Developmental Regression In the Simons Simplex Collection

Background: It is generally recognized that the development of autism spectrum disorders (ASD) is not uniform across children. Currently, two patterns of ASD onset are cited: early onset, which is characterized by delays and/or unusual patterns of development from birth or shortly thereafter, and regressive onset, in which children experience a period of seemingly typical development, followed by a loss of previously acquired skills. Regressive onset has been reported in approximately one third of children with ASD. However the duration of loss, the types of skills lost, and the length of time that children had the skills prior to losing them are variable. The “gold standard” for assessing regression has been parent report, and the Autism Diagnostic Interview—Revised (ADI-R) is a commonly-used measure for this purpose. Recently, additional questions were created to capture subthreshold skill losses that do not meet the original ADI-R criteria for full losses. New prospective studies suggest that a majority of children with ASD experience regression, particularly in social engagement, but this is often missed by current ADI-R questions. The expanded questions may allow detection of more subtle losses that would not have been captured previously.

Objectives: (a) To provide descriptive information about the types and duration of skill losses among a large sample of children with ASD; (b) To determine whether children differ in terms of cognition, communication, indices of adaptive functioning, and/or ASD symptom severity based on whether they experienced full, subthreshold, or no skill losses.

Methods: Data from the Simons Simplex Collection (SSC; https://sfari.org/simons-simplex-collection) were analyzed for probands with ASD who received both the ADI-R and the expanded loss insert (N =1657). Frequencies of regressions were calculated by type of skill loss and whether they were full losses, as defined by the original ADI-R questions, or subthreshold losses detected by the expanded questions. T-tests were used to compare ages of onset for and duration of full and subthreshold losses. One-way ANOVAs were computed to determine effects of type of loss (full, subthreshold, no loss) on cognition, communication, indices of adaptive functioning, and ASD severity.

Results: Full losses were experienced by 17.1% for language and by 28.8% for other skills, whereas an additional 11.7% experienced subthreshold losses in language and 3.1% in other skills. Subthreshold losses appeared at significantly later ages than full losses. While the duration of other-skill losses was significantly longer for subthreshold losses than full losses, this was not the case for language losses. Children who experienced any degree of loss scored significantly lower than those without loss on indices of cognition, communication, and adaptive functioning but not ASD severity.

Conclusions: The original ADI-R questions may not capture subtle regressions or other, non-word language losses. While the expanded questions increased the number of regressions detected, regressive onset continued to be reported less frequently than early onset. Future research should (a) compare the expanded loss-insert questions prospectively with observational data to further explore their accuracy and (b) examine the influence of any degree of loss on developmental outcomes.