Fractures, falls, and stroke are all risks for patients with dementia who take higher dose “Z-drugs,” including zopiclone, zaleplon, and zolpidem for sleep disturbances, results of a large observational study show.
The risks with higher doses of Z-drugs were similar or greater in magnitude than those for higher dose benzodiazepines.
Z-drugs “should be avoided in people with dementia. Guidance indicates they should be avoided but they are still prescribed despite this,” Chris Fox, MD, from University of East Anglia, Norwich, England, told Medscape Medical News.
The study was published online November 24 in BMC Medicine.
Safer Than Benzos?
Z-drugs, a class of non-benzodiazepine gamma-aminobutyric acid agonists with a shorter half-life, were initially thought to be safer than benzodiazepines, but growing evidence has revealed significant adverse effects in this patient population.
Fox and colleagues examined the association between first Z-drug prescription and subsequent risk of falls, fractures, death, infection, ischemic stroke, and venous thromboembolism (VTE) in 27,090 people (mean age, 83; 62% women) diagnosed with dementia between January 2000 and March 2016.
To reduce confounding, they compared 3532 Z-drug users with 1833 nonsedative users with sleep disturbance; 10,214 nonsedative users with a recent GP visit matched on age, sex, and antipsychotic use; and 5172 new benzodiazepine users.
Of 3532 dementia patients prescribed Z-drugs, 584 (17%) were started on higher doses (≥ 7.5 mg zopiclone or equivalent).
In the fully adjusted model, patients started on higher dose Z-drugs had a significantly increased risk for fractures (hazard ratio [HR], 1.67; 95% CI, 1.13 to 2.46), hip fractures (HR, 1.96; 95% CI, 1.16 to 3.31), falls (HR, 1.33; 95% CI, 1.06 to 1.66) and ischemic stroke (HR, 1.88; 95% CI, 1.14 to 3.10), vs nonsedative users with sleep disturbance.
Similar associations were found when higher dose Z-drug users were compared with nonsedative users with a recent GP visit.
“Minimal or inconsistent” excess risk for adverse outcomes were observed when Z-drugs were prescribed at lower doses (≤ 3.75 mg zopiclone or equivalent) daily, the investigators report.
There were no consistent or clinically significant increased risks of mortality, infection, or VTE with Z-drug use.
Patients with dementia who were prescribed higher dose Z-drugs were also more likely to be admitted to the hospital, visit their GP, and receive prescriptions for antipsychotics, antidepressants, and antibiotics.
A Necessary Reminder
There were no differences in adverse events when Z-drug users were compared with the 5172 benzodiazepine users, except for lower mortality rates with Z-drugs (HR, 0.73; 95% CI, 0.64 to 0.83).
“Our findings serve an important caution regarding the harms of sleeping tablets in people with dementia,” Clive Ballard, MD, University of Exeter Medical School, England, who collaborated on the study, said in a statement.
“This research is a very timely and unfortunately necessary reminder that sedative medications are not a helpful way to manage social isolation during COVID-19,” he added.
“We need to improve sleep management treatments in dementia,” Fox told Medscape Medical News.
Options included Improving sleep hygiene and reviewing medicines that may be causing the sleep problem.
“If one has to prescribe medication this should be very time-limited. Patients on Z-drugs should not stop but should consult their health provider to tailor off them,” said Fox.
Reached for comment, Marc Gordon, MD, chief of neurology, Zucker Hillside Hospital, Glen Oaks, New York, said, “This was a study that was worth doing. They looked at [this issue] more systematically, and I think they did a good job in trying to control for, or at least address, the potential confounding effects.”
The study adds to the literature and provides validation of what we have suspected — that it’s probably not a good idea to give sedative hypnotics to elderly people with dementia,” said Gordon.
However, he cautioned, “with a study like this, you can determine there is an association, but you can’t absolutely determine causality.”
The study was funded and commissioned by the National Institute of Health Research (NIHR). Fox and Gordon have disclosed no relevant financial relationships.
BMC Med. Published online November 24, 2020. Full text