At present, the only way to diagnose autism is by observing behavior, which is time-consuming and involves an element of subjectivity. (Contrary to myth, brain scans are not yet a diagnostic tool.) Shirley Wang writes in The Wall Street Journal about a potential blood test that might distinguish kids with autism from those with other developmental delays.
The 20-site, 660-patient project, expected to be launched Wednesday, is thought to be one of the biggest studies to date to examine a biological marker for autism-spectrum disorders, which affect one in every 50 children in the U.S.There is reason for caution about such tests, as previous posts have explained.
The blood test aims to speed the diagnosis of autism, a condition characterized by poor social interaction and repetitive behaviors that can be hard to recognize when a child is very young.
The average age of diagnosis in the U.S. is about 4 years, older than is optimal, according to experts, because therapies are more effective when begun early.
"Time is the enemy here in improving outcomes," said Stan Lapidus, chief executive of SynapDx, a lab-services company in Lexington, Mass., that is sponsoring the trial.
The new study on autism is based on early work from several academic groups suggesting that it may be possible to distinguish the condition by looking at the quantity of particular molecules in the blood.
Isaac Kohane, a professor of pediatrics, health sciences and technology at Boston Children's Hospital and Harvard Medical School, who led some of this work, said his team found a "substantial and persistent difference" in molecules produced by about 55 genes that could identify about 75% of the children with autism. The accuracy is "not the best you've ever seen, but it may be good enough to start with," said Dr. Kohane, who serves as a consultant to SynapDx but isn't involved with its blood test.