REVIEW OF: Lagerberg et al, Eur Neuropsychopharm 2020;26:1–9

TYPE OF STUDY: Analysis of Swedish national registries

Soon after the introduction of SSRIs in 1988, case reports began to emerge suggesting that they might trigger violence in a small subset of patients. At the time, such reports were mostly dismissed and attributed to the fact that patients receiving antidepressant therapy presented during times of crisis, so correlation was thought to be mistaken for causation. So far, the epidemiological studies have provided conflicting results on this issue.

In this study, researchers utilized various Swedish databases to examine rates of violent acts (such as homicide, robbery, assault, and threats) for individuals aged 15–60 who received SSRI therapy between 2006 and 2013. They specifically focused on the timing of violent acts in relation to SSRI treatment by demarcating 6 phases of treatment: 3 on SSRI therapy (days 0–28, days 29–84, and days > 84) and the same 3 time frames after the SSRI was discontinued. The researchers attempted to control for a variety of confounding variables.

In total, 785,337 individuals were included and followed for an average of 7.3 years. Violent crime convictions occurred in 2.7% (n = 20,821) of the cohort. Overall, SSRI use was associated with a modest increase in violent crime (risk = 1.10, confidence interval 1.06–1.13). Further analyses suggested that most of this increase occurred in the subset of individuals who (1) were under the age of 35, and (2) had a prior history of violence. Benzodiazepines were also associated with an increased risk of violence (risk = 1.32, confidence interval 1.21–1.55) in this analysis, but other psychotropic medications were not.

In examining the 6 phases of treatment, no pattern emerged other than that the risk of violence dissipated about 3 months (84 days) after SSRI discontinuation.

TCPR’s Take
This study is the most rigorous take yet on the controversial link between SSRIs and violence. However, it still can’t confirm causality because it’s always possible that SSRI therapy is a proxy for distress, and that patients come off SSRIs when their lives become more stable, causing a false association. Still, paradoxical reactions do occur in psychiatry—benzodiazepines usually calm agitation but can make the occasional patient disinhibited; SSRIs generally lower irritability but may do the opposite if they cause mania or akathisia. Nonetheless, even if the causality implied here is true, it would only represent an increase of roughly 4 violent acts per every 10,000 SSRI trials. While still worthy of pursuit for epidemiological research, such a low risk does not warrant a change in clinical practice.

To learn more, listen to our 2/15/21 podcast, “Six Depressions That Can Worsen on Antidepressants.” Search for “Carlat” on your podcast store.