View commentaries on this research

This is a plain English summary of an original research article

Taking long-term antidepressants can prevent depression recurring (relapse). But new research shows that almost half of those who stop taking the medication do not relapse.

Depression is a major cause of ill health and disability worldwide. It causes emotional distress and interferes with everyday life. Many people with depression continue taking antidepressant drugs for months or even years after their symptoms have resolved. This so-called maintenance therapy aims to reduce the risk of relapse.

The numbers of people taking maintenance therapy for depression is increasing. However, there is little research to show how effective these drugs are in preventing relapse in people who have been taking them long-term.

This study included people who had two previous relapses of depression. Researchers compared rates of relapse in those who continued on antidepressants with those who stopped. They found that people who stopped medication were more likely to relapse. However, more than 4 in 10 people who stopped taking antidepressants had no relapse of their depression.

More information about depression is available on the NHS website.

What’s the issue?

Most people feel down from time to time, but people with depression have a low mood that lasts for weeks or months. They typically have little interest in things they used to enjoy and can withdraw socially, sleep and eat more or less than usual, and have poor concentration.

More than 300 million people worldwide have depression. Most recover with the right treatment and support. But around half of those who have recovered go on to relapse.

The condition is usually treated with antidepressants. However, these drugs sometimes have side-effects including indigestion, and feeling agitated. Side-effects usually improve over time.

Guidance from the National Institute of Health and Care Excellence recommends that antidepressants are used as ‘maintenance’ treatment for up to 2 years to prevent their depression returning (relapse). It also recommends cognitive-behavioural therapy to change habits of thought and behaviour.

An increasing number of people are taking antidepressants as long-term maintenance. The chance of relapse is reduced by staying on the drugs for a few months after depression has cleared. However, there is little evidence that these drugs prevent relapse when taken for longer than 8 months.

In this study, researchers explored relapse rates among people who had taken them for more than 9 months.

What’s new?

The study included 478 people aged 18 - 74 years, who were treated at 150 general practices in Bristol, London, Southampton, and York. Participants had all had two previous relapses of depression. They had been taking one of the 4 most commonly prescribed antidepressants (citalopram, sertraline, fluoxetine, or mirtazapine) for at least 9 months, and were well enough to consider stopping.

Half (238 people) continued taking their usual antidepressant (treatment group), while the others in the discontinuation group (240 people) took dummy pills (placebo). These dummy pills looked identical to people’s usual pills and initially contained their normal dose of antidepressant. This dose was gradually reduced over 1-2 months until the placebo contained no antidepressant at all.

Participants’ depressive symptoms were assessed at the start of the trial, and then again at various time points (12, 26, 39 and 52 weeks).

The study found that:

  • by 52 weeks, more people on placebo relapsed (56%) than those who continued with treatment (39%)
  • more than 4 in 10 people (44%) on placebo did not relapse
  • most relapses occurred 12 to 26 weeks after the study started
  • people in the placebo group had worse anxiety and depression scores than those in the treatment group, particularly at 12 and 26 weeks
  • discontinuing treatment cost marginally more than continuing over 12 months (because people who re-start treatment have more healthcare appointments, which cost more than continuing on treatment).

Not everyone who relapsed went back on to medication. Of 134 people in the placebo group who relapsed, most (53%) went back on to antidepressants prescribed by their doctor. But 49 people (37%) remained on placebo in the trial, and a further 14 (10%) chose not to take any medication.

More people on placebo had withdrawal symptoms (which can feel similar to anxiety and depression and include sleeping problems and restlessness) than in the treatment group.

Why is this important?

The study suggests that some people who would have remained free of depression with treatment, relapse when they discontinue. But more than 4 in 10 are likely to remain well without treatment. These findings should inform discussions between doctors and people who have had depression. Both need to be aware of the likely benefits and harms of stopping long-term treatment with antidepressants.

Follow-up during the study identified the time points at which people might have worse anxiety or depression or be more vulnerable to relapse. Knowing in advance that these symptoms may appear - and then resolve - is helpful. For instance, many people discontinuing medication felt their mood worsening at 12 weeks, but this may have not been severe enough for them to restart treatment.

What’s next?

The researchers say their results should be interpreted with care since the study included only people who had had two previous relapses. The findings might not apply to people receiving treatment for their first depressive episode.

Participants were mostly White, married and employed, and were recruited from moderately sized general practices in urban areas. Whether the same patterns are seen in other groups (people who are younger, unemployed or from minority ethnic groups) needs to be explored.

Outstanding questions include:

  • does reducing the dose of antidepressant more gradually (over a longer period) have benefits
  • what is the relapse rate after 12 months
  • since this study looked at 4 antidepressants, would other antidepressants that work in different ways have different relapse rates.

It would also be useful to explore the reasons behind individual decisions to stop medication during the study.

