Measuring Callous-Unemotional Traits in Autism Spectrum Disorders

Background:  The presence of callous unemotional traits (CU) has been an important specifier in defining the severity and prognosis of individuals with Conduct Disorder. However, little is known about whether these traits can be reliably measured in individuals with ASD. The potential overlap between the ASD phenotype and children presenting as seemingly callous, unemotional and uncaring may impact on the ability to measure CU traits in this population.  The inventory of callous-unemotional traits (ICU) has been widely validated in the general population and is thought to provide an efficient, reliable measure of CU traits in young people. Although research suggests certain deficits seen in ASD (e.g. cognitive perspective taking) are qualitatively different to the deficits seen in CU (e.g. affective empathy deficits), it remains unclear as to whether these qualitative differences between CU traits and ASD can be captured in an informant report questionnaire such as the ICU.

Objectives:  To test the extent to which the ICU measures a construct above and beyond ASD trait severity. To identify the proportion of an ASD sample who are regarded as having psychopathic tendencies (as measured by the ICU) in order to inform appropriate cut offs for ICU in an ASD population.

Methods:  Cross-sectional data were examined for 56 (87.5% males) young people. A well-standardised parent report interview (3Di) was used to measure ASD. Parent and teacher report SDQ scores for conduct problems were also captured. The Inventory of Callous-unemotional Traits (ICU) parent and teacher scores were used to measure CU traits.

Results:  Correlation analysis indicates a relationship between parent reported measures of ASD severity and CU traits (Social reciprocity r =.56, df = 55, p <.05; Communication r =.40, df = p <.05; Repetitive behaviours r = .22, df = 55, p =.11, ns). However, the relationship between ASD and teacher reported CU traits is not indicated (Social reciprocity r =.084, df = 55, p =.54, ns; Communication r = -.09, df = 55, p =.50, ns; Repetitive behaviours r = -.04, df = 55, p =.79, ns). The proportion of young people with ASD reported to have CU traits (>33 on ICU) was higher than the general population (Parent report CU = 58.9%; Teacher report CU =42.9%). CD symptoms were reported in a significantly lower proportion of the sample (Parent report CD = 17.9%; Teacher reported CD = 12.5%).

Conclusions:  Using gold-standard CU trait measures in an ASD clinical sample probably leads to a significant over-reporting of CU traits. This could be partially explained by the strong relationship between ASD severity and CU traits, which may reflect conceptual and phenotypic overlap between these two constructs. Cut offs for the ICU need to be modified for an ASD population; and different items may be needed to accurately identify CU traits in the ASD population. Clinicians and researchers should be cautious using the ICU to measure CU traits in ASD populations.