Objectives To investigate the relation between the risk of Alzheimer's disease and exposure to benzodiazepines started at least five years before, considering both the dose-response relation and prodromes (anxiety, depression, insomnia) possibly linked with treatment.

The short-term effects of benzodiazepines on memory are well established and are suspected in the long term. Eleven studies have been published so far concerning benzodiazepine use and the risk of dementia disorders; nine of these studies concluded these drugs have a deleterious effect, one found a protective effect, and one (the most recently published) observed no effect. The positive association found in some studies could be due to a reverse causation bias since the main indications for benzodiazepines (e.g. sleep disorders, anxiety) can also be prodromes of dementia disorders. This bias is less likely for treatments started more than 10 years before the diagnosis. Among others, three mechanisms could underlie the potential influence of benzodiazepines on the development of dementia disorders. First, benzodiazepines can decrease beta-site amyloid precursor protein-cleaving enzyme 1 (BACE-1) and γ-secretase activity and slow down the accumulation of Aβ oligomers in the brain. This potential positive effect has never been confirmed; the same is true for the prevention of excitotoxicity through benzodiazepine anti-glutamatergic action. Second, since astrocytes located in the area of amyloid plaques could have gamma-aminobutyric acid (GABA)-secreting activity, patients with pre-dementia lesions could be at increased risk of presenting with more pronounced deleterious cognitive effects of benzodiazepines. Finally, owing to the neural compensation and cognitive reserve concepts, some subjects could cope with initial lesions by using/developing alternative networks. By lowering the brain activation level, benzodiazepines could limit this capacity. In conclusion, it is essential that animal studies explore the mechanistic hypotheses of this association found by most of the pharmacoepidemiological studies conducted on this topic.