Benzodiazepine use in relation to long‐term dementia risk and imaging markers of neurodegeneration: a population‐based cohort study.

Abstract

AbstractBackgroundBenzodiazepine use is common, particularly in older adults. Benzodiazepines have well‐established acute adverse effects on cognition. However, long‐term effects on dementia risk remain uncertain, and on subclinical imaging markers of neurodegeneration largely undetermined.MethodWe included 5443 cognitively healthy participants (MMSE≥26) from the population‐based Rotterdam Study (57.4% women, mean age 70.6 years). Benzodiazepine use from 1991 until baseline (2005‐2008) was derived from ATC‐coded pharmacy records, from which we determined drug type (anxiolytics vs. sedative‐hypnotics vs. both) and cumulative dose. We determined the association of benzodiazepine use with dementia risk until 2020 using Cox regression, and with change in neuroimaging markers during 5‐yearly repeated brain MRI using linear mixed models. Models were adjusted for demographics, lifestyle factors and comorbidity, including presence of anxiety, depression and sleeping problems.ResultDuring a mean follow‐up of 11.2 years, 726 participants (13.3%) developed dementia. Benzodiazepine use occurred in 1675 (30.8%) participants, of whom 845 (50.4%) had used anxiolytics, 390 (23.3%) had used sedative‐hypnotics, and 440 (26.3%) had used both. Current use of any type of benzodiazepine at baseline, but not former use, was associated with an increased dementia risk during follow‐up (HR[95%CI] for current users: 1.56[1.17‐2.09]; and past users: 1.07[0.87‐1.31]). Excess risk in current users persisted after excluding a 2‐year lag‐time of benzodiazepine use prior to baseline. Dementia risk was dependent on cumulative dose (>median: HR[95%CI] 1.31[1.05‐1.63] vs. < = median: HR[95%CI] 1.01[0.78‐1.31]). Use was not associated with volume of grey matter, white matter, hippocampus, amygdala or thalamus at baseline. During follow up, high cumulative dose was associated with accelerated decrease in hippocampal volume (p = 0.021). Regarding drug type, dementia risk was increased with anxiolytics (overall HR[95%CI]: 1.37[1.10‐1.71]), which was paralleled by accelerated atrophy of grey matter (p = 0.036), albeit no dose‐response relationship was observed. Sedative‐hypnotics were not associated with dementia risk or brain atrophy. Combined use of anxiolytics and sedative‐hypnotics at baseline was associated with increased dementia risk (HR[95%CI]: 1.70[1.05‐2.76]), but this attenuated after excluding use during a 2‐year lag time before baseline (HR[95%CI]: 1.08[0.71‐1.64]), and was not associated with brain atrophy.ConclusionChronic use of benzodiazepines in a population of cognitively healthy older adults is associated with increased dementia risk.