“Prescription sleeping pills … can raise chance of developing Alzheimer's by 50%,” reports the Mail Online.
This headline is based on a study comparing the past use of benzodiazepines, such as diazepam and temazepam, in older people with or without Alzheimer’s disease. It found that the odds of developing Alzheimer’s were higher in people who had taken benzodiazepines for more than six months.
Benzodiazepines are a powerful class of sedative drugs. Their use is usually restricted to treating cases of severe and disabling anxiety and insomnia. They are not recommended for long-term use, because they can cause dependence.
It’s also important to note that this study only looked at people aged 66 and above, therefore it is not clear what the effects are in younger people. Also, it is possible that the symptoms these drugs are being used to treat in these older people, such as anxiety, may in fact be early symptoms of Alzheimer’s. The researchers tried to reduce the likelihood of this in their analyses, but it is still a possibility.
Overall, these findings reinforce existing recommendations that a course of benzodiazepines should usually last no longer than four weeks.
Where did the story come from?
The study was carried out by researchers from the University of Bordeaux, and other research centres in France and Canada. It was funded by the French National Institute of Health and Medical Research (INSERM), the University of Bordeaux, the French Institute of Public Health Research (IRESP), the French Ministry of Health and the Funding Agency for Health Research of Quebec.
The Mail Online makes the drugs sound like they are “commonly used” for anxiety and sleep disorders, when they are used only in severe, disabling cases. It is also not possible to say for sure that the drugs are themselves directly increasing risk, as suggested in the Mail Online headline.
What kind of research was this?
This was a case control study looking at whether longer-term use of benzodiazepines could be linked to increased risk of Alzheimer’s disease.
Benzodiazepines are a group of drugs used mainly to treat anxiety and insomnia, and it is generally recommended that they are used only in the short term – usually no more than four weeks.
The researchers report that other studies have suggested that benzodiazepines could be a risk factor for Alzheimer’s disease, but there is still some debate. In part, this is because anxiety and insomnia in older people may be early signs of Alzheimer’s disease, and these may be the cause of the benzodiazepine use. In addition, studies have not yet been able to show that risk increases with increasing dose or longer exposure to the drugs (called a “dose-response effect”) – something that would be expected if the drugs were truly affecting risk. This latest study aimed to assess whether there was a dose-response effect.
Because the suggestion is that taking benzodiazepines for a long time could cause harm, a randomised controlled trial (seen as the gold standard in evaluating evidence) would be unethical.
As Alzheimer’s takes a long time to develop, following up a population to assess first benzodiazepine use, and then whether anyone develops Alzheimer’s (a cohort study) would be a long and expensive undertaking. A case control study using existing data is a quicker way to determine whether there might be a link.
As with all studies of this type, the difficulty is that it is not possible to determine for certain whether the drugs are causing the increase in risk, or whether other factors could be contributing.
What did the research involve?
The researchers used data from the Quebec health insurance program database, which includes nearly all older people in Quebec. They randomly selected 1,796 older people with Alzheimer’s disease who had at least six years’ worth of data in the system prior to their diagnosis (cases). They randomly selected four controls for each case, matched for gender, age and a similar amount of follow-up data in the database. The researchers then compared the number of cases and controls who had started taking benzodiazepines at least five years earlier, and the doses used.
Participants had to be aged over 66 years old, and be living in the community (that is, not in a care home) between 2000 and 2009. Benzodiazepine use was assessed using the health insurance claims database. The researchers identified all prescription claims for benzodiazepines, and calculated an average dose for each benzodiazepine used in the study. They then used this to calculate how many average daily doses of the benzodiazepine were prescribed for each person. This allowed them to use a standard measure of exposure across the drugs.
Some benzodiazepines act over a long period as they take longer to be broken down and eliminated from the body, while some act over a shorter period. The researchers also noted whether people took long- or short-acting benzodiazepine, those who took both were classified as having taken the longer acting form.
People starting benzodiazepines within five years of their Alzheimer’s diagnosis (or equivalent date for the controls) were excluded, as these cases are more likely to potentially be cases where the symptoms being treated are early signs of Alzheimer’s.
