Article Text

Download PDFPDF

Review: benzodiazepines increase sleep duration but also lead to adverse effects in adults with insomnia
  1. Toshiaki Furukawa, MD, PhD
  1. Nagoya City University Nagoya, Japan

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

OpenUrlAbstract/FREE Full Text

QUESTION: In adults with insomnia, are benzodiazepines effective and safe for improving sleep outcomes?

Data sources

Studies were identified by searching Medline (1966–98), the Cochrane Controlled Trials Register, and reference lists, and contacting manufacturers of benzodiazepines.

Study selection

English language studies were selected if they were randomised controlled trials (RCTs) that compared a benzodiazepine with a placebo or another active drug in patients with insomnia.

Data extraction

RCTs were assessed for the quality of study methods. Data were extracted on study design, conditions treated, patient characteristics, setting and duration of RCT, and outcomes (including adverse effects).

Main results

45 RCTs (2672 patients) (mean age range 29–82 y, 53% men) met the selection criteria. Study duration ranged from 1 day to 6 weeks (median 7.5 d). Patients who received benzodiazepines had greater sleep duration than those who received placebo (2 RCTs) (table). A statistically non-significant decrease occurred in benzodiazepine group sleep latency (8 comparisons in 4 RCTs) (table). Patient estimates of sleep latency were shorter in the benzodiazepine group than in the placebo group (8 RCTs), but heterogeneity existed. When RCTs were grouped by quality of study methods, heterogeneity was no longer statistically significant; the effect was smaller for high quality RCTs than for low quality RCTs (table). Patient estimates of sleep duration were greater in the benzodiazepine group than in the placebo group (8 RCTs) (table). Overall adverse effects (7 RCTs) and daytime drowsiness (8 RCTs) were more common in the benzodiazepine group than in the placebo group; dizziness or lightheadedness was also more common in the benzodiazepine group (4 RCTs), but this difference was not statistically significant (table). Zopiclone was not superior to benzodiazepines on any outcome measured.

Benzodiazepines v placebo in adults with insomnia*


Benzodiazepines improve sleep duration but also lead to adverse effects in adults with insomnia.


What information do we need to consider when treating patients with insomnia? We need to know (1) by how much a medication is better than placebo or non-pharmacological treatment for improving objective measures of sleep, subjective quality of sleep, and overall quality of life; and (2) by how much that medication is worse than placebo or non-pharmacological treatment with respect to drowsiness, cognitive impairment, dependence, tolerance, accident proneness, teratogenicity, and cost. Furthermore, we need to know these estimates not only for several days but also for much longer periods of time.

After reading this methodologically rigorous, comprehensive meta-analysis by Holbrook et al, are we in a better position to answer these questions? Firstly, this meta-analysis does not show how benzodiazepines affect the patients' quality of sleep or quality of life because these were not measured in the original RCTs. Secondly, no comparison between benzodiazepines and non-pharmacological alternatives was presented. Finally, none of the meta-analysed RCTs examined long term use. The authors note all of these problems.

Importantly, a few RCTs are now available that directly compare pharmacotherapy and non-pharmacological treatment. These studies suggest a greater commitment of time, but equal effectiveness and superior durability for behavioural techniques.13 Some problems of benzodiazapine pharmacotherapy can now be quantified. A cohort study suggested that approximately 1 in 2900 people will be admitted to hospital for a traffic accident within 2 weeks after their first prescription of a benzodiazepine.4 For every 100 elderly people who take benzodiazepines, approximately 1 person is admitted to hospital for a fall.5

As a clinician, I must confess that it is difficult to balance this information against an approximate 1 hour increase in total sleep time as well as increased drowsiness and dizziness as shown by this meta-analysis. I concur with the authors that we need further pragmatic studies before we can make strong evidence-based recommendations.



  • Sources of funding: in part, Canadian Pharmaceutical Association and Canadian Medical Association.

  • For correspondence: Dr A M Holbrook, Centre for Evaluation of Medicines, St Joseph's Hospital, 50 Charlton Avenue East, Hamilton, Ontario L8N 4A6, Canada. Fax +1 905 521 6136.