Project Title:

Unintended effects of medications for sleep disturbance in people with dementia

Type of Research:

Secondary data analysis

Background & Scientific Rationale:

Poor sleep is common for people with dementia (PwD). It can lead to poor quality of life for them and their carers. Z-drugs (zolpidem, zopiclone and zaleplon) are sleep-drugs often given to PwD to improve their sleep. Some studies suggest that Z-drugs may be harmful, but no studies have looked into the effects of Z-drugs for PwD. In addition, Z-drugs are believed to be a safer alternative to benzodiazepines, but this has not been examined in PwD.

The Z-drugs for sleep disturbance in people with dementia study (“ZED”) was funded by NIHR HTA from April 2016-July 2018 to perform secondary data analysis on this topic. I was funded for part of this study to examine adverse events of Z-drug use in PwD in primary care data. We used primary care data in a fairly small specific population (N=4,603) and found some interesting border-line significant associations with Z-drug use and increased risk of fractures. However, we were limited by identifying patients with sleep disturbance recorded in their primary care record, which turned out to be poorly recorded. As such the study was under-powered to examine our primary outcome of fractures. We were also unable to compare to benzodiazepine use.

This funding application is to support work extending the ZED study to perform a larger primary care study to specifically compare the risk associated with Z-drugs compared to benzodiazepines in all people with dementia (without the requirement of a record of sleep disturbance in their primary care record).

Research aims:

To estimate the effects of first prescription of (i) Z-drugs and (ii) benzodiazepines in PwD compared to no treatment. Patient outcomes include: (a) Incidence of falls and factures, (b) Mortality, (c) Incidence of infection, (d) Incidence of ischaemic stroke and venous thromboembolism, (e) Additional medication (sedatives, antipsychotics, antidepressants and antibiotics), and (f) Healthcare utilisation (GP visits and hospital admissions.


Design: Observational cohort studies using Clinical Practice Research Datalink (CPRD)
Setting: 651 primary care practices in the UK.
Participants: UK primary care patients aged over 55 years diagnosed with dementia from January 2000 to March 2016, followed for up to 2-years.
Exposures: First prescription of a Z-drug or benzodiazepine in PwD (ie no prescriptions in past 12 months). Each exposed PwD will be matched to 2 PwD not prescribed Z-drugs /benzodiazepines on age, sex, country, antipsychotic use, and dementia diagnosis year.
Main outcomes: Fall, fracture, infection, stroke, venous thromboembolism, and mortality

Expected Output of Research / Impact and added value:

These findings are likely to influence guidance for sleep management in dementia and future research in this area. Doctors will better understand the potential harms associated with medications they are prescribing. Patients and carers of people with dementia will be able to weight up the benefits and harms of specific treatments. Ultimately, this work will lead to improved care and outcomes for PwD, which will improve the quality of life of patients and carers, reduce hospital admissions, mortality and institutional placement.


For further information on this project, please contact Kathryn Richardson

Share this article
Tweet about this on Twitter