REVIEW OF: Lewis G et al, N Eng J Med 2021;385(14):1257–1267

TYPE OF STUDY: Randomized controlled trial

Most of us have been taught that long-term antidepressant therapy is crucial for patients who have had three or more episodes of depression. This idea is based on trials in which patients in remission were randomly assigned to medication continuation vs placebo. Those switched to placebo had a higher rate of relapse, especially if they’d had three prior episodes of depression.

However, these studies have a number of flaws that may bias the results. Generally, these patients were in remission for only a short amount of time—typically from three to eight months. Second, antidepressant discontinuation was typically done rapidly, making it hard to tell whether patients actually relapsed or were suffering from withdrawal symptoms. Third, the switch to placebo took place at a fixed time instead of when patients felt ready to come off their antidepressant. That leaves open the question: Can patients who have been in sustained remission from depression and feel ready to come off medication safely discontinue antidepressants?

In this study, researchers recruited 478 adult patients from 150 general practices across England, all of whom had a history of at least two prior depressive episodes, were currently in remission, had been taking their antidepressant for at least nine months, and felt well enough to consider stopping their medication. Only antidepressants that are commonly prescribed and known to have low rates of withdrawal were included: citalopram, sertraline, fluoxetine, and mirtazapine. Consenting patients were randomized in a double-blind manner to either remain on their antidepressant or have it slowly replaced with a placebo over two months (fluoxetine was tapered over only one month due to its long half-life). Patients were then followed every three months for one year to ascertain relapse rates.

In total, 92 of 238 patients (39%) in the maintenance group relapsed compared to 135 of 240 (56%) in the discontinuation group over the course of 52 weeks (HR 2.06; p < .0001). Differences in rates of depression, anxiety, and quality of life all emerged within 12 weeks and persisted throughout the trial. Of the patients who stopped their trial medication, 20% of the maintenance group and 39% of the discontinuation group elected to resume antidepressant medication.

One potential limitation of the study is that reported rates of medication withdrawal symptoms were higher in the discontinuation group (3.1) than the maintenance group (1.9; 95% CI 1.5–2.3). Although this could represent an unmasking of depressive symptoms, it also suggests that despite the slow taper, physiological withdrawal cannot be ruled out as a contributing factor.

CARLAT TAKE
This landmark study puts the risks of antidepressant discontinuation in perspective. On the one hand, staying on the medication lowers the relapse risk by 17% over the course of a year. On the other hand, 44% of discontinuation patients did taper off their antidepressant without relapsing. Those figures challenge the black-and-white recommendation to remain on an antidepressant indefinitely after more than two episodes and move the question into the realm of collaborative decision-making between clinician and patient.