Mediation analysis of antidepressant use, depressive symptoms, and recurrent falls in community-dwelling older fallers: an exploratory study

Objectives: This study aimed to explore whether there might exist an interaction between using antidepressants
and the influence of depressive symptoms on the recurrence of falls.
Design: Cross-sectional study using secondary data from a randomized clinical trial.
Setting and participants: Community-dwelling older adults (n = 609, aged 73.4 ± 7.4 years) who had experienced
at least one fall in the past 12 months.
Methods: Depressive symptoms were measured using the Geriatric Depression Scale, and information about

antidepressant usage was collected. Mediation models were built to decompose the effects of depressive symp-
toms on fall risk into direct effects and indirect effects mediated by antidepressant use.

Results: Depressive symptoms were reported by 29.1 % of the participants, and 27.4 % were using antidepres-
sants. Those with depressive symptoms had 1.86 times the likelihood of being recurrent fallers (ORTE: 1.861, 95

% CI: 1.197, 2.895), and there was no significant interaction between depressive symptoms and antidepressant
use on recurrent falls (P interaction = 0.989). Antidepressant use might be a significant mediator in the relationship
between depressive symptoms and recurrent falls (ORNIE: 1.140, 95 % CI: 1.007, 1.291), accounting for 21.1 % of
the total effect.
Conclusions/implications: Antidepressants probably do not add a significant risk of recurrent falls beyond what is
already contributed by the presence of depressive symptoms. A longitudinal study could clarify whether it might
be safe to use antidepressants to treat older people with depressive symptoms without increasing the risk of falls
the disease leads by itself.