If you are struggling right now: in Ireland and the UK, call Samaritans free on 116 123, day or night. In the US, call or text 988. If you or someone else is in immediate danger, call 112 or 999 in Ireland, or 999 in the UK.
Most people can come off an antidepressant successfully, but not by stopping suddenly. Withdrawal (also called discontinuation) symptoms are real. They include dizziness, electric-shock sensations, poor sleep, anxiety and low mood. How common they are is genuinely disputed. A major 2024 meta-analysis puts the risk caused by the drug itself at roughly one in six or seven people. Other researchers put it at more than half. For many people symptoms are mild and fade within a couple of weeks; for some they last months. Paroxetine and venlafaxine carry more risk than most. NICE and the Royal College of Psychiatrists both advise reducing gradually, in shrinking steps, at a pace you agree with your prescriber.
This guide does not give a dosing schedule, because that has to be built around you.
NICE's 2022 depression guideline lists the main symptoms as:
The Royal College of Psychiatrists adds vivid dreams, a feeling that things are not real, and "cotton wool in your head". It also lists akathisia, an inner restlessness that makes it hard to stay still. Brief "brain zaps" can be made worse by turning your head.
The College describes this as physical dependence, not addiction. Your body has adapted to the drug and reacts when it is withdrawn. Addiction, as the College defines it, involves craving, loss of control and a "high".
Three recent analyses reach very different numbers. We cover the wider debate in what SSRIs actually do; here is what matters if you are planning to stop.
Why the gap? Henssler's team could compare against people stopping placebo, who also report symptoms. Inside the randomised trials alone, the difference between drug and placebo was 8 percentage points. Severe symptoms followed 2.8% of antidepressant discontinuations, against 0.6% after placebo.
The critics argue most of Henssler's data came from industry efficacy trials where withdrawal was a minor consideration. They also say withdrawal may have been misrecorded as relapse. Henssler's team, in turn, say returning depression or anxiety must be considered when reading withdrawal figures. Both points can be true in different patients.
Two caveats. Davies and Read are co-authors of the 2025 reanalysis, so it is not an independent third view. And in that reanalysis, four of the five studies involved 12 weeks or less of prior treatment. People on antidepressants for years are poorly represented in all of this research.
Official patient information varies too. The Royal College says a third to a half of people get symptoms; the NHS says most people get some.
NICE now says withdrawal symptoms usually go within one to two weeks. It adds that they can sometimes last several weeks, and occasionally several months. They can be severe, particularly after a sudden stop.
Davies and Read had challenged the older "self-limiting" wording. In their review, seven of ten studies on duration found a significant share of people had symptoms for more than two weeks. The four studies that calculated a mean duration found 5 days, 10 days, 43 days and 79 weeks, in very different groups of patients.
The Royal College says some people have symptoms for months, or more than a year. It also says nobody can yet predict who will be affected badly. The NHS notes that symptoms can sometimes start weeks or months after a dose change.
NICE singles out paroxetine and venlafaxine as more likely to cause withdrawal symptoms, so particular care is needed with them. One reason is how quickly they leave the body.
NICE says antidepressants with a short half-life need tapering more slowly. It also notes that fluoxetine's prolonged action means it can sometimes be stopped more simply. The catch, says the Royal College, is that fluoxetine symptoms can appear days or even weeks after a reduction.
The Royal College groups antidepressants by withdrawal risk. Its highest-risk group includes desvenlafaxine, duloxetine, mirtazapine, paroxetine and venlafaxine, plus the older MAOI drugs. Sertraline, citalopram, escitalopram and fluoxetine sit in its medium group. Agomelatine and lofepramine are in its lowest.
Henssler's analysis broadly agrees on venlafaxine, desvenlafaxine and paroxetine, but linked escitalopram to more frequent symptoms. Rankings like these start a conversation; they do not predict your experience.
The Royal College says risk seems higher after a high dose taken for a long time, and after a sudden stop or fast reduction.
