Regression Histories in Young Children with Autism: Timing and Associations

Background: Developmental regression is not uncommon in autism, yet little is known about the timing and acuity of the loss of skills.  In particular, the acuity at which loss occurs (i.e. sudden versus gradual) has rarely been studied.  The acuity of regression may provide clues to etiology or pinpoint associated events, such as physical illness, psychosocial trauma or other environmental factors. 

Objectives: To explore the timing and acuity of language and social losses in young children with autism, and to determine if physical illnesses/medical conditions or psychosocial stressors were associated with the developmental regressions.

Methods: Parents of 185 children (146 autism, 39 PDD-NOS) with a mean age 53 months (range 21-153 months) were interviewed using the Autism Diagnostic Interview-Revised [ADI-R; a Toddler version of the ADI was administered for children under age 48 months (n=93)]. For this study, regression was defined as loss of at least 5 words and/or a significant loss of social engagement/responsiveness (score of 2 on ADI-R).  Acuity of loss was explored through a question on the Early Development Questionnaire (EDQ; Ozonoff et al., 2005), “Was your child’s loss of skills gradual or sudden?  Please describe.” Potential relationships with physical illnesses or psychosocial stressors were explored through answers to open-ended questions on the Early Development Questionnaire.

Results: Twelve of 185 children met criteria for language loss only, 30 children met criteria for social engagement/responsiveness loss only, and 38 met criteria for loss in both areas. Mean age of language loss was 20.7 months (SD  = 6.7, range = 12-42), and mean age of loss of social engagement was 18.8 months (SD =6.3, range, 10-51 months).

Among the 80 cases with significant regression, parents of 26 children indicated physical illness/medical issues may have been related to loss of skills, and 21 responses indicated an associated  psychosocial stressor (in 7 of these cases, both illness and psychosocial stressors were reported).  However, only 8 reports clearly indicated a sudden loss of skills. Five of the 8 cases with sudden loss included both loss of language loss and social engagement/responsiveness.  Two of the sudden loss cases were reported to be associated with a physical illness/medical issue and 3 were reported to be associated with a psychosocial stressor (with 1 reported to have both).

Conclusions: Most regression was reported to have been gradual in onset, making it difficult to identify illnesses or events that may have precipitated the developmental losses.  Although a causal relationship can only be determined by prospective study of at-risk children, further clues to the nature, timing and events associated with regression may be provided by examination of the children’s medical records, which is currently underway.

.