The effect of continuous use of antidepressants, with or without anticholinergic properties, for at least 4 months versus no treatment on the incidence of dementia: a target trial emulation. Forns J, Aguado J, Rivero-Ferrer E, Plana E, Dickerman BA, Garcia de Albeniz X.
Researcher Joan Forns discusses research on the effect of continuous use of antidepressants, with or without anticholinergic properties, for at least 4 months versus no treatment on the incidence of dementia: a target trial emulation.
Our study evaluated whether the use of antidepressants with anticholinergic properties were associated with dementia, as prior studies showed. We estimated the per-protocol effect of adhering to 3 different treatment strategies on the 12-year risk of dementia. The first two strategies involved the initiation of antidepressants either with or without anticholinergic properties for at least 4 months. The third strategy involved no initiation of antidepressants.
Our target trial emulation design differed from previous studies in several ways.
- Active comparator
- Inverse Probability Weighting
- Exposure assignment and Eligibility at time zero
An active comparator was included to reduce confounding by indication. The study design appropriately handled treatment-confounder feedback by including inverse probability weighting to adjust for time-varying confounders. Both treatment assignment and the determination of eligibility occurred simultaneously at time 0 preventing selection bias from inclusion of prevalent users and immortal time bias from the use of postbaseline treatment information to define baseline treatment status.
Our results showed both an increase of 1.6% and a decrease of approximately 2% in the risk of dementia after 12 years of follow-up when comparing initiators of antidepressants with anticholinergic properties with initiators of antidepressants without anticholinergic properties. Long-term exposure to antidepressants, regardless of their anticholinergic properties, slightly increased the risk of dementia.