Autism: Earliest Diagnosis Ever

Autism spectrum disorders affect about one in 150 children. Often doctors don't diagnose the disability--which is characterized by impairments of social interaction and communication--until age 3. And yet, experts say earlier diagnosis is critical, since it can lead to earlier intervention and better outcomes. The good news: in a study appearing this week in the Archives of General Psychiatry, researchers say they have successfully identified autism in children as young as 14 months--the earliest the disorder has ever been diagnosed. The authors say the findings indicate that about half of autism cases can be diagnosed within months of the first birthday. Researchers at the Kennedy Krieger Institute in Baltimore, Md., evaluated social and communication development among 107 high-risk kids  (children whose siblings have autism) and a control group of 18 low-risk kids (no family history of autism). Speech pathologist Rebecca Landa, director of Kennedy Krieger's Center for Autism and Related Disorders,  and lead author of the study, spoke with NEWSWEEK 's Karen Springen about the implications for the 1.5 million Americans with autism. Excerpts:

NEWSWEEK: What does this mean for the theory that vaccines and other environmental factors may play a role in autism?
Rebecca Landa:
This study doesn't really touch on that. There may come a time when we can look at these children's records and see what kind of vaccines they were exposed to. There are different onset patterns. Basically, in the one pattern, there are clear, clear problems in social development and communication development at 14 months of age. The children are quite distinguishable. [But] for about half of the children we studied, their autism symptoms didn't really show up until later, sometime after 14 months, but certainly before 24 months. The point is that some parents are concerned that when their child gets a vaccine at around 15 months, like the MMR [measles, mumps, rubella], that their child is suddenly changing. What these data are indicating is that there is going to be a progressive phase to the disorder of autism in at least half of the children, where you're going to see that around 14 months, they look pretty healthy. It begins to gradually dwindle. They gradually begin to look pretty different. This is probably not an incident that comes upon a child instantaneously. This is an ongoing process.

Why were you able to diagnose just half the kids at 14 months?
That's a very good question, and we are going to be looking at that to see if there are certain very subtle predictors of children who are going to have this downward turn. Most of the children who ended up with this downward turn did show some mild, subtle developmental disruptions at 14 months, but nothing that would be alarming.

How did you pick the kids for your study?
This study was designed to try to understand the very earliest possible markers of autism in life and also how autism develops over time. We did pick babies who were at high risk for developing autism. We had a sample of 107 baby sibs of kids with autism. Somewhere around 30 of them developed autism.

Why is early diagnosis so important?
Because we believe that we can prevent certain aspects of the autism behavioral picture from either emerging or from becoming a major problem.

I'm not saying we can do this for every child, but you really contour how they learn to interact with people, how they play, how they learn to learn. In a typically developing infant, everything is a learnable moment. In an infant or toddler with autism, their attention gets hyperfocused on things that aren't important--like for example, the letters on a wooden block.

How do you stop such behavior?
You redirect the child's attention, you engage them in other toys. You also teach them how to pay attention to really important social signals, like people's eyes, people's faces.

So early intervention can affect how a child with autism functions later in life?
It very well can impact how someone functions later. We've just been funded to follow them [the 125 kids in this study and another 150 kids who entered the study later] until they're 8 years old. We have a separate treatment study. If you're not giving these kids real intervention, you see what happens to them. They pretty much stop smiling at people. If you can catch them early and really engage them, their growth curves might look really different.

Does autism begin to develop even sooner than 14 months?
Sometime between 6 and 12 months, there is this period of brain overgrowth. In a baby's brain, there are too many cells, nerve cells. Some of them have to be very systematically pruned away. In the kids with autism, that's not happening properly. The neurobiological underpinnings are probably already there. There's probably something underlying going on, but it hasn't really affected behavior until later. [Heather Cody] Hazlett talks about brain overgrowth not really being detectable until the time of the first birthday. [A 2005 University of North Carolina and Duke University study found that by age 2, kids with autism show a 5 percent enlargement of their brains.]

Could vaccines be a catalyst?
We're going to go back and look at the vaccine histories of these children.

You talk about kids with autism exhibiting irregularities in how they play. For example, they would not pretend to eat with a toy fork. How do we make sure parents whose kids do anything "abnormal" don't automatically suspect autism?
The thing to help parents not be paranoid is this: if there are warning signs for autism, there's more than one. It's not going to just be a fork. No. 1, there have to be multiple signs, and No. 2, the signs have to persist over weeks, months.

How did you diagnose these kids so early?
I had to learn to retrain my eyes when I started to see the 14-month-olds. I thought autism at 14 months was going to look like autism at 36 months, the age at which people normally diagnose it. It's the same flavor--the social system is disrupted, the communication system is disrupted. But it's different in that it's not as pervasively disrupted. What I mean is that at 14 months, you can get kids with autism to give you a beautiful response to peekaboo. But you can't get the child to engage with you around more novel, new activities. At 14 months, you see more flickers of interaction. They were doing some looking at people and smiling.

Do you think you can diagnose autism even earlier?
It is possible. The problem is that very early in development, there's a very rapid change in child behavior. If you did see these signs in 9 to 12 months of age, I would probably put out an alert. If there was enough of a delay, I might even put the child in intervention. I would start doing some preventive interaction, like making sure there was a variety of toys available to the child, and teaching this child to do a diversity of things with each toy.

So the kids don't fixate on one toy?
That's right.

How many false positives did you get?
Maybe three. They still had problems. But it wasn't autism.

There's no blood test for autism yet, but is it realistic that some day we could get a definitive test that could be used very early?
I have a colleague, Dr. Carlos Pardo-Villamizar at Johns Hopkins. We are now collecting blood from the infants at each visit. He has some specific theories about what might be fruitful as a blood test. He's looking at how these certain things change in the brain that you can measure through the blood as the child gets older.

So some day kids may get tested at birth, then? How long do you think it will take to get such a test?
It's possible. I'm going to guess a minimum of five years.

Editor's Pick