NPC Archive Item: Z-drugs increase the risk of road traffic accidents

NOTE – This is an archive post from the NPC and has not been updated since first publication. Therefore, some hyperlinks may no longer be working.
MeReC Rapid Review NPC Logo

A study in Norway has found that prescription of Z-drugs is associated with around a doubled increased relative risk of road traffic accidents compared to those who don’t take them. This is similar to that seen with of nitrazepam.

Prescribers should follow NICE guidance for the use of Z-drugs in the management of insomnia. After non-drug therapies have been explored, hypnotics should be used in the lowest dose possible for no more than 4 weeks in the case of benzodiazepines and between 2 and 4 weeks with Z-drugs. Prescribers should be aware that clinical evidence to differentiate Z-drugs from benzodiazepines is weak, and in practice there appears to be little to choose between them in terms of efficacy, side-effects (including next-day effects) and dependency. Patients should be advised that there is around a doubling of the relative risk of road traffic accidents associated with taking a Z-drug, and they need to be very cautious about driving the day after taking any hypnotic.

What is the background to this?
A previous MeReC Rapid Review blog highlighted the widespread inappropriate prescribing of hypnotic drugs contrary to NICE guidance, and questioned the perceived view that Z-drugs have any advantage in practice compared with benzodiazepines with regards to efficacy, adverse effects or in causing dependency. One of the commonly perceived advantages of Z-drugs over benzodiazepines is that they have less of a “hangover” effect the next day and may reduce the risk of accidents and falls. This recent study in Norway has examined the risk of road traffic accidents following prescription of Z-drugs, and suggests that this perception is false.

What does this study claim?
The study found that Norwegian drivers, aged 18 to 69 years, were at around double the relative risk of being involved in a traffic accident if they had been prescribed a hypnotic in the previous week compared with those who hadn’t. Z-drug use (as indicated by dispensing of a prescription of zopiclone or zolpidem in the previous 7 days) had a standardised incidence ratio [SIR] of 2.3; 95% confidence interval [CI] 2.0 to 2.7) –  similar to that of nitrazepam (SIR 2.7; 95%CI 1.8 to 3.9). Flunitrazepam use was associated with an SIR of 4.0 (95%CI 2.4 to 6.4). Absolute rates of traffic accidents associated with hypnotic prescription were around 5, 5, 6 and 9 accidents respectively per exposed 1000 person years and around 2 per 1000 in the group not exposed to hypnotics.

How does this relate to other studies?
One of the main problems associated with the use of hypnotics is the risk of “hangover” effects, e.g. residual daytime sleepiness and impairment of psychomotor and cognitive functioning the day after bedtime administration. However, the severity and duration of these effects may differ between hypnotics and is strongly dependent on the dose administered (Vemeeren 2005). Summaries of product characteristics for both benzodiazepines and Z-drugs, include warnings about possible effects which may impair the ability to drive or operate machinery. The warning for zopiclone states that patients should not drive or operate machinery the day after treatment “until it is established that their performance is unimpaired, whereas that for zolpidem recommends a resting period of 7 to 8 hours between taking zolpidem and driving “in order to minimise the risk”.

Two earlier studies, discussed by Bandolier, of the effect of prescribed benzodiazepines on road traffic accidents, suggests that the risk is reduced with drugs that have a shorter half-life, or when lower doses are used. Z-drugs, because of their shorter half-life, are commonly perceived to have less of a “hangover” effect than longer-half life benzodiazepines. This suggests that they may be associated with a lower risk of traffic accidents. However, this has not been established in studies of accident rates in the “real world”. A previous study in Norway found that hypnotic benzodiazepines were the prescribed drugs associated with the highest risk of road traffic accidents (SIR 3.3, 95%CI 2.1 to 4.7 compared with non-use in the first 7 days after dispensing). The study reported here, extends the findings of this earlier study, by estimating the risk associated with use of Z-drugs.

So what?
These new results may help dispel any remaining myths that Z-drugs (zopiclone and zolpidem) cannot influence next-day driving ability and cause road traffic accidents. Although the increased risks of road traffic accidents with Z-drugs were not compared directly with those of nitrazepam, the mean estimates are similar and the 95% confidence intervals [CI] around the mean values suggest that there is no significant difference between them.

There are many limitations to this study. Traffic accidents are only a surrogate marker for adverse side-effects of hypnotics, albeit a very serious one, and the absolute increased risk of road traffic accidents associated with taking a hypnotic is very small. It is also not possible to say with certainty that the accidents were caused by the taking of the drugs; it was not known, for example, if the driver taking the drugs was to blame for the accident. Another limitation is that the study did not take account of the doses of drugs prescribed or the time between the drugs being taken and the accident, or whether the drugs that had been dispensed had actually been taken. Although the study did take account of people who received prescriptions for other psychoactive drugs, it did not adjust for concomitant alcohol intake, which would be expected to increase the risk.

Despite these limitations, the study reinforces the need to warn people about the increased risk of accidents and to be very cautious about driving the day after taking any hypnotic, including Z-drugs, regardless of the dose used or the period between driving and taking the drug.

Study details

Gustavsen I, et al. Road traffic accident risk related to prescriptions of the hypnotics zopiclone, zolpidem, flunitrazepam and nitrazepam. Sleep Med 2008;9:818-22

Design: Retrospective cohort study
Participants: 3.1 million Norwegians aged 18-69 years (data taken from Norwegian population-based registries).
Intervention: Receipt of a prescription of zopiclone, zolpidem, nitrazepam or flunitrazepam. Periods of exposure were 9,894, 1,487, 2,082 and 1,043 person-years, respectively.
Comparison: The incidence of accidents, within just over 7 and 14 days of dispensing of the hypnotics, was compared to the incidence of accidents in unexposed people.
Outcomes: SIR compared to non-exposed people; SIR excluding those taking concomitant psychoactive drugs; SIR following use of zopiclone after a 180-day washout period.
Results: There were 129 accidents reported following zopiclone use, 21 following zolpidem use, 27 following nitrazepam use, and 18 following flunitrazepam use (the total number of exposed person years were about 28,000, 3,500, 4,500 and 2,000, respectively). These equate to absolute rates of about 0.005, 0.006, 0.006 and 0.009 accidents per exposed person years compared with around 0.002 accidents per person year in the control group. Use of Z-drugs (zopiclone or zolpidem), as indicated by dispensing of a prescription in the previous 7 days, was significantly associated with an SIR of 2.3 (95%CI 2.0 to 2.7). SIRs for individual Z-drugs were 2.3 (95%CI 2.0 to 2.8) for zopiclone and 2.2 (95%CI 1.4 to 3.4) for zolpidem. The SIR for nitrazepam was 2.7 (95%CI 1.8 to 3.9) and the SIR for flunitrazepam was 4.0 (95%CI 2.4 to 6.4). Using a 14-day period following dispensing as the indicator for use of the hypnotics, the SIRs were slightly lower; however, the mean values for all drugs were still statistically significantly greater than zero, and their CIs  overlapped. Results were similar when concomitant use of psychoactive drugs was excluded or following exposure to zopiclone after 180 days of washout.
Sponsorship: Authors were from the Norwegian Institute of Public Health and Universities within Canada and Norway. No sponsorship was reported.

Please comment on this blog in the NPC discussion rooms, or using our feedback form