High-functioning autism spectrum disorder: phenotypic subgroups, diagnostic instruments, and predictors of behavioural and emotional functioning

OBJECTIVES: Autism spectrum disorder (ASD) is highly heterogeneous, with phenotypic variability and behavioural complexity greatest within higher-functioning individuals without intellectual disability (i.e., ‘high-functioning ASD’). Historically, differentiating between high-functioning ASD subgroups, Asperger’s disorder (AS) and autistic disorder without intellectual disability (i.e., high-functioning autism, HFA), has been extensively debated. While distinction between AS and HFA was found to be unreliable, the phenotypic variability in high-functioning ASD has questioned the utility of DSM-5’s dimensional diagnostic approach for this subgroup of ASD. The phenotypic heterogeneity also impacts the reliability of standardised diagnostic tools, making the clinical diagnostic process difficult and challenging in high-functioning ASD. Further to this, the high prevalence of behavioural and emotional difficulties in childhood and early adolescent years adds to the variability and clinical complexity. The association between behavioural and emotional dysfunction in high-functioning ASD, and the core ASD diagnostic features and/or other aspects of neurocognitive functioning is unclear. The current thesis comprised three main aims: (1) to investigate whether childhood high-functioning ASD subgroups could be identified based on profiles of core ASD symptomatology; (2) to evaluate the relationship between the ‘gold standard’ ASD diagnostic instruments (the Autism Diagnostic Observation Schedule-Second Edition, ADOS-2; and Autism Diagnostic Interview-Revised, ADI-R) throughout childhood in high-functioning ASD; and (3) to explore predictors of behavioural and emotional functioning within the school setting in childhood high-functioning ASD. These thesis aims were addressed in three separate studies; Aim (1) was the principle aim of this thesis, and Aims (2) and (3) were subsidiary aims. METHOD: Sixty-one children (5-14 years; 51 male) without cognitive impairment and with parent reported clinical diagnosis of DSM-IV-TR autistic disorder (n=25) or AS (n=29), or DSM-5 ASD (n=7), participated in Study One. Data from the ADOS-2 and ADI-R were subject to exploratory cluster analysis in this study. Study Two (n=57) was limited to verbally fluent participants (i.e., evaluated with Module 3 ADOS-2) and examined the level of agreement between diagnostic classifications according to the ADI-R and ADOS-2 in two childhood age groups (5-8 and 9-13 years). The relationship between ratings of social interaction (SocInt), communication (Comm), and restricted, repetitive behaviours and interests (RRBI) were also explored. Study Three (n=38) used stepwise linear regression to examine whether teacher ratings of behavioural and emotional functioning could be predicted by factors that represented ASD symptomatology and aspects of neurocognitive functioning. RESULTS: In Study One, two subgroups with unique profiles across core ASD symptom domains were identified: (i) a Severe Social Impairment subgroup with greater SocInt (ADI-R Md=18.5; ADOS-2 Md=8.5) and Comm (ADI-R Md=16.0; ADOS-2 Md=4.0) deficits, but lower lifetime severity of RRBI (ADI-R Md=3.5); and (ii) a Moderate Social Impairment subgroup with the reverse pattern of lower SocInt (ADI-R Md=14; ADOS-2 Md=5.0) and Comm deficits (ADI-R Md=12; ADOS-2 Md=2.0), but greater lifetime severity of RRBI (ADI-R Md=5). In Study Two, poor to fair agreement (κ=-.21 to .24) was found between diagnostic classifications according to the ADI-R and ADOS-2 algorithms across the two age groups examined. The strength of the association between ratings of SocInt and Comm across the measures was weaker for older (9-13 years: rs(SocInt)=.23, p>.05, rs(Comm)=.24, p>.05) compared to younger children (5-8 years: rs(SocInt)=.59, p<.01, rs(Comm)=.60, p<.01). Ratings of RRBI were not significantly correlated for either age group (5-8 years: rs=-.01, p>.05; 9-13 years: rs=.24, p>.05). In Study Three, teacher reported Externalising Problems (F=5.13, Adjusted R2=.23, p<.05), School Problems (F=9.23, Adjusted R2=.25, p<.01), and Adaptive Skills (F=10.40, Adjusted R2=.32, p<.01) were significantly predicted by one factor representing working memory, perceptual reasoning, and expressive and receptive language. In contrast, the domains of neurocognitive functioning and ASD symptomatology assessed in this study did not significantly predict Internalising Problems. CONCLUSION: This thesis found preliminary evidence supporting that phenotypic subgroups can be defined within high-functioning ASD based on unique profiles of impairment across the social communication and repetitive behaviour dimensions characteristic of the disorder. While similarities between the newly identified subgroups and previous conceptualisations of AS and HFA were evident, the clinical characteristics found to differentiate between the Moderate and Severe Social Impairment subgroups differed to those previously thought to be important discriminators between AS and HFA. Findings suggest that a combination of dimensional and categorical approaches may be informative in the classification of ASD; subdividing high-functioning ASD into subgroups would reduce the variability represented within the single diagnostic group, thereby providing greater clarity in clinical and educational settings by conveying additional information beyond the diagnostic label. Further, this thesis demonstrated that the ADI-R and ADOS-2 (Module 3) evaluate different phenotypic aspects in high-functioning ASD, particularly in older children. This variability between the ‘gold-standard’ diagnostic tests adds complexity to the diagnostic decision making process, demonstrating the need to develop assessment methods to evaluate diagnostic characteristics in this population. In particular, the ASD diagnostic algorithms require further development with high-functioning children before they can be relied upon for diagnosis. Lastly, findings suggest that cognitive and language difficulties place individuals at greater risk for teacher reported behavioural difficulties and adaptive skill deficits than ASD specific symptomatology. Therefore, cognitive and communication strategies may be more beneficial in improving functioning within the school setting than interventions targeting ASD symptomatology. Further, evidence that neurocognitive functioning and ASD symptomatology were poor predictors of internalising problems in this population suggests that screening emotional wellbeing and targeting this domain in intervention is important to ensure support needs are identified and managed