Subtyping the Autism Spectrum Disorder: Comparison of Children with High Functioning Autism and Asperger Syndrome

First, in the present study, the results suggest that in both AS and HFA, autistic core features (i.e. deficits in social communication and social interaction and restricted, repetitive patterns of behavior, interests, or activities) are present, even if with subtle differences.

About the communication features, the present study revealed a significant difference in the presence of a history of delay in language development, prior to 3 years of age, between subjects with AS and subjects with HFA, even if also in AS a mild language delay was reported. In the DSM-IV-TR, the “absence of clinically significant general delay in language (e.g., single words used by age 2 years, communicative phrases used by age 3 years)” was a relevant criterion for AS diagnosis. In contrast with this theoretical construct, extensive literature data showed that language delay could not be considered as a diagnostic exclusion criterion for AS because subtle language problems are expected in AS and also because the recognition of a language delay is not always enough reliable, since it is often dependent on parents recall (Ghaziuddin and Welch 2013). In the DSM-5, the decision to merge the single diagnostic entities in a unique broader category of ASD and the introduction of language impairment as a “specifier” solved this controversy. The present study suggests that even if language delay may not be a differentiating feature of AS/HFA, it can be useful to differentiate developmental language profiles in terms of clinical approach. In the present study, communication skills resulted overall overlapping between AS and HFA, revealing statistically significant differences only in “over-precise or pedantic speech” (characterized by overly formal speech, similar to an in-depth monologue about a topic of special interest, verbose and tangential, lack of the normal prosody), more frequent in AS, and in “literal interpretation of comments”, more common in HFA. The first finding (“over-precise or pedantic speech”) is unsurprising and represents one of the most typical clinical feature of AS, firstly described by Hans Asperger’s original work (1944, in which AS were called “little professors” for their way of talking), then included in the following proposed diagnostic criteria sets (Gillberg and Gillberg 1989; Szatmari et al. 1989) and finally confirmed by subsequent studies (Ghaziuddin and Gerstein 1996; Gilchrist et al. 2001; Macintosh and Dissanayake 2004; Woodbury-Smith and Volkmar 2009). The second finding (“literal interpretation of comments”) may be related to the HFA worse cognitive and comprehension abilities, which made them unable in recognizing the speaker’s communicative intention and in social understanding, unlike patients with AS, who adopted intellectual strategies of compensation.

Clumsiness has been originally considered a distinctive feature of AS (Ghaziuddin et al. 1994); in our experience, it appeared in both groups, without significant difference, as reported by more recent studies (Fournier et al. 2010; Nayate et al. 2012).

Even if sensory abnormalities were well-represented in our sample, no significant difference between subjects with AS and subjects with HFA was observed. Sensory abnormalities were included in DSM-5 within the second cluster of core symptoms (previously omitted). Sensory overload can be experienced as painful and correlated with lower participation in leisure activities and lower performance at school, therefore considering sensory features in intervention is essential (Grapel et al. 2015).

In this study, Full Scale Intelligence Quotient mean values were statistically significant higher in AS than HFA. This result is in accordance with previous literature data (Ghaziuddin and Mountain-Kimchi 2004; Macintosh and Dissanayake 2004; Koyama et al. 2007; Ghaziuddin 2008; Witwer and Lecavalier 2008; Noterdaeme et al. 2010; Chiang et al. 2014), although, on this topic, conflicting results were reported. Wilson et al. compared cognitive characteristics of individuals with AS and HFA reporting no significant differences on the cognitive performance (2014). Most reports suggested that patients with AS have a distinct cognitive profile, characterized by a higher verbal IQ and a lower performance IQ, whereas in most cases with HFA, the pattern is reversed (Ghaziuddin and Gerstein 1996; Saulnier and Kin 2007; Volkmar et al. 2009; Noterdaeme et al. 2010; Planche and Lemonnier 2012; Chiang et al. 2014). This profile has been linked to the better verbal abilities of AS patients. Other studies did not confirm the existence of these typical profiles, reporting mixed cognitive patterns that are not consistent enough to enable diagnostic differentiation between the two groups (Ghaziuddin and Mountain-Kimchi 2004; Macintosh and Dissanayake 2004; Ghaziuddin 2008). In the present study, on the basis of the cognitive evaluation, it is not possible to distinguish the two groups: all the IQ sub-values resulted statistically higher in AS than in HFA and differences between verbal and performance sub-values are not statistically significant within the two groups. These results support the hypothesis that the two conditions could be distinguished at a quantitative cognitive level, represented by FSIQ, rather than at a qualitative level, represented by specific cognitive profile. Nevertheless, as for intellectual disability, the presence of very different IQs could denote distinct levels of severity and clinical profile, in the same way, a large discrepancy between AS and HFA intellectual scores could have a qualitative effect on general function.

In the present study, another remarkable topic of differentiation between HFA and AS is the prevalence of learning disorders that resulted significantly more frequent in HFA, except to dysgraphia that emerged frequently in both groups, without a significant difference. Low performances in hand-writing, in terms of forming letters or using the paper space were noted in the original description of AS (Asperger 1944) and in recent literature (Jansiewicz et al. 2006; Fuentes et al. 2009). To our knowledge, until today, academic functioning and education needs in HFA have still not been sufficiently explored in literature. In accordance with our findings, Grimm et al. (2015) reported schooling of 45 patients exhibiting ASD without mental retardation, finding that patients with childhood autism and atypical autism respect to those with AS were more likely to receive the support of a special educational assistant and underwent a higher number of consultations and treatment episodes.