You may be interested to read

The paper this NIHR Alert was based on: Duffy L, and others. Antidepressant medication to prevent depression relapse in primary care: the ANTLER RCT. Health Technology Assessment 2021;25:69

National Institute for Health and Care Excellence (NICE). Depression in adults: recognition and management. NICE guideline [CG90]. 2009

The cost-effectiveness analysis from the ANTLER study: Clarke CS, and others. Cost utility analysis of discontinuing antidepressants in England primary care patients compared with long term maintenance: the ANTLER study. Applied Health Economics and Health Policy 2022;20:269-282

Funding: This study was funded by the NIHR Health Technology Assessment programme.

Conflicts of Interest: Several authors report competing interests.

Disclaimer: NIHR Alerts are not a substitute for professional medical advice. They provide information about research which is funded or supported by the NIHR. Please note that views expressed in NIHR Alerts are those of the author(s) and reviewer(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

  • Share via:
  • Print article


Study authors

Antidepressant prescriptions have been rising steeply since the mid-1990s and people are staying on them for longer. Most prescriptions are to prevent relapse in people who have recovered from depression. Many studies have looked at how effective antidepressants are at preventing relapse, but only in people who have been taking them for a short time. For people taking antidepressants long-term, there is no evidence that antidepressants prevent relapse.

This study has shown that long-term maintenance therapy with antidepressants reduces the risk of relapse over 52 weeks. However, being on antidepressants does not guarantee freedom from relapse; and discontinuing them does not mean people will relapse. Almost half (44%) of people who discontinued their medication did not relapse.

People with depression and their doctors need to come to joint decisions. They need to weigh up the risks of continuing with antidepressants (side-effects) and discontinuing (withdrawal) against the risk of relapse.

Glyn Lewis, Professor of Epidemiological Psychiatry; Gemma Lewis, Sir Henry Dale Fellow and Larisa Duffy, ANTLER Trial Manager, Division of Psychiatry, University College London 


This study is an important contribution to our understanding of the benefits of antidepressants for people receiving treatment for depression, for whom the risk of relapse can be an ever-present concern.

The stigma surrounding depression and other conditions thrives in ignorance. It is only by improving our insight that we can hope to truly tackle it, by demystifying treatments and determining the best routes to recovery and remaining well.

Marjorie Wallace, Chief Executive of SANE, the mental health charity 

Mental Health Foundation 

This research provides really useful information for anyone considering withdrawal from long term antidepressant use. It helps to quantify the costs and benefits associated with withdrawal and will support people to make informed choices. It remains clear that many people can safely withdraw and also that maintenance does not remove the risk of relapse. Though this research cannot identify in advance who is likely to relapse on withdrawal, it does help identify the most vulnerable times in the withdrawal process. This might help people to find appropriate support if they decide to withdraw.

It is interesting to note that the cost of withdrawal diminished over time and the reduction in physical medication side effects became apparent later in the trial period. I hope that future research can follow people for even longer as this research seems to show that risk of relapse reduces over time and it seems plausible, therefore, that the cost/benefit balance will improve. Although the research notes that longer term horizons are rare in depression research when the issue under investigation is long term antidepressant use, twelve months is a relatively short time to draw conclusions if reduction in side effects and reduced risk of relapse both improve over time.

David Crepaz‑Keay, Head of Applied Learning, Mental Health Foundation  


This paper contributes to the wider discussion on discontinuing long-term psychiatric medications (for example see Research into Antipsychotic Discontinuation And Reduction [RADAR] trial). It is an important and difficult area of research, especially considering the prevalence of depressive disorders, limited resources for interpersonal/talking therapies such as CBT, and the high number of people remaining on antidepressant therapies for long periods of time.

Some evidence suggests that slower tapering of antidepressants (to doses much lower than what is considered ‘therapeutically beneficial’) or changing to fluoxetine before tapering may lead to fewer side effects or withdrawal syndromes. It is possible the symptoms or relapses seen in this study could be related to the tapering programme, even though it is within current clinical guidelines and this is something to consider when generalising these findings.

Future research (or further exploration of this study’s results) should consider validation of the study’s novel measure of relapse in relation to other validated measures, and could explore whether certain clinical variables (such as length of time on antidepressants, which antidepressant, number of previous relapses, and use of talking therapies) predict higher incidence of relapse.

Victoria, Bristol 

Lived experience

As someone who has taken sertraline for 24 years, this research is relevant to me, but my experience of relapse is so awful that it doesn’t inspire me to try discontinuation again. I’m not aware of any negative effects of taking 50mg sertraline daily, so where is the incentive? As a widow living alone, I must prioritise healthy independence.

I’m not confident that these findings will outweigh the patient’s personal experience in deciding what would be best. I won’t be consulting my GP about changing my prescription. But the study might encourage GPs to explore more options with patients. I very much hope the findings do not lead to public policy decisions to make antidepressants more difficult to obtain on the NHS.

Public Contributor, North-East England

Lived experience 

I have a strong family history of depression and long-term medications, dealing with side effects and the results of reducing doses and changing medications. Conversations and shared decision-making with patients could help people have the confidence to try to reduce or stop their medication.

Sarah Bittlestone, Public Contributor, County Durham 

Back to top