To investigate the effect of drugs acting on the endogenous opioid system, we studied 10 adults with Tourette's syndrome who received propoxyphene hydrochloride (260 mg/day), naltrexone hydrochloride (50 mg/day), and placebo in a double-blinded, randomized clinical trial. Using a self-report scale (Tourette's Syndrome Symptom List), subjects noted a significant (p less than 0.04) lessening of tics after treatment with naltrexone when compared with placebo. An improvement in performance on the Trail Making B test, a measure of attention and visuomotor sequencing and planning, occurred after receiving naltrexone when compared with placebo (p less than 0.08) or propoxyphene (p less than 0.02). The Trail Making B test best discriminated the treatments (p less than 0.02, analysis of variance). No other treatment effects were observed for several other measures of tic severity, attentional ability, or obsessive-compulsive symptoms. Our findings indicate that pharmacological manipulation of the endogenous opioid system does influence symptoms of Tourette's syndrome.