Background: Chronic insomnia affects between 9 and 12 percent of U.S. adults and can impair mental health and quality of life. Pharmacologic treatments include benzodiazepines, non-benzodiazepines, and antidepressants. The only insomnia medication that is approved by the U.S. Food and Drug Administration for long-term usage is the non-benzodiazepine eszopiclone (Lunesta). Only one previous meta-analysis has evaluated these agents in the treatment of chronic insomnia. Buscemi and colleagues evaluated the effectiveness and safety of these treatments in the management of chronic insomnia in adults.

The Study: In this meta-analysis of randomized controlled trials, outcomes included sleep onset latency (the most commonly measured), wakefulness after sleep onset, sleep efficiency, and total sleep time. Methods of measurement were laboratory-based studies (polysomnography) or self-reported (sleep diary).

Results: The review included 105 trials, mostly short-term studies (≤ 4 weeks), in patients with insomnia lasting between 1.1 months and 17.7 years. The combined weighted mean difference (WMD), measured polygraphically, showed a significant reduction in sleep latency for medications compared with placebo. The WMD for benzodiazepines was −10.0 minutes (95% confidence interval [CI], −16.6 to −3.4); for non-benzodiazepines, −12.8 minutes (95% CI, −16.9 to −8.8); and for antidepressants, −7.0 minutes (95% CI, −10.7 to −3.3). The sleep latency reduction results for sleep diary were more prominent in favoring medication over placebo. For other sleep quality parameters, statistically significant results generally favored benzodiazepines and non-benzodiazepines. In antidepressant studies, only polysomnography measurements favored antidepressants; sleep diary measurements gave nonsignificant results favoring antidepressants or placebo, depending on the outcome.

A combined risk difference consistently demonstrated a higher rate of adverse events with all medications compared with placebo. These included dizziness, fatigue, headache, nausea, and somnolence. No major adverse events such as falls or death were reported. In indirect comparisons of the three medication categories, differences in effectiveness were nonsignificant. Non-benzodiazepines had a significantly better safety profile than benzodiazepines.

Conclusion: Benzodiazepines and non-benzodiazepines are effective in treating chronic insomnia. Indirect comparison suggested a greater safety profile for non-benzodiazepines. Antidepressants also appear to be effective, although studies are lacking. The findings are limited by the short-term nature of the studies and the publication bias likely associated with the private funding of many benzodiazepine and non-benzodiazepine studies. Further research is needed to study the effects of these medications on daytime functioning and their use by vulnerable groups, such as older adults.