"The Phenomenon of Loss." Early Development and Functional Outcome in Children with Autism Spectrum Disorder and Reported Developmental Regression.

Background:  Although in most children with ASD an early onset of symptoms can be observed, about one-third appears to show a loss of previously established skills somewhere in the second year of life, a phenomenon called regression. Recently, additional patterns of onset have been described such as early onset+regression and plateau. However, due to methodological issues and different views on the definition of regression, still little is known about early pathways, causes, predictors and prognosis.

Objectives: The first purpose is to explore the trajectories of different onset patterns in ASD. Second, we want to examine if pre-regression development is really typical. Third, outcomes in terms of severity of ASD-symptoms, language and behavioural functioning will be investigated.

Methods: Participants are 100 children (4-10y) with a diagnosis of ASD. Parent report is used to measure early ASD-symptoms, classification to onset groups (EDQ, ADI-R, RSQ) and current behavioural problems (CBCL). Non-verbal intelligence is measured through the WNV. Current severity of ASD-symptoms is both reported by parents (SCQ, SRS) and clinically evaluated (ADOS-2). Outcomes in language reception and production are measured through the CELF-4 and by parent report (NCDI).

Results: Preliminary results based on 54 children revealed an early onset (n=23), a later onset (n=15) and an early onset+regression group (n=6). Other children had a typical development followed by regression (n=8) or plateau (n=3). Regression involved in 92% of the children loss of language skills, in most cases combined with loss of other skills such as social skills. In 38% of the children loss or a stagnation in motor skills (mean onset: 18m) was reported to precede loss in other skills (mean onset: 24m).
Analysis of early development shows similar scores of social and stereotyped behaviours in the first 18 months of life in both the regression (ASD-R) and non-regression group (ASD-NR). Further, ASD-R showed even less early communicative behaviours (U(50)=105.5;p<.01) and received the clinical diagnosis of ASD at a younger age than ASD-NR (38m vs 67m; t(52)=4.2;p<.001).
Analysis of current functioning shows lower non-verbal intelligence scores (67 vs 92; U(54)=83.5;p<.001) in ASD-R. Examination of current severity of ASD-symptoms and behavioural problems based on parent report shows no significant differences. However, on the ADOS-2 there are significantly more restricted and repetitive behaviours (U(54)=486;p<.001) in ASD-R, especially more sensory interests (χ²(1)=14.6;p<.001) and mannerisms (χ²(1)=27.6;p<.001). Analysis of current language outcome shows both an increased delay in receptive and expressive language (37m vs 11m; t(51)=3.3;p<.01 and 32m vs. no delay; t(51)=3.6;p<.01, respectively) in ASD-R.

Conclusions: We found support for different onset patterns previously described in the literature. Since some parents report motor atypicalities preceding loss of other skills, some evidence is provided for the over-pruning hypothesis as a neuropathological mechanism of regression. Further, the majority of ASD-R appears to have pre-existing social-communicative difficulties and received their clinical diagnosis earlier than ASD-NR, supporting that regression may provide a useful “red flag” in identifying children who are at risk for ASD. Furthermore, ASD-R display more severe impairments later in life as measured by non-verbal IQ, clinical evaluation of ASD-symptomatology and language abilities.