Autism Diagnosis in Younger Kids Holds Up Over Time

ANN ARBOR, Mich., June 5 ? A diagnosis of autism in a two-year old child stands a pretty good chance of holding up when that child is nine, according to researchers here.

On the other hand, a two-year-old diagnosed with pervasive developmental disorder not otherwise specified (PDD-NOS) is more likely to be re-classified as having autism or another disorder seven years down the road, according to Catherine Lord, Ph.D., and colleagues in North Carolina, England, and Israel.

Getting the early diagnosis right may be important for researchers developing therapies for autism, and for educators developing early intervention programs targeted at the social, behavioral and adaptive deficits of children with autism, the authors suggested in the June issue of the Archives of General Psychiatry.

"Several intervention projects reported diagnostic changes and extraordinary levels of improvement in a substantial minority of young children with autism," the investigators wrote. "Other reports found little diagnostic change and fewer marked improvements. Possible explanations for these conflicting results are diagnostic instability or the lack of age-appropriate diagnostic criteria for very young children."

The authors undertook a study looking at the diagnostic utility of standard instruments for detecting autism in children under three years old. They also attempted to identify the features seen in two-year-olds that best predict a later diagnosis, whether that is autism or another disorder.

They conducted a prospective study of 192 children referred before their third birthdays for possible evaluation of autism, plus an additional 22 children with developmental delays but no diagnosis of autism from the same referral sources, who served as controls.

The children enrolled in the study were referred from both a state-supported community autism program in North Carolina, and from a privately funded university hospital in Chicago. Controls came from the same referral sources as the children in the North Carolina program.

The children were evaluated at two, five, and nine years with standard instruments, including the Autism Diagnostic Interview-Revised (ADI-R, a parental interview), the Pre-Linguistic Autism Diagnostic Observation Schedule/Autism Diagnostic Observation Schedule (an observational scale), and with independent clinical diagnoses made at ages two and nine.

The results were compared with those of a clinical research team's standard, best-estimate diagnoses.

A total of 172 children were available for examination at age nine, and 58% of these kids had a best-estimate diagnosis of autism, up from 49% at age two.

Overall, two-thirds (67%) of best-estimate diagnoses remained the same from age two to age nine. Among those with an original (age two) diagnosis of autism, 76% were diagnosed autism again at age nine.

Diagnostic stability was even greater among children with an original diagnosis of an autism spectrum disorder, with 90% of such children having a diagnosis of either autism or an autism spectrum disorder such as Asperger syndrome, atypical autism or PDD-NOS at age nine.

In a simple additive logistic regression mode for best-estimate autism diagnosis, the investigator found that the parental interview (ADI-R), observational scale, and clinician diagnosis at two years were all statistically significant, independent predictors for an autism diagnosis at nine years.

In cases where the diagnosis shifted from age two to age nine, the change was for the better (i.e. more accurate) in 8% of the cases, but worse in 18% of the cases, and an early diagnosis of autism was more likely to be right seven years later than an early diagnosis of PDD-NOS.

Whereas only one in 84 children diagnosed with autism at age two was determined at age nine not to have the disorder, more than half the children with the more vague diagnosis of PDD-NOS were diagnosed with autism when they were nine.

"There are real questions about the usefulness of pervasive developmental disorder not otherwise specified as a categorical diagnosis," the investigators wrote. "However, especially for very young children, having a way for experienced clinicians to acknowledge their uncertainly about some two-year-olds was ultimately helpful as a means of flagging children who by age 9 years had a range of difficulties from autism to very mild social deficits."

They additionally caution that "health care professionals should be wary of telling parents that their young children do not have autism, only pervasive developmental disorder not otherwise specified."

The authors also noted that judgment of clinicians who were trained on specific diagnostic instruments can help to improve upon the information obtained from parental interviews and observational studies.