Evaluating placebo responses to intranasal oxytocin in autism: findings from the placebo lead-in phase of a randomised controlled trial.

Abstract

BACKGROUND The placebo effect is established in clinical trials, but for paediatric research, questions remain about how to best manage its influence. Within the autism field, data on these issues is sparse. This is particularly important in the oxytocin field where placebo responses are thought to play an important role. This study reports on data from the single-blind, placebo lead-in phase of a randomised controlled trial (RCT) to investigate the placebo response and its relationship to treatment response in autistic children. METHODS Eighty-seven autistic children aged 3-12 years (M = 7.27; SD = 2.69; 85.1% male) were consecutively recruited into a multi-site RCT evaluating the efficacy of oxytocin for improving social responsiveness. Participants underwent a 3-week, single-blind placebo lead-in before randomisation into a 12-week double-blind treatment phase (oxytocin, n = 45; placebo, n = 42). The Social Responsiveness Scale, 2nd Edition (SRS-2) Total Raw Score was used to measure change from baseline to post-placebo lead-in. A ≥10-point improvement defined placebo responders. RESULTS Nearly half the sample (n = 42, 48.3%) were identified as placebo responders during the lead-in phase, showing a clinically significant degree of change on the SRS-2. Caregiver treatment guess did not significantly impact the placebo response (p = .534). Placebo response was associated with greater symptom severity (r's > -.23; p-values < .037) and higher cognitive ability (r = -.35, p = .004). Smaller placebo responses during the lead-in phase were associated with larger responses during active treatment in participants receiving oxytocin (r = -.36, p = .017). Placebo responses during the lead-in phase were observed across all caregiver-reported measures (Cohen's d = .19-.65). CONCLUSIONS This study provides important information about placebo effects and placebo lead-in designs for clinical trials in the autism field. We show widespread clinically significant improvement during placebo lead-in, utility of identifying placebo responders for informing clinical trial analyses, similarities in symptom measure effect sizes for placebo effects, and a lack of influence of caregiver beliefs on placebo responses.