Widely used sleeping pill 'increases dementia' risk

Sleeping pills taken by millions are linked to dementia, according to The Daily Telegraph. Given that an estimated 10 million to 11 million prescriptions for benzodiazepines are reported to be issued each year in the UK, could we be at risk of “sleepwalking” into a public health disaster?

The reports stem from the results of a French study that followed just over a thousand elderly adults (average age of 78) for 15 years. The participants were initially free from dementia but those who started taking benzodiazepines after the first three years of the study were 60% more likely to develop dementia than those who did not use the drugs.

The main difficulty in this study is in establishing the exact cause of dementia and what role benzodiazepines play. Benzodiazepines are a commonly used group of sedatives prescribed for sleeping problems and anxiety.

Although the researchers have taken into account several potential confounding factors that may be involved in the relationship, it is difficult to exclude the possibility that the apparent dementia risk may not be caused directly by the drugs themselves. Instead, it could be related to whatever underlying conditions or biological processes in the brain are causing the person to require sleeping tablets in the first place.

Furthermore, disturbed sleep can be an initial sign of dementia, so the use of sleeping tablets may be triggered by early dementia and not vice versa.

The use of benzodiazepines appears to be relatively common in France so the results may not apply to the UK.

Where did the story come from?

The study was carried out by researchers from Université Bordeaux Segalen and other institutions in France, and Brigham and Women’s Hospital, Boston, USA. The research received financial support from several sources, including the Institut National de la Santé et de la Recherche Médicale (INSERM) and Université Bordeaux Segalen.

The study was published in the peer-reviewed British Medical Journal.

The study was reported accurately by the media and many newspapers should be praised for highlighting the other health risks associated with long-term use of benzodiazepines.

What kind of research was this?

This was a cohort study which aimed to look at the association between benzodiazepine use and risk of new onset of dementia in a group of elderly people followed for six years.

A cohort study is a good way of looking at whether a particular exposure is associated with the risk of developing a particular disease outcome over time.

Some potential limitations of this study are that, despite the researchers’ attempts to take into account potential confounders, it is difficult to ensure that all of these have been taken into account and to establish that early dementia was not the cause of insomnia.

The researchers tried to offset this by making sure they selected recruits who did not start taking sleeping tablets until at least until the third year.

By using this method, people who may have had initial signs of dementia at the start of the trial, such as insomnia and anxiety, were excluded from the study.

This helped to minimise the potential for what is known as reverse causation from distorting the results of the trial (in other words, people were actually taking sleeping pills because they were developing the first signs of dementia).

Despite the best efforts of the researchers, as little is known about the early stages or pre-symptoms ("prodrome") of dementia, it is unclear whether the three-year gap was long enough to offset this potential completely.

What did the research involve?

The research included participants enrolled into a study designed to look at brain ageing in both normal and diseased states. Adults aged over 65 years old were randomly sampled from the French community between 1987 and 1989. 

At the start of the study and at follow-up interviews every two to three years, trained researchers collected information on:

  • personal characteristics
  • sociodemographics
  • lifestyle
  • medical conditions
  • medication use
  • functional abilities
  • depressive symptoms
  • brain function

The presence of dementia was assessed using valid diagnostic criteria which were carried out by psychologists trained to diagnose dementia.

The researchers looked at 1,063 men and women (average age 78) who were free from dementia at the start of the study. People using one of 23 types of benzodiazepines were defined as those who started taking the drugs for the first time somewhere between the first three-year and five-year follow-ups. This was when they were still confirmed to be free of dementia. The researchers collected data on the specific benzodiazepine drugs used.

The researchers then looked at the risk of dementia in follow-up after first reported benzodiazepine use, compared with those with no reported benzodiazepine use.

In their analyses, the researchers took into account potential confounders that could also affect the participants' risks of dementia, such as:

  • age
  • gender
  • school education
  • marital status
  • wine consumption
  • depression
  • use of blood-pressure medication
  • use of diabetic medications
  • use of cholesterol-lowering medications 
  • use of anti-blood-clotting drugs (such as warfarin)

What were the basic results?

The researchers found that during the first six years of follow-up there were 253 new cases of dementia. Dementia affected 23% of non-users of benzodiazepines compared with 32% of those who started using benzodiazepine between three to five years of follow-up.

In the adjusted analyses the researchers found that new use of benzodiazepines (reported at five-year follow-up) was associated with a 60% increased risk of dementia compared with non-use (hazard ratio 1.60, 95% confidence interval 1.08 to 2.38).

The researchers also found broadly similar risk associations when looking at the first report of benzodiazepines at later follow-up points (participants who reported first benzodiazepine use at 8, 10, 13 or 15 years when they were still free of dementia). Looking at all of these risk associations they found that people who used benzodiazepines had, roughly, an almost 50% higher risk of dementia compared with non-users.

How did the researchers interpret the results?

The researchers reported that in their cohort study of elderly adults, new use of benzodiazepines was associated with increased risk of dementia. They said: “Considering the extent to which benzodiazepines are prescribed and the number of potential adverse effects of this drug class in the general population, indiscriminate widespread use should be cautioned against.”


This research demonstrates an association between new benzodiazepine use among elderly adults and the risk of developing dementia. The study has several strengths, including its long follow-up period and valid assessment of dementia development. Some limitations are that:

  • Despite the relatively large sample size, the actual number of new users of benzodiazepines after the third year of study was quite small (95), therefore limiting the ability to detect reliable differences in dementia risk among users and non-users (the smaller a sample size the more likely it is that any detected effects are the results of chance).
  • The researchers did attempt to adjust for confounders, but it is difficult to exclude the possibility that the apparent dementia risk may not be caused directly by the drugs themselves, but be related to whatever underlying conditions or biological processes in the brain are causing the person to require sleeping tablets. For example, although the study looked at dementia, it did not look at other mental health problems such as anxiety.
  • People who were excluded from the study because they were taking benzodiazepine during the three-year “run-in” period tended to be less well-educated and more likely to live alone, so removing these types of people from the study may mean that the results are not representative of the population at large or those who may have higher risk of developing dementia.
  • It is possible that although a three-year window before drug use was chosen to ensure that  early symptoms of dementia were not appearing, this may not have been long enough. This could have led to reverse causation – a common problem with these studies – in which it is the early symptoms of dementia itself that lead to use of sleeping tablets.

Nevertheless, this is a well-conducted study that adds to the growing body of opinion that benzodiazepines should only be a “treatment of last resort” for severe acute insomnia or anxiety and should be taken for no longer than two-to-four weeks at a time.

Analysis by NHS Choices. Follow Behind the Headlines on twitter.

NHS Attribution