Six weeks in, the fog lifts. Three months in, you feel like yourself. And then the obvious question arrives: if the problem is solved, why am I still taking this?
The answer turns on a distinction that is easy to miss. Feeling better means the treatment is working. It does not mean the episode is over.
The short answer
The standard minimum is 4 to 9 months after you reach remission, at the same dose that got you there. Not 4 to 9 months from when you started the medication.
If you have had three or more episodes of depression, or your depression has been chronic, guidance shifts toward longer-term maintenance treatment measured in years rather than months.
The reason is relapse risk, and the numbers behind it are unusually clear for psychiatry.
Why feeling better is not the finish line
Clinicians divide treatment into three phases, and the confusion almost always comes from mistaking the first for the whole thing.
- Acute phase. From starting treatment until symptoms remit. Usually 6 to 12 weeks.
- Continuation phase. From remission onward, to prevent the same episode from returning. This is the 4-to-9-month window.
- Maintenance phase. Longer-term treatment to prevent a new episode, for people at elevated risk of recurrence.
The distinction between continuation and maintenance is not academic. Continuation treats the episode you are still technically in. Maintenance prevents the next one. They have different justifications and different durations.
An untreated depressive episode typically runs its course over many months. Medication suppresses the symptoms; it does not shorten the underlying episode. Stopping at three months often means stopping in the middle of something that was still active underneath.
The minimum: 4 to 9 months past remission
American Psychiatric Association guidance is specific on two points. Patients in whom pharmacotherapy has been successful should continue treatment at the same dosage for four to nine months, and depression-focused cognitive behavioral therapy is also recommended during this phase.
The same-dose instruction is worth emphasizing, because reducing the dose once you feel well is intuitive and counterproductive. The dose is what is holding the improvement in place. A common pattern is a half-dose reduction at month three followed by a return of symptoms at month five, which then gets read as "the medication stopped working."
Other guidelines converge on a similar floor. NICE guidance in the UK recommends continuing for at least six months after remission, and longer where relapse risk is high.
When longer treatment is recommended
The threshold most often cited is three or more prior episodes, or chronic major depressive disorder. Those patients are advised to move from the continuation phase into maintenance rather than stopping.
Additional factors that push toward longer treatment:
- Residual symptoms. Partial rather than full remission is one of the strongest predictors of relapse.
- Ongoing psychosocial stressors. Bereavement, caregiving, financial strain, or an unresolved living situation.
- Early age at onset.
- Family history of mood disorders.
- Severity of past episodes, particularly any involving suicidal ideation or hospitalization.
Residual symptoms deserve particular attention because they are so often tolerated. Sleep that never fully normalized, persistent anhedonia, or ongoing concentration problems are not the finish line. Our guides to anhedonia and depression-related brain fog cover two of the most commonly overlooked residual symptoms.
What does the evidence actually show?
The foundational analysis is Geddes and colleagues, published in The Lancet in 2003. It pooled 31 randomized trials covering 4,410 participants and found:
- 41% average relapse on placebo
- 18% average relapse on continued antidepressant treatment
- Roughly a 70% reduction in the odds of relapse from continuing
That is a large effect by the standards of almost any preventive treatment, and it held across drug classes and follow-up durations. It is the basis for essentially every continuation-phase recommendation in current guidelines.
The ANTLER trial: what happens when people stop
The more directly relevant study for anyone already well is ANTLER, published in the New England Journal of Medicine in 2021.
It enrolled 478 UK primary care patients who had been on a stable antidepressant dose, most for at least two years, and who felt well enough to consider stopping. Half were maintained on their medication; half were tapered to placebo. Everyone was followed for 52 weeks.
Results at one year:
- Maintenance group: 39% relapsed (92 of 238)
- Discontinuation group: 56% relapsed (135 of 240)
- Hazard ratio 2.06 (95% CI 1.56 to 2.70), roughly a doubling of risk
Two findings deserve equal weight. Stopping doubled the relapse risk. And roughly one third of the discontinuation group completed the full year without restarting.
That second number matters. It means stopping is a genuine option for a substantial minority of people, even those on long-term treatment. It also means it is a gamble with roughly even odds over a year, which is exactly why it should be planned rather than improvised.
Factors that shift the answer for you
Work through these before raising the question:
- How many episodes have you had? One, two, or three or more. This is the single largest input.
- How long since full remission? Count from when symptoms resolved, not from the prescription date.
- Is remission actually complete? Sleep, energy, concentration, and interest all back to your baseline, or only some of them?
- What is happening in the next six months? Discontinuation during a known stressful period is poor timing.
- What season is it? If your episodes have a seasonal pattern, stopping in early autumn is the worst available choice. See our seasonal depression timeline.
- Do you have psychological tools in place? Relapse prevention skills work far better learned in advance than assembled during a relapse.
How to stop when the time comes
Three principles, none of them complicated:
Taper slowly, and slower at the end. The final steps down are usually the hardest, because the relationship between dose and receptor occupancy is not linear at low doses. A taper that is comfortable from 20 mg to 10 mg can become difficult from 5 mg to zero.
Distinguish withdrawal from relapse. Discontinuation symptoms typically appear within days of a reduction, often include physical features such as dizziness or electric-shock sensations, and improve over one to three weeks. Relapse tends to arrive later, builds gradually, and reproduces your familiar depressive symptoms. Confusing the two leads people to conclude they cannot manage without medication when they were experiencing a time-limited withdrawal effect. Our guide to what happens when you stop taking antidepressants covers this distinction in detail.
Plan the monitoring in advance. Agree with your prescriber what early relapse would look like for you specifically, who is watching for it, and what the restart threshold is. Most people have a recognizable early signature: a particular sleep change, a specific withdrawal from contact, a familiar shift in thinking.
Questions for your prescriber
- When did we agree I reached full remission? The continuation clock starts there.
- Given my episode history, am I a continuation case or a maintenance case? This is the decision that sets the timeline.
- Do I have residual symptoms we have been tolerating?
- What would a taper look like, and over how many weeks?
- What is my restart plan if symptoms return?
Stopping an antidepressant is a legitimate goal and, for many people, an achievable one. The evidence argues against doing it early, quickly, or on the basis of feeling well, because feeling well is the treatment working rather than the reason to end it.
Last updated: August 2026. This article is for informational purposes only and does not constitute medical advice. Never stop or change an antidepressant without speaking to your prescriber. If you are in crisis, call or text 988 (Suicide & Crisis Lifeline) or go to your nearest emergency room.