October 07, 2019
2 min read

Antidepressants not associated with adverse health outcomes

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Marco Solmi

Current evidence does not support claims of adverse health outcomes associated with antidepressant use, and such claims are likely exaggerated by confounding by indication, according to findings from a systematic umbrella review published in JAMA Psychiatry.

“Antidepressants are overall safe, and according to objective quantitative criteria, there is no strong evidence that they have any serious adverse health outcomes,” Marco Solmi, MD, PhD, from the department of neurosciences at the University of Padova in Italy, told Healio Psychiatry. “Depressive disorder can have adverse health outcomes, such as cardiovascular disease among other conditions, and antidepressants might instead help people to take care of their physical health.”

According to Solmi and colleagues, antidepressant drugs are ranked third among prescribed and fourth among sold medications. Further, up to 8% to 10% of adults in the United States take at least one antidepressant drug. The researchers noted that following the FDA’s introduction of the black box warnings that associated selective serotonin reuptake inhibitor use with a higher risk of suicidal behavior in children and adolescents, controversy arose regarding the safety profile of antidepressants.

The researchers reviewed 45 meta-analyses of observational studies and 252 full-text articles. They included no age, sex, population or medical condition restrictions.

Solmi and colleagues reported that these studies described 120 associations between antidepressant use or exposure and adverse health outcomes, including data from 1,012 individual effect size estimates. Of the associations, the researchers considered 61.7% as nominally significant (P .05) using random effects models. They found that 43.4% had large heterogeneity (I2 > 50%), 14.2% had small-study effects and 7.5% had excess significance bias. They found convincing evidence for associations between antidepressant use and suicide attempt or completion among patients aged younger than 19 years, and between antidepressant use before and during pregnancy and autism risk among offspring. After conducting a sensitivity analysis that adjusted for confounding by indication, the researchers found that no associations remained at the convincing evidence level.

“People who might feel the need to see a psychiatrist but are avoiding it because of bad information on antidepressants might be reassured by the findings of our study,” Solmi said. “Second, people who have already seen a psychiatrist but are not taking medications despite a prescription might be helped in having a second thought based on our findings. Third, people who are taking medications but are afraid of negative effects and are planning to withdraw them are now informed that there is no reason to drop antidepressants, in general.”

In a related editorial, José M. Flores, MPH, MD, PhD, from the Yale Child Study Center at Yale University School of Medicine, and colleagues highlighted the risk of reductionism that accompanies meta-analysis and umbrella reviews.

“When meta-analysis is conducted on observational studies that are susceptible to systematic confounding or bias, the meta-analysis will yield more precise estimates of the strength of an association, but that measured association may still differ systematically from the true causal effect (eg, through narrower confidence intervals around confounded or biased point estimates.”

Thus, this umbrella review may “report more precise estimates of associations that are subject to confounding and bias, leading to false accuracy and certainty,” they wrote. – by Joe Gramigna

Disclosures: Flores reports funding from the NIH. Solmi reports no relevant financial disclosures. Please see the full study/editorial for all other authors’ relevant disclosures.