BACKGROUND: insomnia is a common problem encountered by family physicians. Benzodiazepines have been the treatment of choice despite questions about their efficacy and their potential to cause adverse effects, such as confusion, falls, and memory loss. The authors of this systematic review summarize the literature on the efficacy and common adverse effects of bezodiazepines compared with both placebo and other treatments for insomnia.
POPULATION STUDIED: Using MEDLINE and the Cochrane Controlled Trials Register, the authors identified randomized controlled trials of benzodiazepines compared with placebo or other active drugs for treatment of insomnia. Only studies in English were included in the evaluation. The manufacturers of the medications were contacted for any studies that may not have been published. The 45 studies included in the meta-analysis represented 2672 patients, 47% of whom were women. The mean age of the patients studied ranged from 29 to 82 years. The duration of the studies ranged from 1 day to 6 weeks with a mean of 12.2 days and median of 7.5 days. The drugs studied included flurazepam (14 studies), temazepam (13), midazolam (5), nitrazepam (4), estazolam (2), and 1 each used lorazepam, brotizolam, quazepam, loprazolam, and flunitrazepam. Nonbenzodiazepines studied included zopiclone, diphenhydramine, glutethimide, and promethazine.
STUDY DESIGN AND VALIDITY: This was a meta-analysis that combined data from the 45 studies that met the inclusion criteria. The quality of the individual reports was assessed on a scale from 0 to 5 on the basis of the quality of randomization, blinding, follow-up, and control for baseline differences in the groups. Interrater reliability was assessed and overall agreement was 98%. The meta-analysis of end points was appropriately limited to those presented in a comparable way. Tests for homogeneity were applied, and if study results were heterogeneous, the studies were subdivided into predetermined groups. This was a well-executed meta-analysis.
OUTCOMES MEASURED: The primary outcomes reported were sleep latency, sleep duration, and patient reports of adverse effects.
RESULTS: The pooled mean difference in sleep latency between benzodiazepines and placebo was 4.2 minutes (95% confidence interval [CI], -0.7 to 9.2 minutes). These data were taken from 4 studies of 159 patients. Two studies of 35 patients measured total sleep duration and were combined; patients in the benzodiazepine group slept 61.8 minutes (95% CI, 37.4-86.2) longer than those given placebo. Patient estimates of sleep latency were measured in 8 studies of 539 patients. Benzodiazepines were superior to placebo, with a perceived reduction in sleep latency of 14.3 minutes (95% CI, 10.6-18.0). Seven studies involving 821 patients evaluated patient reporting of adverse drug effects. The authors found that patients randomized to the benzodiazepine group were more likely than those taking placebo to report adverse drug effects (odds ratio=1.8; 95% CI, 1.4-2.4). The absolute rates of any adverse event during a week of treatment were 59% for patients taking benzodiazepines and 44% for those taking placebo (number needed to harm=6.8). Zopiclone was compared with a benzodiazepine in 3 studies of 96 patients. There was no difference between these drugs in sleep latency or total sleep duration, but there was no comparison with placebo. Only one study followed patients for more than 2 weeks.
This meta-analysis shows that benzodiazepines help patients sleep longer, but it also shows that adverse drug effects are more likely than with placebo. The short duration of most of the studies, usually 2 weeks or less, limits the usefulness of these data because treatment of insomnia is often needed for much longer periods. Benzodiazepines are reasonable for short-term use while further research is done to clarify the most efficacious treatment for insomnia. The authors point out many areas where further research is needed, especially in the area of nonpharmacologic therapy.