Review of: Osler M and Jorgensen MB, Am J Psych 2020;177(6):497–505

Study TYPE: Epidemiologic case-control

Few psychotropics stir controversy like the benzodiazepines. While they work well for anxiety and insomnia, their risks of abuse and dependence have always nagged at us. More recently, some research has suggested that long-term use increases the risk of dementia. Those studies did not control for subjects’ underlying diagnoses, and this new study partially overcomes that problem by focusing on a homogenous diagnostic group.

This cohort and nested case-control study drew its data from Danish hospital and pharmacy registries. The cohort of roughly 235,000 subjects was comprised of patients who presented for their first hospitalization for a mood disorder. These patients were assessed for dementia at that first visit and followed for 3–11 years. The primary outcome was the difference in the rate of dementia between those who were started on a benzodiazepine or z-hypnotic and those who were not.

Dementia can present with symptoms of depression, so it’s possible that some of the patients who converted to dementia were in the early phase of cognitive decline at the start of the study. To control for that, patients who converted to dementia in the first two years were analyzed separately.

The researchers adjusted for a number of variables including gender, age, marital status, education, depression subtype, year of diagnosis, psychotropic medication use, and comorbidity. The thoroughness of their adjustments and generalizability are what make this study stand apart from prior, similar case-control and prospective studies. The data collected from the registries allowed the researchers to see all diagnoses, prescribed medications, amounts of medications filled, hospitalizations, timing of diagnoses, and even data prior to study entry.

In their analysis, the authors found that out of the study cohort, 4% were diagnosed with dementia. Unexpectedly, there was a decreased risk of dementia in the first 2 years after study entry if a benzodiazepine or z-drug was prescribed (hazard ratio 0.70; range 0.66–0.74). For years 2 through 20 after study entry, there was no association between dementia and use of benzodiazepines or z-drugs, even when stratifying based on number of prescriptions, duration of use, combined use, and half-life.

TCPR’s Take
While it does not completely close the door on the controversy, this study significantly weakens the links between benzos, z-hypnotics, and dementia. However, these medications do have other cognitive side effects separate from any dementia risk, and their use still warrants caution in the elderly due to their risks of falls, traffic accidents, and respiratory suppression. Low doses of short-acting benzodiazepines like lorazepam and oxazepam minimize those risks.