What are antidepressants?
Antidepressants are the most commonly prescribed class of psychiatric medication in the UK and worldwide. They are used primarily to treat depression and anxiety disorders, but also prescribed for conditions including OCD, PTSD, chronic pain, and eating disorders. Around one in six adults in England are currently prescribed antidepressants, a figure that has risen steadily for two decades.
The main classes include selective serotonin reuptake inhibitors (SSRIs, such as fluoxetine, sertraline and citalopram), serotonin-norepinephrine reuptake inhibitors (SNRIs), tricyclics, and newer agents such as mirtazapine and vortioxetine. Debates about their efficacy, safety, mechanisms of action, long-term use, and how to stop them safely remain active and sometimes contentious, making this one of the most important areas for mental health professionals to stay up to date on.
What the research tells us
Our research summaries are based on blogs published since 2018, each written by a clinician or researcher who has read and critically appraised the original study.
Do antidepressants work?
The clearest answer to date came from a landmark 2018 network meta-analysis by Andrea Cipriani and colleagues, the largest ever meta-analysis in psychiatry, pooling 522 trials across 21 antidepressants for adults with major depression. The conclusion was unambiguous: all 21 were more effective than placebo for short-term treatment. This was a pivotal moment for the field, cutting through years of media controversy and providing a solid evidence base for prescribing decisions. Read our blog: Antidepressants can help adults with major depression (Tomlin, Feb 2018).
That said, effect sizes vary. A 2024 survey on the “Smallest Worthwhile Difference” found that patients typically set a higher bar for benefit than clinical trials measure, a reminder that statistical significance and clinical meaningfulness are not the same thing.
Getting the dose right
A 2019 dose-response meta-analysis found that higher doses of antidepressants bring maximum side effects with only marginal additional gains. The common clinical instinct to escalate doses when a patient is not responding is not well supported by the evidence. Lower, optimised doses deserve more attention in practice.
Side effects matter: variations by drug
Emotional blunting is reported by nearly half of people taking antidepressants, according to a 2018 survey, though it remains difficult to disentangle from depression itself. A 2023 Danish RCT added to this picture, finding that escitalopram reduced reinforcement sensitivity compared to placebo, likely a neurological correlate of the blunting patients describe.
Weight gain is a meaningful long-term concern: a 2019 ten-year cohort study confirmed a significant population-level impact. A major 2025 Lancet review compared 30 antidepressants on physical health outcomes. cardiovascular disease, stroke, diabetes, and found substantial variation between drugs. The choice of antidepressant matters for physical health, and these differences should inform prescribing, particularly for people with existing physical health conditions.
The serotonin story
The 2022 Moncrieff review claiming the serotonin hypothesis of depression was unfounded generated enormous media coverage. A 2023 blog by Rebecca Wilkinson and Sameer Jauhar offered a more balanced reading: the picture is nuanced, and the evidence does not justify telling patients their antidepressant “won’t work” or encouraging abrupt cessation. Debate about mechanism does not undermine the evidence for efficacy.
Withdrawal and stopping safely
This has been one of the most active and contested areas since 2018. A 2018 review by Joseph Hayes and Sameer Jauhar found that the widely-reported Davies and Read withdrawal paper was seriously flawed and did not accurately portray the data. Meanwhile, Mark Horowitz’s influential 2019 Lancet Psychiatry piece argued for slower, hyperbolic tapering to mitigate withdrawal symptoms, a view further developed in a 2021 blog drawing on lived-experience expertise.
The most definitive recent evidence comes from a 2025 network meta-analysis of 76 trials, which found that slow tapering (over at least four weeks) combined with psychological support was the most effective strategy for preventing relapse when stopping antidepressants. Abrupt discontinuation and fast tapering substantially increased relapse risk and should be avoided.
A 2024 systematic review and meta-analysis put numbers on discontinuation symptoms: 8–14% of patients will experience them, with around 2% experiencing severe symptoms. A 2024 survey of 708 people with lived experience reinforced the case for much better withdrawal support in clinical settings.
Long-term use and relapse
The ANTLER trial (2021) found that patients in primary care who discontinued antidepressants were at substantially higher risk of relapse than those who continued, an important counterweight to the narrative that long-term use is invariably problematic. A 2024 secondary analysis of ANTLER data identified specific clinical factors associated with relapse risk, helping clinicians identify who most needs to continue medication.
Antidepressants and suicide risk
The relationship between antidepressants and suicide is genuinely complex. A 2018 study in the British Journal of Psychiatry found SSRIs remain safe and effective and do not increase suicidality. However, a 2019 meta-analysis found a small but significant increase in suicide risk in adults with major depression. A 2025 network meta-analysis offered further nuance: SSRIs may reduce suicidal ideation in the short term, but this effect is not sustained beyond eight weeks of treatment. Antidepressants are not a complete solution to suicidal risk and must be accompanied by appropriate monitoring and support.
Young people
The evidence base for antidepressants in children and adolescents is thinner and more contested. A 2021 Lancet Psychiatry review found meaningful but uncertain benefits alongside real safety concerns. A 2025 analysis argued that treatment decisions should be guided by patient values as much as evidence, given that trials of antidepressants and psychotherapy in young people are often too methodologically different to compare meaningfully.
Predicting who will respond
A 2024 study explored whether artificial intelligence could predict antidepressant response, promising but preliminary results. A 2018 systematic meta-review of predictors of outcome (clinical, demographic, genetic, and neuroimaging) found we are not yet at the point of reliable personalised prescribing, but the field is moving in that direction.
Antidepressants in pregnancy
A 2024 umbrella review found no convincing evidence of adverse health outcomes for babies born to mothers who took psychotropic medication during pregnancy, but the authors were clear that this reflects limited and uncertain evidence rather than established safety. Decision-making in this area remains difficult and should always be individualised.
Explore related topics
Related conditions
Depression,
Anxiety,
OCD,
PTSD,
Eating disorders,
Bipolar disorder
Related treatments
CBT,
Psychotherapy,
Mindfulness,
Self-management