NICE's 2022 guideline (NG222) tells prescribers to:
If symptoms are severe, NICE says to consider going back to the previous dose, then reducing more slowly with smaller steps.
The Royal College's 2024 patient guide goes further. It says everyone should taper. People on an antidepressant for only a few weeks may manage a couple of steps, but over at least four weeks. People on one for many months or years are best tapering over months or longer. Some need to reach a very small dose, as low as 2% of the original, before stopping.
The College describes "hyperbolic" tapering for higher-risk drugs. The idea is that small doses have a bigger effect on the brain than you would expect. So each cut should get smaller as the dose falls. The same case was made in a 2019 analysis of brain-imaging data by Horowitz and Taylor. They argued for slow tapers to doses well below the usual minimum. Both are among the College guide's expert authors.
Be aware that the research base for tapering methods is thin. A 2021 Cochrane review found most trial tapers lasted four weeks or less. It could draw no firm conclusions, partly because withdrawal symptoms could have been counted as relapse.
Mixing these up matters both ways. Withdrawal mistaken for relapse can keep people on medicine they no longer need; relapse mistaken for withdrawal can delay help. The Royal College suggests three clues.
NICE adds that relapse does not usually happen as soon as you stop or lower the dose. Even after restarting, withdrawal symptoms may take a few days to go. Fluoxetine blurs the timing clue, because its symptoms can start late.
Relapse is a real risk too. The UK ANTLER trial enrolled 478 people in general practice who felt well enough to consider stopping. Over a year, 56% of those tapered onto placebo relapsed, against 39% who stayed on their antidepressant. The placebo group also reported more withdrawal symptoms, which complicates the comparison.
NICE says relapse is more likely after repeated episodes, incomplete recovery, severe depression, or if the stresses behind your depression are still there. For people at higher risk who want to stop, it suggests considering group CBT or mindfulness-based cognitive therapy. Our guide to what actually works for depression covers those options.
Every source we checked says the same. NICE, the Royal College, the NHS, and the UK and US product information all advise against suddenly stopping. Missed doses and partial doses can trigger symptoms too.
The venlafaxine product information reports suicidal thoughts and aggression during dose changes, including stopping. The US label for the same drug advises watching for suicidal thoughts whenever the dose is adjusted up or down. If you notice suicidal thoughts while reducing, the Royal College says to speak to your prescriber immediately.
Rarely, serious side effects justify a fast stop. That decision belongs with a clinician, seen straight away.
Do not add herbal remedies to "help" a taper without checking. Product information for paroxetine, venlafaxine and fluoxetine all list St John's wort among products that can interact. Benzodiazepines are a different case again, where a badly managed stop can be dangerous; we explain why in what benzodiazepines actually do.
If you take an antidepressant for anxiety, the same principles apply. See what actually works for anxiety for the wider treatment picture.
Brand names differ by country: venlafaxine is Efexor XL in the UK and Effexor XR in the US. This article draws mainly on UK guidance from NICE and the Royal College. Your prescriber in Ireland or the US may follow other guidance, but the UK and US product information both advise gradual reduction. Tapering options depend on which tablet and liquid strengths exist where you live, so ask your pharmacist.
The Royal College says this is not quite the same as addiction, which involves craving, loss of control and a high. It calls it physical dependence: your body has adapted and reacts when the drug is withdrawn.
The Royal College advises against it, because drug levels rise and fall and can trigger symptoms. Fluoxetine is a possible exception, but only with your prescriber's agreement.
No. The Royal College says you can still stop, but you may need smaller reductions over a longer time. It suggests going back to the last comfortable dose and waiting until you feel ready.
NICE says treatment may need to continue for at least six months after symptoms go, with regular review. People with repeated or severe episodes may be advised to stay on longer.
This article is general information, not medical advice. Do not start, stop or change any medicine without talking to the clinician who prescribed it.
Reviewed by: Proco Team
Last updated: November 2026