DOI: https://doi.org/10.36347/sjmcr.2025.v13i09.013

Pages: 2013-2016

Objective: To review recent epidemiological and mechanistic data on the association between benzodiazepine (BZD) use and dementia risk, discuss methodological biases, and propose adapted clinical recommendations, including for the Moroccan context. Methods: A targeted narrative review of literature (published up to July 2024), including cohort studies, case-control studies, recent meta-analyses, neuroimaging studies, and animal mechanistic studies. Databases consulted: PubMed/PMC, Embase, BMJ/BMC, J Clin Neurol, Int Psychogeriatr. Priority was given to meta-analyses and large prospective cohorts. Results: Studies remain discordant. Some prospective and case-control studies report an increased risk (e.g., Billioti de Gage et al., 2012; meta-analyses 2018–2019), while other large cohorts with better temporal controls find no clear overall effect (e.g., Gray et al., 2016; vom Hofe et al., 2024). Recent mechanistic work in mice shows that prolonged diazepam treatment can promote synaptic loss via the TSPO receptor and impair plasticity, providing biological plausibility for persistent cognitive effects. The role of protopathic bias (prescription for prodromal dementia symptoms: anxiety, insomnia) and confounding by indication largely explain the heterogeneity of results. In some analyses, cumulative duration, drug half-life, and dose appear to modulate the observed signal. Moroccan data show significant consumption and reported cases of intoxication, suggesting an important local public health issue. Conclusion: There is currently no definitive proof of an unambiguous causal link between BZDs and dementia for all populations. However, the body of evidence (epidemiology + experimental mechanisms + imaging) argues for caution, especially in the elderly. Practical recommendations should aim to: limit prescriptions, favor non-pharmacological alternatives, and implement structured deprescribing when indicated.