Saturday, May 22, 2010: 2:30 PM
Grand Ballroom AB Level 5 (Philadelphia Marriott Downtown)1:15 PM
Background: As part of the STAART network funded by the National Institutes of Health, a multi-site clinical trial designed to assess both the efficacy and safety of the SSRI, citalopram, for a population of children and adolescents with autism and high levels of repetitive behavior was conducted. Objectives: Citalopram did not show benefit over placebo in this trial, however a third of each treatment group did experience clinical improvement. We thus sought to evaluate whether there were factors that predicted the likelihood of a placebo response at study entry. Methods: One hundred forty-nine subjects 5 to 17 years old inclusive were randomized to receive citalopram (n = 73) or placebo (n = 76). Participants had autistic spectrum disorders, Asperger disorder, or pervasive developmental disorder, not otherwise specified; illness severity ratings of moderate or greater on the Clinical Global Impressions, Severity of Illness Scale; and moderate or greater scores on the Children's Yale-Brown Obsessive Compulsive Scales modified for pervasive developmental disorders (CY-BOCS). The study lasted 12 weeks achieving a mean final dose of citalopram of 16 mg daily. Bivariate associations between responders and non-responders for potential baseline predictors were performed as well as between responders and non-responders within treatment group. These potential predictors included demographic factors such as IQ, age, race, gender, autism clinical global severity rating (CGI-S), behavioral ratings from the Aberrant Behavior Checklist, the Child and Adolescent Symptom Inventory, Repetitive Behavior Scale, Caregiver Strain Index and Vineland Adaptive Behavior Scale. Results: A number of significant differences emerged which differentiated placebo and citalopram responders from non-responders at baseline including scores on the hyperactivity and irritability subscales of the ABC, scores on the CASI ADHD items, and Caregiver Strain. In these and other items that will be presented, subjects who went on to experience a positive treatment response tended to have lower scores at study entry than did non-responders. There was no evidence of randomization bias between citalopram and placebo groups for any of these measures. Conclusions: These findings suggest that it may be possible to reduce the likelihood of nonspecific responses to treatment in clinical trials by identifying and controlling for symptom burden and other factors at study entry.