Association between antidepressant drugs and falls in older adults: A mediation analysis in the World Health Organization's pharmacovigilance database.

Objectives: The objective is to investigate the association between antidepressant drugs intake and falls reporting, as well as the potential mediators in-between, in older adults.

Methods: In VigiBase®, the World Health Organization's pharmacovigilance database, we performed a disproportionality analysis to probe the putative associations between each antidepressant drugs class (non-selective monoamine reuptake inhibitors [NSMRIs], selective serotonin reuptake inhibitors [SSRIs], serotonin-norepinephrine reuptake inhibitors [SNRIs], alpha-2-adrenergic receptor antagonists, and "other antidepressants") and reports of falls in people aged 65 and over (NCT05628467). The reporting odds ratios and their 95% confidence interval were derived from logistic regression models with adjustment for confounders. We studied the falls-inducing mechanisms (delirium, hyponatremia, hypotension) by using causal mediation analyses and by using a disproportionality analysis for the co-occurrence of falls and these events.

Results: Our main analysis included 86,200 cases of falls reporting in older adults (of which 57% were 75 and over). A significant association was found between falls and every antidepressant drugs class, except for NSMRIs. According to causal mediation analysis, a direct effect on the falls reports was shown for alpha-2-adrenergic receptor antagonists and for "other antidepressants". According to the co-reports analyses, all antidepressant drugs classes except SNRIs were associated with the co-event fall-delirium; SSRIs, alpha-2-adrenergic receptor antagonists, and "other antidepressants" with fall-hypotension; all antidepressant drugs classes except NSMRIs with fall-hyponatremia.

Conclusions: In multivariate disproportionality analyses, all antidepressant drugs classes were associated with signals of disproportionate reporting of falls in older adults, except for NSMRIs. In mediation analyses, a direct effect on the falls reports was only found for alpha-2-adrenergic receptor antagonists. Single-mediators based models seem insufficient to explain the diversity of clinical settings resulting in falls. These findings underline the necessity of a comprehensive analysis of all clinical and pharmacological features in older falling adults treated with antidepressant drugs.