Developmental Regression In Children with Autism Spectrum Disorders

Background: Recent studies have found that as many as 50% or more of children with Autism Spectrum Disorders (ASD) have a history of regression in development, most commonly in language, social skills, or both. Prevalence of regression depends on the sample studied and the definition used (Wiggins et al., 2009).

Objectives: This study describes developmental regression in a sample of children with ASD identified through the South Carolina Autism and Developmental Disabilities Monitoring program (SC-ADDM), which is part of the larger Centers for Disease Control and Prevention’s (CDC) ASD monitoring program.  Specifically, we examine (1) prevalence; (2) timing; (3) skills affected; (4) characteristics of affected children; and (5) subsequent developmental outcome.

Methods: Children with ASD born in 1992, 1994, 1996, and 1998 were identified via the SC-ADDM network methodology during study years 2000, 2002, 2004, and 2006. Records of those children who had been identified through SC-ADDM as having a history of regression were requested, and details regarding regression were re-coded to better characterize the age at which the child lost skills, the type of skills lost, and the child’s developmental status prior to the regression. For this study, regression is defined as a loss (not simply a plateau in skill acquisition) of previously acquired skills in any domain not due to a clearly identified precipitating event such as physical or psychological trauma.

Results: Results indicated that 20.7% of children with ASD in this sample had documentation of developmental regression in their records (126/609). The mean age of regression was 22.8 months (SD ± 14), with a range of 6 – 84 months.  Regression was most commonly reported in the domain of language/communication skills, followed by social skills. Regression in motor and adaptive skills was rarely reported. Significant differences were noted for intellectual disability status (X² = 15.8, P<0.0001), community diagnosis (X²=35.6, p<0.0001), and educational placement at age 8 (X²=45.1, p<0.0001).  Gender and race were not associated with presence/absence of regression.

Conclusions: Regression was documented in one out of five children with ASD in this sample. Children who evidenced developmental regression were more likely to meet criteria for an intellectual disability. They were more likely to have an educational classification of Autism, and were more likely to have a community ASD diagnosis than children with ASD without a history of regression. No increased risk appeared to be associated with the child’s gender or race. These results may suggest that a history of regression may lead educators and clinicians to consider an ASD diagnosis.  This is one of the larger samples of children with ASD in which developmental regression has been studied.