In a double-blind placebo-controlled study, following a 1 week washout, 75 outpatients suffering from generalized anxiety disorder with severe insomnia as the target symptom were randomly assigned to 4 weeks of treatment with zopiclone 7.5 mg, triazolam 0.5 mg or placebo at bedtime. Zopiclone was significantly better than placebo on most sleep parameters. Triazolam tended to be superior to placebo, but its superiority was significant only on the sleep induction factor. Triazolam-treated patients presented significantly more day-time-interdose anxiety than zopiclone as assessed by the weekly HARS and Clinical Global Assessment of Anxiety. Although daytime-interdose anxiety was observed with both drugs, this treatment emergent symptom was more frequent and severe with triazolam. Side-effects were of a mild to moderate intensity for both zopiclone and triazolam; however, taste perversion frequently appeared with zopiclone. Although both drugs share similar pharmacological properties and bind to benzodiazepine receptors, they differ significantly with respect to side-effects and daytime anxiety.