The purpose of this study was to determine the safety and efficacy of the anticonvulsant levetiracetam in the treatment of children with autism. A previous open-label study in autistic children treated with levetiracetam demonstrated effectiveness in hyperactivity, impulsivity/aggression, and mood lability. Twenty patients with autism ranging from 5 to 17 years of age were entered into a 10-week, placebo-controlled, double-blind trial of levetiracetam versus placebo. The mean maximum dosage for levetiracetam was 862.50+/-279.19 mg/day. We evaluated global improvement of autism with the Clinical Global Impression-Improvement (CGI-I) Scale and aggression and affective instability with the Aberrant Behavior Checklist (ABC) parent and teacher ratings. We measured repetitive behaviors using the Children's Yale-Brown Obsessive-Compulsive Scale (CY-BOCS) score and impulsivity and hyperactivity with the Conners' Rating Scale-Revised: Long Version for parent and teacher. No significant difference was found between levetiracetam and placebo groups comparing the change in CGI-I (t=0.350, d.f.=13.621, P=0.765), nor on change in ABC, CY-BOCS or Conners' scales. These findings suggest that levetiracetam does not improve behavioral disturbances of autism, but are limited by the small sample size and lack of stratification of the autistic sample at baseline.