Rates of comorbid disorders are significantly higher in AS than HFA. The most prevalent comorbidity in both groups is ADHD, but comparing AS and HFA, its prevalence did not statistically differ. An association between ASD and ADHD and other externalizing disorders have been reported and supported by clinical population research and twin studies, suggesting substantial genetic overlap (Taylor et al. 2013, 2015; Cooper et al. 2014; Salazar et al. 2015; Antshel et al. 2016). Considering that comorbidity between these disorders was a relevant and frequent occurrence, differently from the previous edition, DSM-5 allowed the diagnosis of ADHD in the context of an ASD. People with AS displayed more anxiety and depressive symptoms compared to HFA, consistently with previous literature (Mazzone et al. 2012; Simonoff et al. 2013; Ung et al. 2013; Chiang and Gau 2015; De-la-Iglesia and Olivar 2015; Roy et al. 2015; Gillberg et al. 2016).

Overall, the results of this study suggested a complex aged-related different developmental profile between HFA and AS, characterized by quantitative and qualitative differences such as language development and communication style, academic achievement, comorbidities profile and cognitive functioning. Since language development and academic abilities play a central role in global functioning during the first years of life, HFA require more support in terms of rehabilitative treatments and school educational needs than AS. Consistently, in this study, when DSM-5 specifiers of severity levels were applied to the sample, HFA were more frequently classified in level 2 (“requiring substantial support”) than AS, while AS were more frequently classified in level 1 (“requiring support”) than HFA.

Nevertheless, despite this aged-related definite pattern of differentiation and severity level classification, in middle childhood and adolescence, AS presents higher internalizing comorbidities rate (overall anxiety and depressive disorders) than HFA, as resulted in this study. The presence of high rate of comorbidity is also underlined in previous literature (Mazzone et al. 2012; Simonoff et al. 2013; Ung et al. 2013; Roy et al. 2015; Gillberg et al. 2016). This could be explained with AS higher cognitive and communicative levels, that made them more able in introspection, more “active but odd” (Ghaziuddin 2008; Wing et al. 2011) in relations attempts, more conscious of social difficulties and so more vulnerable to affective disorders (De-la-Iglesias and Olivar 2015).

At present, DSM-5 allows to use specifiers of language impairment, intellectual impairment and severity levels to distinguish clinical profiles of ASD broader category. Nevertheless, considering the results obtained from the clinical profiles comparison and from the DSM-5 severity levels distinction, it is necessary to consider the risk that, in the severity specifier “Level 1” of the DSM-5, many phenotypic variances could be squeezed making this level very heterogeneous. Because of this, any attempts to predict clinical outcomes, develop individualized treatment targets or identify etiological factors may become unreal (Kim et al. 2015). Lai et al. (2013) underlined that defining autism using the umbrella term ASD, could hide the evident heterogeneity. In order to make progress in autism research and ultimately improve clinical practice, there is the necessity to move forward in the identification of subtypes within the autism spectrum. Maybe, a “subtype specifier” for HFA and AS could be useful to direct specific age-related paths of treatment (i.e. services of traditional rehabilitation, such as speech therapy and psychomotricity, and of assistance tools for school targeted to HFA and social training and psychotherapy targeted to AS).

Clinical judgement of qualified professional experts and the use of specific scales, designed on the clinical characteristics of HFA and AS, could provide data for this subtle distinction. In the present study, MASQ gave significantly different scores in HFA and AS, so that its application could help in discriminating AS from HFA subjects. In its original description, MASQ was used to distinguish, beyond AS and autism, high functioning spectrum disorders from other psychiatric disorders. Authors found a threshold of 22 for AS and intermediate scores for autism/PDD NOS (Ghaziuddin and Welch 2013).

In the present study, the suggested threshold is lower (18 with a sensitivity of 84.6% and specificity of 75.7%).

The second aim of this study was to verify the percentage of individuals maintaining diagnoses when DSM-5 diagnostic criteria were applied. Several reviewed studies found that DSM-5 criteria for ASD have a higher specificity but lower sensitivity as compared to the DSM-IV-TR (Gibbs et al. 2012; Worley and Matson 2012; Koh et al. 2014; McPartland et al. 2014). Recent studies indicated a percentage between 50 and 75% of individuals maintaining diagnosis under DSM-5 criteria, with the greatest decreases among high-functioning populations or previous diagnoses of PDD-NOS or AS (Kulage et al. 2014; Young and Rodi 2014; Smith et al. 2015). In the present study, 97.3% of the total sample met DSM-5 criteria for ASD; the remainder, who did not meet the full criteria, consisted exclusively of AS subjects. These AS subject, even if in presence of cognitive-emotional-social features representative of AS, lacked, at the age of re-examination, of significant clinical global impairment, indispensable to justify the diagnosis. In fact, recent theories of neurodiversity that consider AS a normal human difference rather than a disorder (Baron-Cohen 2002; Jaarsma and Welin 2012; Lorenz and Heinitz 2014), their phenotypes may be considered as a sub-clinical AS forms. Nevertheless, also approving the hypothesis of sub-clinical phenotype, in most cases, these subjects also need a sort of support (cognitive-behavioral treatment, parent/teacher psycho-education, personalization of school program). As in ADHD coding of DSM-5, regarding level 1 ASD patients, it could be suggested the use of a “partial remission” specifier, for those patients who do not meet currently the full diagnostic criteria.

As mentioned above, in the present study, the severity levels application showed a distribution of HFA in level 1 and 2, while AS were totally placed in level 1. The concordance between HFA/AS groups and level 1/level 2 groups, measured with Cohen’s K, could be considered poor. Indeed, the choice of evaluating the concordance using McNemar test was forced, since the matched classifications were not concordant: HFA/AS groups pertained to a clinical perspective, while level1/level 2 to an assistential one. In daily practice, clinicians do not strictly address type and intensity of support only to diagnoses, but they have to consider lots of variables, depending on age, comorbidities and environment.