Association Between Regression and Self Injurious Behaviors Among Children with ASD

Background: Self injurious behaviors (SIB) are challenging clinical issues in individuals with autism spectrum disorders (ASD).  McDermott, et al (2008) found that the relative risk of SIB in autism was 7.6 times greater than in typical children.  Risk factors for SIB in developmentally disabled children include male gender, severe intellectual disability, an ASD, and communication disorders (McClintock, et al 2003).  Previous studies show that children with increased severity of autism, lower levels of speech, adaptive skills and younger age are more likely to have SIB (Baghdadli et al 2003).  Developmental regression has not been previously studied as a potential risk factor in self injury.  This study is one of the first and largest to utilize inpatient data to look at risk factors, specifically regression, associated with different types of SIB.

Objectives:  The objectives of the study are to describe the frequency, type and severity of SIB in children with an ASD, and to examine the association of SIB and regression in ASD.

Methods: Patients were eligible if they were treated as inpatients on the neurobehavioral unit (NBU) between January 1995 to December 2004.  Subjects included were between three and seventeen years of age.  Charts were reviewed if they included admitting diagnoses of SIB and a diagnosis of autism or ASD.  Medical records of 132 patients were reviewed; data were collected on the type and frequency of SIB and presence of language, social, or behavioral regression.  Other information gathered were global assessment scale scores, level of intellectual disability, and other medical diagnoses.

Results:  Of the 132 patients, 97 (73.5%) had a diagnosis of autistic disorder and 35 (26.5%) of ASD.  Information on language regression was available for 105 patients, social regression for 81, and behavioral regression for 79.  Among those with data on regression, 42 (31.8%) reported language regression, 26 (19.7%) had social regression, and 22 (16.7%) had behavioral regression.  SIB were not significantly associated with language regression; however, patients with social regression were more likely to have skin picking/pinching (Odds ratio (OR) = 3.11, 95% CI: 1.03-9.36), but less head banging (OR = 0.27, 95% CI: 0.08-0.90), as compared to those who did not have social regression.  Those with behavioral regression showed significantly less hitting (OR = 0.34, 95% CI: 0.12-0.96), elopement (OR = 0.30, 95% CI: 0.10-0.94), and aggression (OR=0.03, 95%CI: 0.003-0.22) as compared to those without behavioral regression.

Conclusions: In children with autistic disorder or an ASD, children with a history of language regression did not have a significant association with SIB.  Children with a history of social regression were more likely to have SIB of skin picking/pinching.  Children with a history of behavioral regression had significantly less hitting, elopement, and aggression.  In our continuing studies, we will look at the frequencies and subtypes of SIB and other medical diagnoses in autism and ASD.