Current issues of ACP Journal Club are published in Annals of Internal Medicine


Use of long-half-life benzodiazepines in older persons was associated with increased risk for injurious motor vehicle accidents

ACP J Club. 1998 Jan-Feb;128:15. doi:10.7326/ACPJC-1998-128-1-015

Source Citation

Hemmelgarn B, Suissa S, Huang A, Boivin JF, Pinard G. Benzodiazepine use and the risk of motor vehicle crash in the elderly. JAMA. 1997 Jul 2; 278:27-31.



To determine whether use of long-half-life and shorter half-life benzodiazepines is associated with risk for injurious motor vehicle accidents in older persons.


Nested case-control study in a cohort followed from 1990 to 1993.


Population-based study in Quebec, Canada.


5579 drivers involved in a motor vehicle accident in which ≥ 1 persons sustained injury, and 55 790 control person-days were selected from a cohort of 19 686 eligible drivers between 67 and 84 years of age in the province of Quebec who had a valid driver's license and had lived in Quebec for at least the past 2 years. Exclusion criteria were residence in a long-term care facility, hospitalization in the past 60 days, or hospitalization for ≥ 30 days in the past year.

Assessment of risk factors

Data on benzodiazepine use were taken from a provincial prescription drug database. Benzodiazepines were classified as having a long elimination half-life (> 24 h) (clonazepam, diazepam, clorazepate, chlordiazepoxide, flurazepam, nitrazepam) or a shorter elimination half-life (≤ 24 h) (alprazolam, bromazepam, lorazepam, oxazepam, temazepam, triazolam). Duration of exposure was classified as 1 to 7 days, 8 to 30 days, 31 to 60 days, or 61 to 365 days. Age, sex, residence, chronic disease score, benzodiazepine dose, other benzodiazepine or central nervous system drug use, and previous motor vehicle accident involvement were controlled for in the analysis.

Main outcome measure

Risk for injurious motor vehicle accidents extracted from police reports.

Main results

Prevalence of exposure to long-half-life benzodiazepines was 6.9% for persons who had accidents and 5.2% for control participants; exposure to shorter half-life benzodiazepines was 14.4% and 14.7%, respectively. Use of long-half-life drugs was associated with an increased risk for accidents (adjusted relative risk [RR] 1.28, 95% CI 1.12 to 1.45), but use of shorter half-life drugs was not (RR 0.96, CI 0.88 to 1.05). Risk for accidents increased during the first 7 days of use of long-half-life drugs (RR 1.45, CI 1.04 to 2.03) and with continued use lasting 61 to 365 days (RR 1.26, CI 1.09 to 1.45).


Use of long-half-life benzodiazepines, but not shorter half-life benzodiazepines, was associated with increased risk for involvement in injurious motor vehicle accidents among older persons. Sub-analyses showed an increased risk during the first 7 days of use and with continued use lasting 61 to 365 days.

Source of funding: National Health Research and Development Program of Canada.

For article reprint: Dr. S. Suissa, Division of Clinical Epidemiology, Royal Victoria Hospital, 687 Pine Avenue West, Ross 4.29, Montreal, Québec H3A 1A1, Canada. FAX 514-843-1493.


Despite widespread publicity about their risks, benzodiazepines are still widely prescribed. Although diminished performance of many tasks, including driving, can be shown, the clinical importance of this is uncertain. We need to know the nature and magnitude of the risks these drugs create and whether such risks are generic to the family of drugs or limited to certain members.

The study by Hemmelgarn and colleagues examined the potential link between benzodiazepine use and motor vehicle accidents. Accidents that result in injury are uncommon (about 1/120 driver-y in this study) and therefore difficult to study prospectively. The ability to link routinely collected data about benzodiazepine prescriptions and motor vehicle accidents allowed a large volume of data to be analyzed.

The drivers involved in accidents were primarily men, presumably because they are more likely to be driving than women. Observational data such as these are susceptible to selection bias, which may explain the association measured. In this case, it may be that the reason for prescribing the benzodiazepine (e.g., dementia, delirium, agitation, or sleep disorders) is associated with the motor vehicle accident rather than the drug itself. The trend for case participants to be slightly older, use more central nervous system drugs of other types, and have more previous motor vehicle accidents and the lack of a dose-response gradient supports this notion.

Nonetheless, this study adds further evidence that the longer-acting benzodiazepines may have harmful effects on the performance of elderly drivers. Despite years of warning about the potential harm of benzodiazepines in elderly persons, we wonder why as many as 36% of elderly Quebec drivers require sedation.

Unfortunately, the designation of 24 hours as the division between long- and shorter-acting benzodiazepines is arbitrary and switching patients from long-acting to shorter-acting drugs has not been shown to improve safety. Thus, while we await definitive randomized controlled trials or meta-analyses, benzodiazepines remain controversial, as their frequent prescription contrasts with their uncertain risk-to-benefit profile.

John D. Holden, MB
Ian Hughes, MBThe Medical CentreHaydock, St. Helen's, United Kingdom
Anne Holbrook, MD, PharmD, MSc
McMaster UniversityHamilton, Ontario, Canada