The incidence of antidepressant discontinuation symptoms appears to be much lower than was previously thought, results from a new meta-analysis of studies assessing this issue showed.
After accounting for placebo effects, results showed that about 15% of patients who discontinue antidepressant therapy had true discontinuation symptoms, with severe symptoms occurring in about 2% of patients.
“Considering all available data, we conservatively estimate that one out of every six to seven patients has truly pharmacologically-caused antidepressant discontinuation symptoms. This might still be an over-estimate, as it is difficult to factor in residual or re-emerging symptoms of depression or anxiety,” the researchers concluded.
The study was published online in The Lancet.
More Reliable Data
“We are not saying all antidepressant discontinuation symptoms are a placebo effect. It is a real phenomenon. And we are not saying that there is no problem discontinuing antidepressants. But these findings suggest that true antidepressant discontinuation symptoms are lower than previous studies have suggested,” study investigator, Christopher Baethge, MD, University of Cologne, Germany, said at a Science Media Centre press briefing.
“Our data should de-emotionalize the debate on this issue. Yes, antidepressant discontinuation symptoms are a problem, but they should not cause undue alarm to patients or doctors,” Dr. Baethge added.
Lead investigator, Jonathan Henssler, MD, Charité – Universitätsmedizin Berlin, Germany, noted that “previous studies on this issue have included surveys which have selection bias in that people with symptoms antidepressant discontinuation are more likely to participate. This study includes a broader range of research and excluded surveys, so we believe these are more reliable results.”
A Controversial Issue
The investigators note that antidepressant discontinuation symptoms can be highly variable and nonspecific, with the most frequently reported symptoms being dizziness, headache, nausea, insomnia, and irritability. These symptoms typically occur within a few days and are usually transient but can last up to several weeks or months.
Explaining the mechanism behind the phenomenon, Dr. Baethge noted that selective serotonin reuptake inhibitor antidepressants increase the available serotonin in the brain, but the body responds by reducing the number of serotonin receptors. If the amount of available serotonin is reduced after stopping the medication, then this can lead to discontinuation symptoms.
However, the incidence and severity of these symptoms remains controversial, the researchers noted. They point out that some estimates suggest that antidepressant discontinuation symptoms occurred in the majority of patients (56%), with almost half of cases classed as severe.
Previous attempts at assessment have been questioned on methodologic grounds especially because of inclusion of online surveys or other studies prone to selection and dissatisfaction bias.
“Medical professionals continue to hold polarized positions on the incidence and severity of antidepressant discontinuation symptoms, and the debate continues in public media,” they wrote.
This is the first publication of a larger project on antidepressant discontinuation symptoms.
For the study, the researchers conducted a meta-analysis of 44 controlled trials and 35 observational studies assessing the incidence of antidepressant discontinuation symptoms including a total of 21,002 patients. Of these, 16,532 patients discontinued antidepressant treatment, and 4470 patients discontinued placebo.
Incidence of at least one antidepressant discontinuation symptom occurred in 31% of patients stopping antidepressant therapy and in 17% after discontinuation of placebo, giving a true rate of pharmacologic-driven antidepressant discontinuation symptoms of 14%-15%.
The study also showed that severe discontinuation symptoms occurred in 2.8% of those stopping antidepressants and in 0.6% of those stopping placebo, giving a true rate of severe antidepressant discontinuation symptoms of around 2%.
There was no association with treatment duration or with pharmaceutical company funding, and different statistical analyses produced similar results, suggesting the findings are robust, Dr. Baethge reported.