Text version
Let’s look at a paper that suggests that antidepressants might increase the risk of dementia. There have been competing sort of hypotheses around what antidepressants might do as far as this risk, and the previous epidemiologic work has had mixed results. This paper is a landmark in terms of advancing us further in assessing that risk.
The first hypothesis says that antidepressants have anticholinergic effects, many of them, and that impairs cognitive function. So, maybe these could present a risk for dementia. Conversely, the other hypothesis says that antidepressants have increased brain-derived neurotrophic factor, anti-inflammatory effects, and have induced hippocampal neurogenesis. So, maybe they reduce the risk of dementia.
One of the reasons the previous epidemiologic studies have been mixed is because they have adjusted by different covariates, one study adjusting for diabetes and hypertension, and sometimes not doing that kind of analysis. We have data from a dementia registry (from a team of investigators in Israel), that’s maintained by one of the health maintenance organizations. They looked at a 4-year window from 2013 through 2016 and took all of the patients who were over 60 years old; some were in their 80s, who received a new antidepressant prescription at that time and continued it for at least 2 months. Then, they compared that group to those who were matched for gender, age, and follow-up time. The striking result was that, in their sample of 71,000 patients roughly, many of them had no antidepressant exposure. However, in those who had antidepressant exposure, the rate of developing dementia was 11% vs 2.6% in those not exposed. That is a huge difference.
What exactly does “exposure” mean? They defined this as no antidepressant in 2012, the year before the observation window, and then monotherapy with a new antidepressant that was “continuously prescribed and purchased for at least 60 days.” That means a 2-month exposure as a minimum. The investigators tried to be careful about these potential covariates and tried to control for multiple comorbid conditions, like obesity and cancer but also anxiety disorder and migraine. Throughout, they got the same highly statistically significant results—a higher incidence of dementia in patients with antidepressant exposure. Remember, this is starting an antidepressant at age 60 or older. They also repeated their analysis, looking at different antidepressant classes, like tricyclics or SSRIs or SNRIs. There was no significant difference there either. All of the antidepressants had the same association.
So, an editorial accompanying the article emphasizes that depression is a significant risk factor for dementia and reminds us that effectively treating late-life depression would theoretically and very likely lower the risk. However, they point out that response rates in older patients are also low. The editorial concludes by saying, prescribe antidepressants only to patients with a reasonable probability of response and avoid perpetuating ineffective antidepressant trials.
In summary, this isn’t the first study to raise a concern about antidepressants potentially increasing the risk of dementia, but it is a replication that has tried to address the complexity of comorbidities and other covariates. It concludes that the risk is substantial.
Abstract
Exposure to Antidepressant Medication and the Risk of Incident Dementia
Arad Kodesh, Sven Sandin, Abraham Reichenberg, Anat Rotstein, Nancy L Pedersen, Malin Ericsson, Ida K Karlsson, Michael Davidson, Stephen Z Levine
Objective: To test competing hypotheses that monotherapeutic antidepressant exposure is associated with an increased versus a decreased risk of dementia.
Methods: A prospective national matched cohort study from Israel (N = 71,515) without dementia (2002-2012) aged 60 and over were followed up for incident dementia from May 2013 to October 2017. Exposure to antidepressant monotherapy was classified with Anatomical Therapeutic Chemical Codes (N06A) from January 1, 2013 to December 31, 2016. The association between antidepressant monotherapy and the risk of incident dementia was quantified with hazard ratios (HR) and their 95% confidence intervals (CI) obtained from Cox regression models unadjusted and adjusted for 42 covariates. The robustness of the results was tested with 24 sensitivity analyses: 19 analyses restricted to subsamples with plausible differential dementia risks (e.g., anxiety and depression), and 5 analyses across and within antidepressant drug classes.
Results: In the primary analysis, the risk of incident dementia for the group exposed to antidepressant monotherapy compared to the group unexposed to antidepressants was estimated with an unadjusted HR = 4.09 (df = 1, 95% Wald CI = 3.64, 4.60) and an adjusted HR = 3.43 (df = 1, 95% Wald CI = 3.04, 3.88). Across the 24 sensitivity analyses the estimated adjusted HR values ranged from 1.99 to 5.47.
Conclusion: In this study, monotherapeutic antidepressant exposure in old age was associated with increased incident dementia. Clinicians, caregivers, and patients may wish to consider this potentially negative consequence of antidepressant exposure and aim to balance the costs and benefits of treatment.
Download PDF and other files
Reference
Kodesh, A., Sandin, S., Reichenberg, A., Rotstein, A., Pedersen, N. L., Ericsson, M., Karlsson, I. K., Davidson, M., & Levine, S. Z. (2019). Exposure to antidepressant medication and the risk of incident dementia. The American Journal of Geriatric Psychiatry, 27(11), 1177-1188.