In their analyses, the researchers took into account whether people had conditions which could potentially affect the results, including:
- high blood pressure
- heart attack
- high cholesterol
What were the basic results?
Almost half of the cases (49.8%) and 40% of the controls had been prescribed benzodiazepines. The proportion of cases and controls taking less than six months’ worth benzodiazepines was similar (16.9% of cases and 18.2% of controls). However, taking more than six months’ worth of benzodiazepines was more common in the controls (32.9% of cases and 21.8% of controls).
After taking into account the potential confounders, the researchers found that having used a benzodiazepine was associated with an increased risk of Alzheimer’s disease, even after taking into account potential confounders (odds ratio (OR) 1.43, 95% confidence interval (CI) 1.28 to 1.60).
There was evidence that risk increased the longer the drug was taken, indicated by the number of days’ worth of benzodiazepines a person was prescribed:
- having less than about three months’ (up to 90 days) worth of benzodiazepines was not associated with an increase in risk
- having three to six months’ worth of benzodiazepines was associated with a 32% increase in the odds of Alzheimer’s disease before adjusting for anxiety, depression and insomnia (OR 1.32, 95% CI 1.01 to 1.74) but this association was no longer statistically significant after adjusting for these factors (OR 1.28, 95% CI 0.97 to 1.69)
- having more than six months’ worth of benzodiazepines was associated with a 74% increase in the odds of Alzheimer’s disease, even after adjusting for anxiety, depression or insomnia (OR 1.74, 95% CI 1.53 to 1.98)
- the increase in risk was also greater for long-acting benzodiazepines (OR 1.59, 95% 1.36 to 1.85) than for short-acting benzodiazepines (OR 1.37, 95% CI 1.21 to 1.55).
How did the researchers interpret the results?
The researchers concluded that, “benzodiazepine use is associated with an increased risk of Alzheimer’s disease”. The fact that a stronger association was found with longer periods of taking the drugs supports the possibility that the drugs may be contributing to risk, even if the drugs may also be an early marker of the onset of Alzheimer’s disease.
This case control study has suggested that long-term use of benzodiazepines (over six months) may be linked with an increased risk of Alzheimer’s disease in older people. These findings are reported to be similar to other previous studies, but add weight to these by showing that risk increases with increasing length of exposure to the drugs, and with those benzodiazepines that remain in the body for longer.
The strengths of this study include that it could establish when people started taking benzodiazepines and when they had their diagnosis using medical insurance records, rather than having to ask people to recall what drugs they have taken. The database used is also reported to cover 98% of the older people in Quebec, so results should be representative of the population, and controls should be well matched to the cases.
The study also tried to reduce the possibility that the benzodiazepines could be being used to treat symptoms of the early phase of dementia, by only assessing use of these drugs that started at least six years before Alzheimer’s was diagnosed. However this may not remove the possibility entirely, as some cases of Alzheimer’s take years to progress, which the authors acknowledge.
All studies have limitations. As with all analyses of medical records and prescription data, there is the possibility that some data is missing or not recorded, that there may be a delay in recording diagnoses after the onset of the disease, or that people may not take all of the drugs they are prescribed. The authors considered all of the issues and carried out analyses where possible to assess their likelihood, but concluded that they seemed unlikely to be having a large effect.
There were some factors which could affect Alzheimer’s risk, which were not taken into account because the data was not available (for example, smoking and alcohol consumption habits, socioeconomic status, education or genetic risk).
It is already not recommended that benzodiazepines are used for long periods, as people can become dependent on them. This study adds another potential reason why prescribing these drugs for long periods may not be appropriate.
If you are experiencing problems with insomnia or anxiety (or both), doctors are likely to start with non-drug treatments as these tend to be more effective in the long term.
Analysis by Bazian
Edited by NHS Website
Links to the headlines
BBC News, 10 September 2014
The Daily Telegraph, 10 September 2014
Daily Mail, 10 September 2014
Daily Mirror, 10 September 2014
Daily Express, 10 September 2014
Links to the science
BMJ. Published online September 9 2014