Cumulative Antidepressant Use and Risk of Dementia in a Prospective Cohort Study

Laura Heath, PhD; Shelly L. Gray, PharmD, MS; Denise M. Boudreau, PhD; Ken Thummel, PhD; Karen L. Edwards, PhD; Stephanie M. Fullerton, PhilD; Paul K. Crane, MD, MPH; Eric B. Larson, MD, MPH


J Am Geriatr Soc. 2018;66(10):1948-1955. 

In This Article

Abstract and Introduction


Objectives: To determine whether antidepressant use is associated with dementia risk.

Design: Prospective cohort study.

Setting: Kaiser Permanente Washington (KPWA), an integrated healthcare delivery system.

Participants: Community-dwelling individuals aged 65 and older without dementia and with 10 years or more of KPWA enrollment at baseline (N=3,059).

Measurements: Primary exposures were selective serotonin reuptake inhibitors (paroxetine vs other), tricyclic antidepressants, and serotonin antagonist and reuptake inhibitors. Using health plan pharmacy data, we calculated cumulative medication exposure, defined as total standardized daily doses (TSDDs), over rolling 10–year windows. Exposure in the most recent year was excluded to avoid use related to prodromal symptoms. The Cognitive Abilities Screening Instrument was administered every 2 years; low scores triggered clinical evaluation and consensus diagnosis procedures. Dementia risk was estimated according to medication use using Cox proportional hazards models.

Results: During a mean follow-up of 7.7 years, 775 participants (25%) developed dementia; 659 (22%) developed possible or probable Alzheimer's disease. Individual antidepressant classes were not associated with differences in dementia risk, although paroxetine use was associated with higher risk of dementia for all TSDD categories than no use (0–90 TSDDs: hazard ratio (HR)=1.69, 95% confidence interval (CI)=1.18–2.42; 91–365 TSDDs: HR=1.40, 95% CI=0.88–2.23; 366–1095 TSDDs: HR=2.13, 95% CI=1.32–3.43; ≥1095 TSDDs: HR=1.42, 95% CI=0.82–2.46).

Conclusion: Most commonly prescribed nonanticholinergic depression medications used in late life do not appear to be associated with dementia risk. Paroxetine and other anticholinergic antidepressants may be exceptions in older individuals. Future studies are warranted to improve scientific understanding of potential associations in other settings and populations.


Antidepressants are among the most commonly prescribed medications in the United States,[1] but little is known about the long-term effects of antidepressant treatment on risk of dementia and late-onset Alzheimer's disease (AD). Depression itself has been characterized as a risk factor for dementia in multiple longitudinal studies[2–8] and as a symptom of prodromal AD,[8] with a recent meta-analysis indicating a greater risk of dementia in individuals with late-life depression (hazard ratio (HR)=1.98, 95% confidence interval (CI)=1.50–2.63).[9] Although the mechanisms are unknown, several pathways linking depression to development of dementia have been hypothesized, including depression as a contributory factor in inflammation, glucocorticoid levels and hippocampal atrophy, and β–amyloid deposition.[4]

It is unknown whether pharmacological treatment for depression affects the risk of developing dementia by interrupting pathways between depression and dementia or due to direct risk-altering effects of antidepressant medications themselves. Animal studies suggest that multiple classes of antidepressants, including selective serotonin reuptake inhibitors (SSRIs), tricyclic antidepressants (TCAs), and serotonin antagonist and reuptake inhibitors (SARIs), appear to have neurogenic and neuroprotective properties,[10–13] but anticholinergic medications, which include TCAs and the SSRI paroxetine, are not recommended in older adults because of the risk of dementia and impaired cognitive and physical outcomes associated with anticholinergic medication exposure.[14–16]

Several studies have examined the link between antidepressants and dementia directly,[17–26] with conflicting results. Our previous study focused on anticholinergic medications, including anticholinergic antidepressants (TCAs and paroxetine). We found a greater risk of dementia and AD in heavier users of these drugs, even when we accounted for antidepressant use possibly prescribed for prodromal symptoms of dementia.[22] In the current study, we extended this prior analysis within the Adult Changes in Thought (ACT) cohort[27] and focused on 3 major classes of antidepressants: SSRIs, TCAs, and SARIs. We hypothesized that antidepressants with anticholinergic properties (paroxetine, TCAs) would be associated with greater risk of dementia and AD and that nonanticholinergic antidepressants (other SSRIs other than paroxetine, SARIs) would be associated with lower risk of dementia and AD.