How Long To Take Antidepressants

Counted from feeling well, not from starting

The standard answer is at least six months after symptoms have remitted, with regular review. The critical detail is the starting point: six months from the day you felt well, not six months from the day you started the tablet. Someone who took four months to respond is looking at ten months of treatment, not six.

Indian Psychiatric Society guidance is slightly longer and asks for six to nine months after remission, at the same dose that produced it. Dropping to half the dose as a compromise is not part of the recommendation. Continuation means continuation.

The size of the difference

The landmark synthesis of thirty-one randomised trials covering 4,410 patients found relapse in forty-one per cent of those switched to placebo against eighteen per cent of those who continued treatment, over an average of about a year. That is a difference of roughly twenty-three percentage points, meaning about one relapse prevented for every four or five people who continue.

In older adults, continuing reduced recurrence at twelve months from sixty-one per cent to forty-two per cent. In persistent, chronic depression the difference was fourteen per cent against thirty-four per cent.

Every one of those trials enrolled people who had already responded. Feeling well was the entry condition, not the protection.

The trial that answered the actual question

The ANTLER trial is the most directly relevant study, because it enrolled the people who actually ask this question: 478 UK primary-care patients who had been on an antidepressant for at least nine months, had two or more previous episodes, and were well enough to consider stopping. Half continued, half were tapered onto placebo, and nobody knew which.

By fifty-two weeks, fifty-six per cent of those who stopped had relapsed, against thirty-nine per cent of those who continued. That is the headline finding, and the other half of it matters just as much: about forty-four per cent of people who stopped were still well a year later. The authors’ own conclusion was that people who discontinue are at increased risk, and that a substantial proportion can discontinue without relapse.

Two caveats belong with those numbers. Withdrawal symptoms were significantly commoner in the discontinuation group, peaking around twelve weeks. And about thirty-nine per cent of that group had restarted an antidepressant by the end of the study.

What moves the recommendation

The factors that lengthen recommended treatment are reasonably consistent across guidelines.

For long-term maintenance, Indian guidance is explicit that people with three or more relapses or recurrences need long-term treatment, and that many will do best continuing indefinitely. Even then, treatment should be reviewed at least every six months, because indefinite is not the same as unexamined.

Continuing has costs too

A conversation about duration is not complete if it only lists the risks of stopping. Guidance requires the risks of continuing long-term to be discussed as well: side effects, increased bleeding risk, long-term effects on sexual function, and the fact that stopping later may be difficult.

Sexual side effects deserve particular mention because they are the commonest reason people stop antidepressants, frequently without telling the prescriber. If that is the issue, there are usually better options than quiet discontinuation, and this article covers them.

In India there is a further practical dimension. Out-of-pocket spending accounts for the great majority of health expenditure, and medicines are consistently the largest single component of the cost of ongoing psychiatric care. A six to nine month course has a real price, and pretending otherwise does not help anyone stay on treatment.

Therapy is the part that changes the odds

This is the most useful finding in the whole area and the least widely known. When stopping was combined with a structured psychological intervention such as preventive cognitive therapy or mindfulness-based cognitive therapy, there was no clear increase in relapse compared with staying on medication, and between forty and seventy-five per cent successfully discontinued.

Mindfulness-based cognitive therapy has been shown in an individual-patient-data meta-analysis of nine trials to reduce relapse compared with usual care, and also compared with maintenance antidepressants, with larger effects in people who were more severely affected at baseline.

Relapse-prevention work is usually about eight sessions over two to three months, focused on identifying warning signs, mapping triggering circumstances and making a specific plan. If you intend to come off medication, arranging this first changes the odds more than any tapering schedule will.

Telling the two apart by timing

Confusing withdrawal with relapse is the commonest reason people conclude that they can never come off. The distinction is mostly about timing and content.

A meta-analysis of seventy-nine studies covering more than 21,000 participants found any discontinuation symptom in about thirty-one per cent of people stopping an antidepressant against about seventeen per cent stopping placebo, with severe symptoms in around three per cent against under one per cent. Rates were lower with sertraline, fluoxetine and citalopram, and higher with venlafaxine and paroxetine. Fluoxetine’s long half-life means it effectively tapers itself.

Proportional steps, getting smaller near the end

Current guidance is specific: reduce the dose step by step, with each step being a proportion of the previous dose rather than a fixed number of milligrams, and use smaller proportional reductions as the dose gets lower. If withdrawal symptoms are severe, the recommended response is to go back to the previous dose, let symptoms settle, and then reduce more slowly in smaller steps.

The pharmacological reason is worth knowing. The relationship between SSRI dose and serotonin transporter occupancy is hyperbolic rather than linear, so equal milligram reductions produce progressively larger biological changes as the dose falls. That is precisely why the last few milligrams are the hardest, and why a taper that went well from 40 mg to 20 mg can fall apart from 10 mg to zero.

Two honest caveats. Most tapering schedules in the trials lasted four weeks or less, so the evidence base for slower tapers is thinner than the argument for them. And a failed taper is information, not a life sentence: it tells you the rate was wrong.