Association of antidepressants with recurrent, injurious and unexplained falls is not explained by reduced gait speed.

Objective: To examine if antidepressants at baseline are associated with falls and syncope over 4 years follow-up and if any observed associations are explained by baseline gait speed.

Design: Longitudinal study (three waves).

Setting: The Irish Longitudinal Study on Ageing (TILDA), a nationally representative cohort study.

Participants: Two thousand ninety-three community-dwelling adults aged ≥60 years.

Measurements: Antidepressants (ATC code "N06A") were identified. Recurrent falls (≥2 falls), injurious falls (requiring medical attention), unexplained falls, and syncope were reported at either Wave 2 or 3. Usual gait speed was the mean of two walks on a 4.88 m GAITRite walkway. Poisson regression analysis was used to examine associations between baseline antidepressant use and future falls adjusting for sociodemographics, physical, cognitive and mental health, and finally, gait speed.

Results: Compared to non-antidepressant users, those on antidepressants at baseline were more likely to report all types of falls (24.8-40.7% versus 9.8-18%) at follow-up. Antidepressants at baseline were independently associated with injurious falls (incidence risk ratio: 1.58, 95% confidence interval: 1.21, 2.06, z = 3.38, p = 0.001, df = 32) and unexplained falls (incidence risk ratio: 1.49, 95% confidence interval: 1.04, 2.15, z = 2.17, p = 0.030, df = 32) independent of all covariates including gait speed.

Conclusion: There was little evidence to support the hypothesis that gait would (partly) explain any observed associations between baseline use of antidepressants and future falls. The underlying mechanisms of the observed relationships may be related to depression, vascular pathology, or direct effects of antidepressants. Clinicians should identify the best treatment option for an individual based on existing risk factors for outcomes such as falls.