Eight weeks of non-nightly use of zolpidem for primary insomnia

Context: Intermittent use (i.e., a few nights per week) of hypnotic medicat ion is often recommended for the treatment of chronic insomnia despite an a bsence of efficacy and safety data using this regimen. Study Objectives: To evaluate the clinical efficacy and safety of intermitt ent pharmacotherapy for chronic insomnia. Design and Setting: Randomized, double-blind, placebo-controlled, parallel groups, clinical trial at six sleep research sites. Patients: One hundred-sixty-three (115 women, 48 men; mean age 44.1+ SE. 0. 9 years), DSM-IV-defined primary insomnia patients were randomized, 134 pat ients completed the study. Interventions: Eight weeks of treatment with eit her zolpidem 10 mg or placebo. Patients were instructed to take medication when they felt they needed it, but at least three and no more than five tim es per week. Main Outcome Measures: Investigator and Patient Global Ratings were the pri mary outcome variables. Secondary measures from daily questionnaires to ass ess efficacy, rebound insomnia and drug taking behavior. Results: The Investigator's Global Rating indicated that intermittent use o f zolpidem produced a significantly better therapeutic effect and significa ntly reduced insomnia severity throughout the 8-week study relative to plac ebo. Zolpidem was found to be effective in initiating and maintaining sleep on nights taken, as compared to placebo, based upon the Patient's Global R atings and all subjective sleep variables. No evidence of rebound insomnia was found on nights that zolpidem was not taken. The number of nights a pil l was taken did not differ between groups, nor did frequency of pill taking change in either group across the duration of the study. There were no sig nificant effects of treatment upon quality of life or neurocognitive measur es. Conclusions: Zolpidem 10 mg is effective in treating insomnia when used int ermittently, without evidence of discontinuation effects or increased frequ ency of pill taking.