New recommendations on autism spectrum disorder

• Download figure
• Open in new tab
• Download powerpoint

Annabella Bluesky/SPL

The Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-V) interim report on autism spectrum disorder was published recently (www.dsmv.com; 26 January 2011). On the basis of a review of the literature, the views of experts, and the requirements of developing a diagnostic manual that satisfies many stakeholders, the Neurodevelopmental Work Group has made several bold and perhaps controversial recommendations. One is to eliminate the term “pervasive developmental disorder” and replace it with the more commonly used “autism spectrum disorder.” A second is to eliminate the various subtypes that were previously described, including Asperger’s disorder. This has been perhaps the most contentious decision because many people with a diagnosis of Asperger’s disorder do not see themselves as having autism and are concerned that they will be denied recognition and services if Asperger’s disorder is eliminated⇑.

The reasons for the work group’s decisions are threefold. Firstly, it is difficult to apply the criteria for the subtypes of pervasive developmental disorder in a systematic way; secondly, most children who are diagnosed with Asperger’s disorder would meet criteria for autistic disorder; and thirdly, the risk factors for autistic disorder and Asperger’s disorder are often similar. On the basis of existing evidence, the committee concluded that autistic and Asperger’s disorders (along with the residual term “pervasive developmental disorder-not otherwise specified) should be combined into one category, called autistic spectrum disorder.

These changes will probably have some positive consequences. Clinical resources will no longer be needed to determine what “type” of autistic spectrum disorder a child has. Disagreements over these distinctions will be eliminated, so that clinicians can focus on getting children into treatment as quickly as possible. People with Asperger’s disorder will no longer be denied services reserved for people with autistic disorder. But there are also some potential dangers. Autistic spectrum disorder advocacy groups must make sure that people with Asperger’s disorder are allowed a voice and do not feel excluded. In addition, some people with “true” Asperger’s disorder might be missed and given an inappropriate diagnosis, such as obsessive compulsive disorder or schizoid disorder. This is because some clinicians will not appreciate the subtleties of differential diagnosis and the wide clinical expression of autistic spectrum disorder, especially in adults where a developmental history might not be available.

Some people will argue with the work group’s decision.1 The more important challenge is to appreciate the heterogeneous nature of the spectrum of autistic disorder phenotypes. This growing appreciation comes from a careful reading of studies on the longitudinal course of the disorder and on its genetic architecture. Recent prospective studies have shown remarkably diverse results,2 3 4 5 6 with, at least in one study, a substantial proportion of subjects (24%) having a “very good” outcome.2 Recent genetic studies have also emphasised heterogeneity in risk factors. Up to 7% of cases of autistic spectrum disorder seem to be the result of genetic variants, many of which arise de novo through the generation of copy number variants.7 Although each copy number variant is extremely rare (less than 1%), this class of risk factor may converge on similar pathways related to synaptic development.8 A better understanding of both de novo and inherited variants will probably show that the genetic heterogeneity of autistic spectrum disorder is much greater than previously thought.

Can we capture this diversity and heterogeneity if subtypes no longer exist? The work group has decided on a bidimensional framework to capture diversity in phenotype. The two dimensions of social communication and repetitive behaviours may not be the most useful for categorising children with autistic spectrum disorder in terms of the causes, outcomes, and response to treatment because variation in these dimensions seems to be only weakly associated with variation in outcome and response to treatment, which are more closely related to cognitive and language abilities.9

The key challenge for research in the next decade is to find clinical markers that provide useful indicators of variation in multiple domains, not just symptoms. In one possible conceptualisation, autistic spectrum disorder represents a “family of dimensional phentoypes” that includes symptoms, level of functioning, and psychiatric and medical comorbidities. Many of these dimensional phenotypes may be correlated and some may be independent from a structural measurement point of view. Some may be useful in predicting outcome and response to treatment, and others may point to genetically homogeneous subgroups. Each person with autistic spectrum disorder represents an overlap of phenotypes, and the degree of overlap represents the clinical profile of any individual child. The idea that autistic spectrum disorder represents a family of phenotypes mediated by a network of genes is one that is becoming more familiar to researchers of complex disorders, such as cardiovascular disease, diabetes, and obesity.10 To think of autistic spectrum disorder in this way is to glimpse a degree of complexity much greater than simple subtypes such as autistic disorder, Asperger’s disorder, and pervasive developmental disorder—not otherwise specified.

The working group has done a good job in reviewing and interpreting the literature, but the quality of the evidence on which their conclusion is based is relatively weak.11 12 The goal of trying to establish the diagnostic validity of Asperger’s disorder is now outdated. Hopefully if we start afresh, the search for useful markers of heterogeneity can begin again with renewed vigor, and be unfettered with stale debates about labels. A greater focus on useful markers that capture diversity in developmental trajectories, in response to treatment and genetic heterogeneity, will be more fruitful in the long run.