Anxiety-like behavior in Rett syndrome: characteristics and assessment by anxiety scales

Katherine V. Barnes, Francesca R. Coughlin, Heather M. O’Leary, Natalie Bruck, Grace A. Bazin, Emily B. Beinecke, Alexandra C. Walco, Nicole G. Cantwell, Walter E. Kaufmann

Rett syndrome (RTT) is a severe neurodevelopmental disorder characterized by regression of language and motor skills, cognitive impairment, and frequent seizures. Although the diagnostic criteria focus on communication, motor impairments, and hand stereotypies, behavioral abnormalities are a prevalent and disabling component of the RTT phenotype. Among these problematic behaviors, anxiety is a prominent symptom. While the introduction of the Rett Syndrome Behavioral Questionnaire (RSBQ) represented a major advancement in the field, no systematic characterization of anxious behavior using the RSBQ or other standardized measures has been reported. This study examined the profiles of anxious behavior in a sample of 74 girls with RTT, with a focus on identifying the instrument with the best psychometric properties in this population. The parent-rated RSBQ, Anxiety, Depression, and Mood Scale (ADAMS), and Aberrant Behavior Checklist-Community (ABC-C), two instruments previously employed in children with neurodevelopmental disorders, were analyzed in terms of score profiles, relationship with age and clinical severity, reliability, concurrent validity, and functional implications. The latter were determined by regression analyses with the Vineland Adaptive Behavior Scales-Second Edition (Vineland-II) and the Child Health Questionnaire (CHQ), a quality of life measure validated in RTT. We found that scores on anxiety subscales were intermediate in range with respect to other behavioral constructs measured by the RSBQ, ADAMS, and ABC-C. Age did not affect scores, and severity of general anxiety was inversely correlated with clinical severity. We demonstrated that the internal consistency of the anxiety-related subscales were among the highest. Test-retest and intra-rater reliability was superior for the ADAMS subscales. Convergent and discriminant validity were measured by inter-scale correlations, which showed the best profile for the social anxiety subscales. Of these, only the ADAMS Social Avoidance showed correlation with quality of life. We conclude that anxiety-like behavior is a prominent component of RTT's behavioral phenotype, which affects predominantly children with less severe neurologic impairment and has functional consequences. Based on available data on standardized instruments, the ADAMS and in particular its Social Avoidance subscale has the best psychometric properties and functional correlates that make it suitable for clinical and research applications.