Boston Children's Hospital researchers have developed a prototype blood test for autism, and preliminary results published Wednesday suggest it could one day be used to help diagnose the disorder when children are very young and respond best to treatment.
The test, which measures the activity of dozens of genes, was able in one group to predict with about 70 percent accuracy whether a boy was at risk for autism or not. Outside researchers cautioned that the work has limitations and that the blood test needs much more study before it is clear whether it could be a useful tool for doctors and parents.
The new study represents one piece of a much larger effort to identify biological hallmarks of autism — whether genes or brain abnormalities detected on scans — that could speed diagnosis of a disorder that causes social problems and developmental delays and is now diagnosed in an estimated 1 in 88 children.
"It's a very worthwhile area of investigation: The hypothesis that one might be able to classify patients based on blood is very worth testing," said Dr. Daniel Geschwind, a professor of neurology and psychiatry at the University of California, Los Angeles School of Medicine who was not involved in the study. But he added, "It's fairly clear that we have a long way to go."
The urgency and need for better tests for autism are increasing as its prevalence grows, in large part because children's brains are most easily molded early in life. Intervening sooner has a greater chance of leading to real and lasting improvements.
But autism spectrum disorders are made up of a suite of behavioral and cognitive problems that have no easy biological litmus test, and the centers that specialize in diagnosis often have lengthy wait lists. The average age of diagnosis is after 4, according to government data.
Dr. Isaac Kohane, a professor of health sciences and technology at Children's Hospital, stumbled on this line of research by chance. More than a decade ago, a neighbor in Brookline asked Kohane, who is also a pediatrician, about one of her triplets. The child tended to play alone and just seemed a little different than the other two.
"I said, 'I know what you're thinking — you think this kid has autism,' " Kohane recalled. He reassured her that he didn't think she had anything to worry about.
Later, when the child was diagnosed with autism, he was stunned, sorry, and motivated to start learning more.
"It's just shocking to be told afterward, 'That kid you told me was fine has autism,' " said Kohane, who led the new study, published in the journal PLOS ONE. "I started reading up a lot on that. . . . This is an amazing problem. I had been unaware of frankly how little we know about this disease and how haphazard our diagnostic process is."
That process is painfully familiar to Stephanie Sourwine, a single mother living in rural Alexandria Bay, N.Y.
Sourwine said she began noticing her son Austin regress when he was 2 years old, but he was almost 3 by the time she got a diagnosis. He
"Gabby was kind of heartbreaking," Sourwine said. "If I could have skipped all that wait, it would have been much better."
Now, she said, it is clear that her youngest, 4-year-old Avery, may have some of the same problems. She called the doctor's office, and was told that he was not taking any new patients; there was a two-year waiting list, despite her family history. Her best option would be to travel to medical centers many hours away.
A Southborough company, SynapDx, has licensed Kohane's technology and is preparing to launch a large study of a blood test that will be informed by his laboratory's research and its own studies, according to the company's president, Stanley Lapidus.
"Behavioral therapy (for autism) is pretty good. But it has to be started earlier — the earlier, the better. Alas, the average age of diagnosis is 4½, and the average age of parental suspicion is 19 months," Lapidus said. "That gap is the tragedy here, and that gap doesn't need to exist."
Kohane said the test was not intended as a general screening tool, but as an aid that could help provide guidance when there was a suspicion of autism. He said it needed further study, and would not supplant behavioral testing.
Researchers not involved in the project had mixed opinions about the findings. The signal was identified by looking for differences in gene activity in the blood samples of 66 boys with autism and 33 boys without it. It was then tested on another sample, correctly predicting autism risk 73 percent of the time in boys, among samples taken from 104 children with autism and 82 without. It was 64 percent accurate for girls.
But differences in how and when the various samples were collected could affect the results, Geschwind said. He also added that one possibility is that the study is not detecting the disorder, but a general signal that exists in the blood of families with autism — meaning it might flag people with autism but also siblings or parents without the disorder.
Andy Shih, senior vice president for scientific affairs at Autism Speaks, a science and advocacy organization, said that so far genetic studies have been limited, identifying mutations that contribute to about a fifth of the cases.
"Anything that will help us push over that threshold . . . would be welcome progress," Shih said. "The caveat here is that it is a complex disorder; when you identify someone at risk . . . it doesn't mean someone's coming down with the disorder."