Thinking About Autism, Asperger Disorder and PDD-NOS

 

By: Dr. Peter Szatmari

 

There have been several important changes to the diagnosis and classification of autism over the years.  Indeed, since 1980 there have been three different classification systems used in North America; the Rutter criteria, DSM-III, DSM-III-R.  It is only with the recent publication of ICD-10 and DSM-IV which contain basically identical criteria, that some consensus has been reached in the diagnosis and classification of the PDD’s.

 

DSM-III contained explicit criteria for autism and several other PDD subtypes including atypical autism, childhood-onset pervasive developmental disorder and residual autism.  The term PDD was used to refer to a triad of impairments in reciprocal social interaction, in verbal and nonverbal communication, and a pattern of repetitive stereotyped behaviours.   It derived from  Wing’s conceptualization of the autistic triad consisting of three types of social deficits; social interaction, social communication and social play.  Autism was now part of a more general group of disorders that shared some essential features but differed on symptom pattern or natural history.   Although the great advantage of the DSM-III criteria were that these could be more easily operationalized than previous criteria the criteria for autism, lacked sensitivity.  In other words, there were a substantial number of children whom clinicians agreed had autism but did not meet the stringent DSM-III criteria.

 

The revisions introduced in DSM-III-R in 1987  were motivated by this low sensitivity and the lack of a developmental framework that separated infantile autism from residual autism.  The other subtypes of PDD were collapsed into a single category known as PDDNOS (not otherwise specified).  The sensitivity and specificity of these new criteria were assessed in a number of studies and it soon became apparent that although sensitivity was now very high, specificity was quite low.  In other words, a large number of children whom clinicians felt did not meet criteria for autism did, in fact, meet the DSM-III-R criteria and so were misclassified.  It also became apparent that a very large number of children now received a diagnosis of  PDDNOS, which was not seen as a useful label by clinicians and parents alike.

 

The DSM-IV criteria for autism attempted to keep the developmental framework and the broader conception of autism contained in DSM-III-R while at the same time improving its specificity and reintroducing several non-autistic PDD subtypes.  The results from the large field trial indicated that the DSM-IV criteria (and the very similar ICD-10 criteria) have quite acceptable levels of sensitivity and specificity and can be reliably applied by clinicians.

 

The non-autistic forms of Pervasive Developmental Disorder differ from autism either on symptom pattern or on developmental course.  For example, for both Rett’s Disorder and disintegrative disorder of childhood, there is a period of normal development followed by regression in social and communication skills.  This regression is followed by the appearance of autistic symptomatology and behaviours and, in the case of Rett’s Disorder, characteristic behavioural and physical abnormalities.  Atypical autism or PDDNOS, differs from autism either in having a later age of onset (i.e. after 36 months), by having subthreshold number of symptoms (that is less than 6 out of 12) or falling below the threshold for one of the major criteria (i.e. social, communication or behaviours).   Asperger Disorder is characterized by the same types of social impairments seen in autism plus the development of very bizarre intense interests such as bus timetables, insects, meteorology, cartography, etc.  It is distinguished from autism by the presence of normal cognitive development and the absence of clinically significant language delay.  In other words, children with Asperger Disorder have normal IQ and usually are speaking in phrases by three years of age.  This does not necessarily mean that they have “normal” language, only that there is no severe delay; mild delays do not constitute exclusion criteria for Asperger Disorder .  Asperger Disorder children tend to have better social skills and fewer examples of repetitive stereotyped behaviours than children with autism.

 

There is considerable controversy concerning the measurement of Asperger Disorder and atypical autism.   One major problem is that it is difficult to apply the DSM-IV criteria for these two disorders.  For example, the DSM-IV criteria state that if a child meets criteria for both autism and Asperger Disorder he/she is given a diagnosis of autism.  Applying these criteria to a sample of PDD children results in few, if any, children meeting criteria for Asperger Disorder.  In our research and clinical experience, it is almost impossible for a child to meet the social and repetitive activity criteria for Asperger Disorder and not also meet the criteria for autism.  Instead, it may make more sense to give precedence to a diagnosis of Asperger Disorder if a child with PDD symptoms has normal cognitive and language development.  While this may reflect actual clinical practice and does provide more coverage it is not consistent with the use of DSM-IV as a diagnostic guide.

