The objective of this study was to determine if an intravenous infusion of synthetic human secretin improves language and behavioral symptoms in children with autism. Forty-two children with the diagnosis of autism were randomized to one of two groups in this double-blind cross-over trial. One group received 2 IU/kg of intravenous synthetic human secretin at the first visit, followed by an equal volume of intravenous saline placebo at week 6. The other group received treatments in the reverse order. All children were evaluated at weeks 1, 3, 6, 9, and 12 with standardized assessments of language, behavior, and autism symptomatology. There were no significant differences in the mean scores on any measure of language, behavior, or autism symptom severity after treatment with secretin compared to treatment with placebo. The results of this study do not support secretin as a treatment for autism.