Kids with autism who eat a diet that eliminates the proteins found in wheat and grains, as well as the protein from dairy don't see significant improvement in their behaviour, new research has found.

Gastrointestinal problems are common in children with autism spectrum disorder. That's sparked the creation of the so-called "autism diet" -- a gluten-free, casein-free eating plan.

Some who advocate for the diet claim that kids with autism have insufficient enzyme activity in their gastrointestinal tracts, which leaves their bodies unable to completely digest the milk protein casein and gluten, the protein found in wheat, rye and some other grains.

They claim that the incomplete breakdown of these proteins is what leads to autism symptoms, so that eliminating these elements can improve symptoms. The diet has become popular, particularly since autism activist and actress Jenny McCarthy said her child's autism symptoms improved when she put him on this diet.

Dr. Susan Hyman, an associate professor of pediatrics at Golisano Children's Hospital at the University of Rochester Medical Center in Rochester, N.Y., set out to test the effectiveness of the diet.

She and her colleagues enroled 22 children with autism spectrum disorder, between the ages of two and five. None of the children had wheat or milk allergies, celiac disease, or iron deficiencies. Fourteen of the kids finished the 18-week study in which they were put on the strict gluten-free, casein-free diet.

The children ate the diet for four weeks, and then were randomly given identical-looking snacks that contained either gluten, casein, both, or a placebo. They then returned to the diet for another four weeks and the routine continued until each child received each snack three times over 12 weeks.

Parents, teachers and research assistants were then questioned about the child's behavior before and after the snacks were eaten. They were asked about sleep problems, which are common in kids with autism, bowel problems, as well as socializing and language behaviour.

The researchers found no "demonstrable improvement" in autism symptoms.

The children showed no differences in attention, activity, sleep or frequency or quality of bowel habits. Some showed a small increase in social language and interest in interaction; however, the differences were not statistically significant, the researchers said.

"It would have been wonderful for children with autism and their families if we found that the diet could really help, but this small study didn't show significant benefits," Hyman said in a news release.

The findings are due to be presented Saturday at the International Meeting for Autism Research in Philadelphia.

Despite the conclusions, Hyman says more studies on how diet might affect autism are warranted, because the study group was small it's possible that there are subpopulations of children that could benefit. A longer study might find different results as well, she said.