Autism Spectrum Disorders Far More Prevalent Than Thought

ATLANTA, Feb. 8 -- Autism spectrum disorders are much more common than previously thought, and could affect as many as one in 150 school-age children, CDC researchers reported today.

"Autism is more common than we believed and is an urgent public health problem," said Catherine Rice, Ph.D., a behavioral scientist at the CDC's National Center on Birth Defects and Developmental Disabilities, in a briefing.

Data from a nationwide surveillance program conducted at sites in 14 states in 2002 suggest that the overall prevalence of ASD is about 6.6 per 1,000 eight-year-old children, ranging from a low of 3.3 per 1,000 in Alabama to a high of 10.6 per 1,000 in New Jersey, CDC investigators reported in the Feb. 9 issue of Morbidity and Mortality Weekly Report.

Even if the lowest prevalence estimates cited are closer to the actual prevalence rates, they are considerably higher than previously thought. Until this report, investigators have typically cited prevalence rates of between four to five per 10,000 and two to three per 1,000, based on the best available data.

Applying the estimates to the United States population, approximately 560,000 children from the ages of birth to 21 years may be living with an autism spectrum disorder, the investigators said.

The CDC investigators cautioned that the new estimates are accurate for the areas studied, but do not represent a national sample. Until wider studies are completed, it will be difficult to know whether there is an upward trend in autism spectrum disorders prevalence, CDC director Julie Gerberding M.D., M.P.H., said in a statement.

"Our estimates are becoming better and more consistent, though we can't yet tell if there is a true increase in autism spectrum disorders or if the changes are the result of our better studies," Dr. Gerberding said. "We do know, however, that these disorders are affecting too many children."

The current MMWR published two reports estimating the prevalence of autism spectrum disorders, a category that includes autistic disorder, pervasive developmental disorder, not otherwise specified (PDD-NOS), and Asperger's syndrome. The reports were conducted as part of the CDC's Autism and Developmental Disabilities Monitoring Network.

The first report estimated prevalence at six surveillance sites in Arizona, Georgia, Maryland, New Jersey, South Carolina, and West Virginia. The investigators looked at data on eight-year olds who were identified as having an autism spectrum disorders through screening or other records as displaying behaviors consistent with one of the three diagnoses under the autism spectrum disorders umbrella.

They chose age eight because studies have consistently shown that nearly all cases of autism spectrum disorders will have been detected by this age.

Clinicians reviewed the abstracted records to ensure that they met diagnostic criteria according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR).

The CDC investigators found that in the year 2000, a total of 1,252 children ages eight were identified as having an autism spectrum disorders, with the overall prevalence ranging from 4.5 per 1,000 in West Virginia to 9.9 per 1,000 in New Jersey. In only one of the six sites (Georgia) was there as statistically significant difference in the rate of autism spectrum disorders between non-Hispanic black and non-Hispanic white children, with white children having a 1.5-fold greater prevalence of autism spectrum disorders than black children.

In the 2000 study, the ratio of male-to-female prevalence ranged from about 2.8-5.5 to 1, consistent with earlier studies showing that autism spectrum disorders affect more boys than girls.

The majority of children with autism spectrum disorders in this survey were receiving special education services, and had a documented history of concerns regarding their development before the age of three.

The median age of earliest autism spectrum disorders diagnosis ranged from four years, four months in New Jersey and West Virginia to four years, eight months (Georgia). But, for 69% to 88% of children with an autism spectrum disorders, parents or caregivers had recorded concerns about the child's development age three.

The 2002 report used data from the same six sites as the 2000 report, plus sites in Alabama, Arkansas, Colorado, Missouri, North Carolina, Pennsylvania, Utah, and Wisconsin.

The investigators of this second report identified a total of 2,685 children with autism spectrum disorders of a total of 407,578 children age eight.

They found that autism spectrum disorders prevalence ranged from 3.3 per 1,000 in Alabama, to 10.6 in New Jersey. The mean overall prevalence was 6.6 per 1,000 or one in 152 children across all sites.

The prevalence in Alabama was significantly lower than in all other sites (P<0.001), and the prevalence in New Jersey was significantly higher than in the rest of the sites (P<0.0001).

As in 2000 and in earlier studies, autism spectrum disorders prevalence was higher among boys than among girls, ranging from a male-to-female ratio of 5-1.4 in Alabama, to 16.8-4 in New Jersey.

Also as in 2000, autism spectrum disorders prevalence was higher among white children than either Hispanic or black children. For whites the prevalence ranged from 3.3 to 12.5 per 1,000, compared with 3.4 to 7.7 per 1,000 for black non-Hispanics, and 0.3 to 9.7 per 1,000 for Hispanics.

The median age of earliest autism spectrum disorders diagnosis among the children who were eight years old in 2002 ranged from four years, one month in Utah, to five years, six months in Alabama. From 50% to 91% of parents/caregivers had recorded developmental concerns about the children before three years of age.

In the four states where the diagnosis was made based only on health care records, the mean prevalence was lower (5.1 per 1,000) than in sites where both health and education records were used to make the diagnosis (7.2 per 1,000).

"It is extremely difficult to accurately estimate the number of children who have an autism spectrum disorders," said Marshalyn Yeargin-Allsopp, M.D., chief of CDC's autism program. "Medical records often do not provide such information, and identification is often made by schools or education specialists"

"It's important to note that these studies don't provide a national estimate, but that they do confirm that autism spectrum disorders in the areas surveyed are more common in these communities studied than previously thought," Dr. Yeargin-Allsopp added. "We need to continue efforts to monitor the prevalence of autism spectrum disorders and to improve our understanding of these disorders. Good estimates of how many children in a community may have an autism spectrum disorders will also help school and health officials in their planning and intervention efforts."

The investigators said that the reasons for the difference in autism spectrum disorders prevalence between current and earlier estimates are unknown, but could be related to better surveillance techniques producing more accurate estimates, or to changes in diagnostic practice, so-called diagnostic shift.

"We cannot make conclusions about trends in autism spectrum disorders prevalence at this tine," Dr. Rice said. "However, continued monitoring of autism spectrum disorders prevalence in these sites will help us answer that question starting with children born in the 1990s."