Benzodiazepines linked to dementia risk: cause for concern or lacking causality?

Benzodiazepines are used widely in older adults and with known side effects. Might Alzheimers be another risk to consider?

This is one of those studies that the media thoroughly enjoy: taking benzodiazepines can “raise chance of developing Alzheimer’s by 50%”.

The Daily Mail leapt upon it as did BBC News. These stories are a bit scary as association can all too easily be mistaken as causation. We may yet see the news presented more calmly and critically in the NHS’ remarkable ‘Behind the Headlines’, but in the meantime, read on.

The study, authored by Billioti de Gage et al, was published in the BMJ and looks at data collected within Quebec, Canada. The researchers wanted to explore the relationship between Alzheimer’s and the use of benzodiazepines in the six years prior to diagnosis.


They used a case control design. This involved comparing a group of people with Alzheimer’s against another group matched for age and gender who didn’t have Alzheimer’s but who’s drug records went back a similar amount of time. By dredging through prescription records (in Quebec this is mostly on an insurance database called RAMQ) they identified 1,796 people with Alzheimer’s and 7,184 controls. By using controls they were trying to single out benzodiazepines as the independent variable. It’s worth noting here that a lot of relevant information isn’t included on RAMQ, like lifestyle and socio-economic status.

By looking through the database they could measure whether those with Alzheimer’s had ever used benzodiazepines and, if so, how much cumulative exposure they had racked up. This was calculated from the strength of daily doses and the length of time they were used.

They also measured to see if there was any difference between those taking the shorter-acting benzodiazepines than the ones that stay in the bloodstream for longer, and doing things to the brain for longer. For their statistical analysis they used a multivariate conditional logistic regression.

Researchers looked at the drug records of a group of people with Alzheimers compared to a similar group without the disease

Researchers looked at the drug records of a group of people with Alzheimer’s compared to a similar group without the disease.


What the researchers found could be a bit spine tingling but remember this is all correlation, not cause.

  • Using benzodiazepines at all was associated with an increased risk of Alzheimer’s of 43-51% (or an adjusted odds ratio (OR) of 1.51 (95% confidence interval (CI) 1.36 to 1.69)
  • A higher cumulative exposure was associated with more risk:
    • 32% (OR 1.32; CI 1.01 to 1.74) for cumulative daily doses of 91-180
    • 84% (OR 1.84; CI 1.62 to 2.08) for more than 180 cumulative daily doses
  • The benzodiazepines that stay in the body longer are also associated with a slightly higher odds ratio (OR 1.70; CI 1.46 to 1.98) compared to those that are shorter-acting (1.43; 1.27 to 1.61). This finding is actually very important because it suggests that there might be some toxic effect to the benzodiazepines (if something’s harmful, it makes sense that you’d see a difference between a longer and shorter acting version). However, I did wonder about the large confidence intervals on this one, undermining the reliability of the finding


There are some interesting associations here, but what does it all mean? The authors conclude that:

The stronger association observed for long-term exposures reinforces the suspicion of a possible direct association, even if benzodiazepine use might also be an early indicator of a condition associated with an increased risk of dementia.

The authors are leaning towards an explanation that features benzodiazepines as having a causal role in dementia, akin to if they were causing some kind of permanent damage to the brain. They place emphasis on the finding that those who used more benzodiazepines were more at risk than those who used them to a lesser extent.

What do other experts have to say? Prof Guy Goodwin, President of the ECNP said:

Non-specific symptoms arise in the 14 years before an Alzheimer diagnosis, so a 5 year study, as in this paper, may not be long enough to exclude what we call reverse causality; in other words symptoms in the early phases of Alzheimer’s disease may increase the probability of being prescribed a benzodiazepine.

Perhaps then, the study is just showing that those exposed to more benzodiazepines were people who had the early signs of Alzheimer’s. On the other hand, might we want to be more careful about prescribing benzodiazepines to older people anyway?

Dr Liz Coulthard, Consultant Senior Lecturer in Dementia Neurology at the University of Bristol said:

This work provides yet another reason to avoid prescription of benzodiazepines for anything other than very short term relief of insomnia or anxiety. In addition to short term cognitive impairment, falls and car accidents already known to be associated with benzodiazepine use.

The study raises questions about the safety of benzodiazepines, but is their really cause for concern?

The study raises questions about the safety of benzodiazepines, but is there really cause for concern?


This research certainly stimulates debate, and more expert opinions can be read on the highly recommended Science Media Centre website.

At best, this is a decent study that provides credible reasons for further investigating whether benzodiazepines might be causing some damage. It also raises questions about whether benzodiazepines are being used excessively in older people and posing other risks like falls or short term cognitive impairment. It’s important to be aware that the cascade of Alzheimer’s disease may start over a decade before a diagnosis is made whereas this study looks at only the five years prior to diagnosis. Is benzodiazepine use then simply indicative of pre-diagnosis symptoms?

There’s no such thing as an Ivory tower and we have to think about the possible consequences of how the findings are interpreted. I work with people experiencing intolerable symptoms of anxiety for whom benzodiazepines, and other drugs, are a lifeline. These are also people who are likely to Google their drugs and every side effect known to man – they suffer from anxiety!

I think this study shows something interesting: that those with Alzheimer’s were more likely to use more benzodiazepines. I also think that Alzheimer’s is a grave and massively important issue and this topic merits further research. However, it also carries responsibility to draw a balanced and cautious conclusion with emphasis that this is a correlation, not a causation.

Short-term benzodiazepine use can play an important role in managing anxiety and insomnia, but we need to be aware of the longer-term risks.

Short-term benzodiazepine use can play an important role in managing anxiety and insomnia, but we need to be aware of the longer-term risks.


Billioti de Gage et al. Benzodiazepine use and risk of Alzheimer’s disease: case-control study. BMJ. 2014. 349; g5205.

Share on Facebook Tweet this on Twitter Share on LinkedIn Share on Google+