very distressing to have a diagnosis of autism, a lifelong condition. Many families are driven to try out interventions which currently have no scientific basis," said Dr Hilary Cass of Great Ormond Street.
"Advocates of the leaky gut theory offer children a casein and gluten-free diet which as yet lacks an evidence base. Our research throws serious scientific doubt on the putative scientific basis of that diet."
But Paul Whiteley of the Autism Research Unit at Sunderland University said while the study appeared to have ruled out one reason why a gluten and casein-free diet may work, that did not mean it was not effective for some sufferers.
"It is very good news that more research is being carried out in this area. But evidence suggest that the diet does have beneficial effects for a proportion of those with autism, many of whom do suffer from bowel problems," he said.
"We need further investigation to find out if there are other reasons why it may work."
Benet Middleton of the National Autistic Society said there was an "urgent need" for more research into the efficacy of special diets for thos with autism.
"We are aware of anecdotal support for some dietary interventions, particularly those involving the exclusion of wheat and dairy products," he said.
"There is limited evidence about whether or not these diets are effective and concerns have been raised about their unregulated use."