Stability and Predictors of the Developmental Course of ASD From Childhood to Adolescence

Background: Studies regarding the stability of ASD from childhood until adolescence are sparse. The few studies available mostly included individuals with classical autistic disorder (AD) combined with mild to severe mental retardation, and concluded that autism is stable throughout life. Currently, the conceptualisation of autism is much broader and more insight is needed on the developmental course of individuals at the higher functioning end of the autism spectrum (ASD).

Objectives: The first aim of this study was to examine the diagnostic stability of ASD from middle childhood to adolescence. The second aim was to investigate putative predictors from childhood for further developmental course into adolescence.

Methods: A follow-up study was performed on a sample of originally 242 clinically referred, 6 to 12 year old children (T1: n=142 ASD, n=100 with subclinical ASD symptoms). Seven years later, at T2, 170 adolescents (n=113 of the original ASD group and n=57 of the ‘sub-ASD’ group) took part in diagnostic assessment with the Autism Diagnostic Observation Schedule (ADOS) and Autism Diagnostic Interview (ADI-R). Four developmental course groups were computed based on best estimate clinical consensus classifications during childhood and adolescence; 1) persistent ASD, 2) individuals who improved from an ASD classification to subclinical ASD symptoms (‘improvers’), 3) persistent sub-ASD, 4) individuals who declined from subclinical ASD symptoms to an ASD classification (‘decliners'). The relative amount of individuals in these groups was regarded (aim 1). To evaluate predictors of developmental course (aim 2), T1 cognitive functioning, language abilities, and severity within symptom ASD domains were compared between the persistent ASD group and improvers (predictors of improvement) and between the persistent sub-ASD group and decliners (predictors of decline).

Results: 74% of the individuals received an ASD classification at both time points, whereas in 26% of the cases functioning improved. 68% of the participants who showed subclinical ASD symptoms during childhood also did not have an ASD classification during adolescence, but 32% declined into a full ASD classification. The individuals who improved showed better language and communication skills during childhood than the persistent ASD group. Individuals that declined into an ASD classification had worse language and communication skills during childhood than the individuals with subclinical ASD symptoms during both assessment waves. Social problem levels during childhood were also marginally different between the groups (p=.06). Cognitive functioning was not predictive of improvement or decline.

Conclusions: In higher functioning individuals, ASD is not as stable from childhood to adolescence as was previously found for more severe autism cases. More than one fourth of the ASD cases no longer meet diagnostic criteria in adolescence, but also one third of the individuals that did not meet full diagnostic criteria during childhood, did meet these criteria during adolescence. Therefore, it is important to re-evaluate the diagnosis of individuals who showed milder forms of ASD during childhood once they have reached adolescence. Language ability is an important prognostic factor regarding further developmental course into adolescence, since limited speech predicts decline, but good pragmatic language ability predicts improvement.