Does the ASD label have validity?

In 1943, a child psychiatrist named Leo Kanner described 11 children with symptoms so unusual that they seemed to constitute a new and previously unrecognized syndrome [1]. Kanner's assessment of these children involved clinical observation and compilation of detailed case histories. Systematic assessment protocols of the type commonly used today [2–4] were not available to Kanner, but his case study approach was sufficient to reveal a cluster of seemingly unique developmental and behavioral characteristics. This newly recognized syndrome seemed to have a certain type of validity in the sense of representing a separate and distinct neurodevelopmental disorder.

Among the more extreme of the characteristics observed and recorded by Kanner were: (a) the presence of stereotyped behaviors, (b) frequent and prolonged tantrums, (c) an obsessive insistence on the maintenance of sameness, and (d) significant communication and social skills deficits. Subsequent reports of infants, children, adolescents, and adults diagnosed with this new autistic syndrome confirmed that this collection of symptoms did seemed to have some coherence or validity. Indeed, from the emerging evidence, Rutter [5] was able to identify a triad of symptoms that appeared unique to autism and which have stood the test of time: (a) qualitative impairment in the ability to relate socially to others, (b) significant deficits in speech, language, and communication development, and (c) aberrant and ritualistic behaviors, such as extreme tantrums, repetitive/stereotyped movements, and an obsessive insistence on sameness. Ritvo and Freeman [6] identified a fourth characteristic (i.e., unusual responses to sensory stimuli) that is now part of the cluster of symptoms that characterize autism.

Today the autistic syndrome constitutes just one disorder within a wider spectrum. Collectively, this wider range of disorders has been referred to pervasive developmental disorder (PDD) or, increasingly, as autism spectrum disorder (ASD). The ASD tag is gaining in popularity as the label to encompass the classic form of autistic disorder described by Kanner, as well as the other PDDs. These other PDDs include: (a) Asperger syndrome, (b) childhood disintegrative disorder, (c) Rett syndrome, and (d) pervasive developmental disorders not otherwise specified [7]. The range of disorders that fall under the PDD or ASD umbrella seem so broad that one might wonder whether Kanner would even recognize the children so classified as having any relation to those he described in 1943.

No one knows what Kanner would make of today's current trends in the classification and diagnosis of ASD. Perhaps he would agree with the move to broaden the definition, perhaps he would object. More important to the readership of this journal is whether the ASD label has validity. That is, does the ASD label cover a cohesive class of disorders that share sufficiently common features? Does Asperger syndrome, for example, have enough in common with the classic features of autistic disorder to warrant including both of these disorders under the ASD umbrella? Perhaps Asperger syndrome is simply a milder version of autism, but then again perhaps its nature and etiology is sufficiently different that it should not be lumped in with autism. The issue raised by this question is an issue of the validity of the ASD category, but is has more general implications for how the wide variety of neurodevelopmental disorders are more generally grouped and classified.

In examining the issue of validity, a useful starting point might be to consider whether the separate disorders under the ASD umbrella are sufficiently similar in terms of their diagnostic features. Here there is some evidence to suggest that the ADS family is perhaps too broad. Specifically, Rett syndrome and childhood disintegrative disorder have features that on face value would seem quite different from the other ASDs.

In the case of Rett syndrome there are several obvious reasons to suggest that it does not fit in too well with the other ASDs. For one, Rett syndrome is diagnosed almost exclusively in females, whereas the other PDDs are more common among males [6]. Second, Rett syndrome is a progressive disorder, characterized by loss of social, communication, and motor functioning that begins at around 6–18 months of age. In its progressive nature it is similar to childhood degenerative disorder, but quite different from the other ASDs. Third, Rett is now known to be caused by X-linked genetic mutations [8], whereas the causes of the other ASDs have not been identified. Finally, whereas children with autism in particular seem to benefit from early intensive behavioral intervention, this form of treatment has been shown to be largely ineffective for children with Rett syndrome [9].

With respect to childhood disintegrative disorder, here too one sees features similar in many respects to Rett syndrome, and quite different from the other ASDs. In childhood disintegrative disorder there is a “marked regression in multiple areas of functioning following a period of at least 2 years of apparently normal development.” [7], p. 77]. In addition, most children with this disorder — as with Rett syndrome — function in the severe/profound range of intellectual disability. Children with other ASDs tend to have more variability in intellectual capabilities. It has been estimated for example that only about 50% of autistic children have IQs in the severely retarded range and many individuals who receive a diagnosis of autism appear to present with borderline or even normal IQ scores [10].

Given these differences, it would seem that at a minimum, Rett syndrome and childhood disintegrative disorder, might not be sufficiently comparable to the other ASDs. The ASD umbrella might thus be at least two disorders too wide. It is of course possible that future research might someday reveal that all of the ASDs share some common neurological/genetic basis, but it is also plausible that the etiologies of the various ASDs will prove to be quite different.

The issue of validity of the diagnostic categorization is of critical importance to researchers investigating the nature, epidemiology, causes, assessment, and treatment of ASD. Getting the validity issue right is of considerable consequence. The huge increase in the number of individuals being identified with autism may stem in part from the use of the much broader ASD umbrella [11]. For some of these individuals it is not inconceivable that their ASD label represents a serious misdiagnosis. If so then some seemingly logical rehabilitation efforts could be contraindicated, as shown by the failed attempt to treat one subtype of ASD (i.e., Rett syndrome) using procedures that have proven highly effective with another subtype (i.e., autism) [9].

To effectively search for causes of neurological disorders will require valid classification schemes. Otherwise researchers could be searching for the causes of a class of neurological impairments (e.g., ASD) that do not in fact share common causes. The search for new and more effective treatments will also be hampered if a seemingly homogeneous clinical sample actually consists of individuals with dissimilar disorders, each of which might have a varying response to treatment. Most likely the ASD label may have validity for some purposes (e.g., screening), but not for others (e.g., etiology, treatment). Research into the validity and the limiting conditions of the ASD label for various purposes is sorely needed.