ASD Concordance of Twins Across DSM-IV-TR and DSM-5 Diagnostic Criteria

Background:  Diagnostic criteria for autism have changed under DSM-5, with the removal of diagnostic subcategories in favor of a single category of autism spectrum disorder (ASD). Previous research has supported a high degree of concordance for ASD among monozygotic twins, but no studies have specifically examined pairwise concordance rates using DSM-5 diagnostic criteria for ASD.

Objectives:  This study aims to examine monozygotic pairwise concordance rates for ASD from the perspective of DSM-IV-TR and DSM-5 diagnostic criteria.

Methods:  Items of a parent report measure of ASD symptoms (Autism Diagnostic Interview–Revised) and a clinical observation instrument (Autism Diagnostic Observation Schedule) were matched to diagnostic criteria, as conceptualized by Huerta et al. (2012). Diagnoses of pervasive developmental disorder and ASD were assigned to 14 pairs of monozygotic twins, in whom at least one twin had an autism diagnosis, using DSM-IV-TR and DSM-5 criteria. McNemar tests were performed to determine whether there was a significant difference in sample pairwise concordance rates between DSM-IV-TR and DSM-5 diagnostic criteria.

Results:  With the use of both parent report and clinical observation measures (Autism Diagnostic Interview-Revised and Autism Diagnostic Observation Schedule), eight of the fourteen twin pairs were classified as concordant using DSM-IV-TR criteria, and five out of the fourteen twin pairs were classified as concordant using DSM-5 criteria, yielding pairwise concordance rates of 57.14% and 35.71%, respectively. The use of either parent report or clinical observation (Autism Diagnostic Interview-Revised or Autism Diagnostic Observation Schedule) resulted in pairwise concordance rates of 85.71% for DSM-IV-TR criteria and 78.57% for DSM-5 criteria. Pairwise concordance rates were not significantly different across DSM-IV-TR and DSM-5 diagnostic criteria using these two methods. DSM-5 pairwise concordance rates were significantly different when evidence of symptoms was required from either measure (78.57%) versus both measures (35.71%).

Conclusions: Pairwise concordance rates using DSM-5 diagnostic criteria were higher if diagnoses were assigned on the basis of information from either the Autism Diagnostic Interview-Revised or the Autism Diagnostic Observation Schedule than they were if information from both sources was used. The finding of a significant difference in concordance rates using either measure as opposed to both measures has important implications for diagnostic practice. Overall, despite concerns regarding changes to the diagnostic construct of ASD with the use of DSM-5 criteria, the present findings suggest that sample pairwise concordance rates for ASD have not substantially changed from DSM-IV-TR to DSM-5 criteria. The composition of the autism spectrum has remained consistent in this sample of monozygotic twins.