OBJECTIVE: To systematically review the benefits and risks associated with the use of benzodiazepines to treat insomnia in adults. DATA SOURCES: MEDLINE and the Cochrane Controlled Trials Registry were searched for English-language articles published from 1966 to December 1998 that described randomized controlled trials of benzodiazepines for the treatment of insomnia. Key words included "benzodiazepines" (exploded), "randomized controlled trial" and "insomnia." Bibliographies of relevant articles were reviewed for additional studies and manufacturers of benzodiazepines were asked to submit additional randomized controlled trial reports not in the literature. STUDY SELECTION: Articles were considered for the meta-analysis if they were randomized controlled trials involving patients with insomnia and compared a benzodiazepine with placebo or another active agent. Of the 89 trials originally identified, 45 met our criteria, representing a total of 2672 patients. DATA EXTRACTION: Data were extracted regarding the participants, the setting, details of the intervention, the outcomes (including adverse effects) and the methodologic quality of the studies. DATA SYNTHESIS: The meta-analyses of sleep records indicated that, when compared with placebo, benzodiazepines decreased sleep latency by 4.2 minutes (non-significant; 95% confidence interval (CI -0.7 to 9.2) and significantly increased total sleep duration by 61.8 minutes (95% CI 37.4 to 86.2). Patient-reported outcomes were more optimistic for sleep latency; those randomized to benzodiazepine treatment estimated a sleep latency decrease of 14.3 minutes (95% CI 10.6 to 18.0). Although more patients receiving benzodiazepine treatment reported adverse effects, especially daytime drowsiness and dizziness or light-headedness (common odds ratio 1.8, 95% CI 1.4 to 2.4), dropout rates for the benzodiazepine and placebo groups were similar. Cognitive function decline including memory impairment was reported in several of the studies. Zopiclone was not found to be superior to benzodiazepines on any of the outcome measures examined. INTERPRETATION: The use of benzodiazepines in the treatment of insomnia is associated with an increase in sleep duration, but this is countered by a number of adverse effects. Additional studies evaluating the efficacy of nonpharmacological interventions would be valuable.