 

There are more serious problems with the reliability and diagnosis of atypical autism.  In a recent study, we had three expert clinicians review an archive of clinical data on a consecutive series of PDD children.  While the clinicians were able to agree on a diagnosis of autism and Asperger Disorder, they were unable to agree beyond chance on a diagnosis of atypical autism or PDDNOS. This diagnosis was often considered in very low functioning children who often did not have enough language to show some of the behaviours in that domain or for children who were not aware enough of their surroundings to show rituals or resistance to change. The other type of child that this diagnosis often refers to are those with severe speech delay and PDD symptoms who then make a remarkable recovery and show fewer symptoms later on. The problem is that both these types of children usually demonstrate between 4-6 DSM-IV PDD behaviours from all three domains and the clinicians had a very difficult time deciding whether they were above or below the cut-off of 6 for a diagnosis of autism. We concluded that the current DSM-IV criteria for atypical autism (PDDNOS) were too vague to be applied reliably.  In clinical practice, this diagnosis should be considered tentative and re-evaluated to see if a more precise diagnosis becomes apparent later on. 

 

There are also few data on the validity of these non-autistic forms of PDD and the extent to which they differ from autism either in terms of cause, course, or response to treatment.  We do know that children with autism can have siblings with Asperger Disorder and atypical autism indicating that common genetic mechanisms may be responsible for all PDD subtypes.  In fact we found that it was level of functioning that showed familial aggregation among sibships with PDD not PDD subtype. For example, children with Asperger Disorder tended to have siblings with higher functioning autism rather than lower functioning autism. If there are etilogical subtypes of PDD they are most likely distinguished on the basis of level of functioning not on the basis of PDD subtypes.

 

  There is however evidence that children with Asperger Disorder do have a different outcome than children with autism.  We recently reported that at 6-8 years of age children with Asperger Disorder had better adaptive behaviour skills in socialization and communication and fewer autistic symptoms than high functioning children with autism. These differences were explained by the differences between the groups already present two years earlier. Thus, we concluded that the groups were following parallel trajectories. The differences in outcome were quite large though and could not be simply explained by initial differences in IQ or in language skills. However, once some children with autism develop fluent language, they tended to resemble more, and more the children with Asperger Disorder, only at an earlier stage in their development. In other words, when children with autism develop a certain level of language ability, they appear to join the developmental trajectory of the children with Asperger Disorder.  Whether they ever actually catch up will need to be determined by further research. We concluded that the clinical distinction between autism and Asperger Disorder is really based on the timing of the development of fluent language. The earlier this skill appears the better the outcome and the more likely the child is given a diagnosis of Asperger Disorder. The distinction is a quantitative one but does have a large influence of outcome and should not be minimized as clinically trivial.

 

It is also probably true that treatment is more closely related to intellectual level and verbal abilities than it is to PDD subtype. For example, it is likely true that children with Asperger Disorder have little need for speech and language therapy that focuses solely of the acquisition of vocabulary and age appropriate grammar. It is more important to focus on the development of  appropriate social and play skills that foster social interaction with other children. For children with autism on the other hand, speech and language therapy is an important part of their treatment program since the acquisition of language is such an essential part of functioning at home, at school and in the community. Neither is it clear that children with Asperger Disorder would benefit from the types of early intervention that focus heavily on discrete trial training as opposed to more naturalistic or incidental learning focusing on social skills.

 

It has been just over 50 years since Kanner originally described the syndrome of infantile autism.  Since that time, considerable evidence has accumulated suggesting that autism is a valid disorder and that the current DSM-IV criteria are quite useful.  However it is important not to think of them as discrete categories. It seems to make much more sense to think of them as different developmental pathways that potentially overlap at key points in development. Much more work still needs to be done however, on the differentiation of autism from the other types of Pervasive Developmental Disorders, refining the diagnostic criteria for atypical autism, and investigating the extent to which the different types of PDD have a different outcome and respond to different treatment modalities.  Currently, from a clinical point of view, it is most important to decide whether or not a child has PDD since that has major implications for understanding etiology, making predictions about outcome and planning treatment.  Given the current state of the evidence, trying to determine the type of PDD a child has is probably not an efficient use of scarce diagnostic and assessment resources.