For adults with unipolar depression: what a ranking built from continued-treatment trials can show, and where its uncertainty begins.
The review by van Haaren and colleagues included 115 double-blind randomised trials of 21 antidepressants in 30,428 adults with unipolar major depressive disorder, treated for more than 12 weeks. In the main head-to-head analysis — 35 continuation and continuation-extension trials, 9,888 people — escitalopram, duloxetine and agomelatine showed higher odds of a sustained response than fluoxetine, which the authors used as the reference. The review by van Haaren, van Gaalen, Ostinelli and colleagues was published online on 10 November 2026 in The Lancet Psychiatry.
This article is for information only and does not replace a consultation with a doctor. It is not a treatment plan and does not tell you which tablet to take or to stop.
How they ranked the drugs
The authors searched trial registers, bibliographic databases and regulator sites from inception to 27 March 2026. They included double-blind randomised trials that compared any of 21 antidepressants with each other or with placebo for long-term treatment of adult unipolar major depression. The primary analysis used head-to-head continuation and continuation-extension studies; enrichment-design studies were added in sensitivity analyses through a chain-of-evidence approach. They analysed the trials as a network, so a drug could be ranked even when it had never sat in the same trial as every other drug. The primary outcomes were a sustained response and all-cause discontinuation — how often people stopped for any reason. Ranked forest plots used fluoxetine as the reference. Mean age in the full set was 45.3 years (SD 8.5); 20,066 of 30,428 participants (66%) were female. Race and ethnicity were not extracted. No people with lived experience took part in the design or interpretation.
What the pooled numbers show
In the 35 head-to-head continuation trials:
Sustained response versus fluoxetine: escitalopram odds ratio 1.62 (95% CI 1.21 to 2.18); duloxetine 1.57 (1.08 to 2.26); agomelatine 1.40 (1.06 to 1.85).
All-cause discontinuation versus fluoxetine: escitalopram 0.60 (0.43 to 0.83); duloxetine 0.64 (0.44 to 0.95) — fewer people stopped.
Bupropion and reboxetine showed lower sustained response rates against four other antidepressants in the pairwise league table; reboxetine and milnacipran showed higher discontinuation against six and four other antidepressants respectively.
Sensitivity analyses that added enrichment-design trials, in which only people who had already responded were randomised, pointed in the same direction.
This is not an acute-treatment ranking
The best-known comparison of the same 21 drugs is the 2018 Lancet network meta-analysis by Cipriani, Furukawa, Salanti and colleagues: 522 trials, 116,477 adults, acute treatment. In that table every drug beat placebo, with odds ratios from 2.13 for amitriptyline to 1.37 for reboxetine, and fluoxetine sat among the less efficacious drugs in head-to-head trials. Several of the same authors, including Cipriani, Furukawa and Ruhé, are on the new paper. The new ranking answers a different question: among adults who stay on treatment past 12 weeks, how often a response held, and how often people stopped. The acute-treatment and long-term reviews measure different periods and outcomes, so their league tables are not interchangeable.
Where the evidence stops
Nineteen of the 35 trials in the primary analysis (54%) were at high risk of bias on the Cochrane RoB 2 tool, and the authors write that this, plus the certainty of the network results, may limit how far the ranking can be trusted. More head-to-head long-term trials of licensed antidepressants are still needed. The published abstract does not report suicide, withdrawal symptoms, sexual function, weight or everyday functioning; acceptability here is only all-cause discontinuation. The analysis is at study level, not person level, so it does not match a drug to your history. It does not say that a person who is well on another antidepressant should switch. Funding for the review: none. Several authors report speaking or consultancy fees from companies outside this work.
What to discuss with your doctor
If you are already taking an antidepressant and the first months have passed, the useful question is not “which drug won a league table” but whether your own response has held, what you are stopping for, and whether a change is worth the disruption. That conversation belongs with your doctor. This review does not give doses, schedules or a reason to start, stop or switch a medicine on your own.
Sources
van Haaren PCF, van Gaalen JM, Ostinelli EG, et al. Comparative efficacy and acceptability of antidepressants in the long-term treatment of major depressive disorder: a systematic review and network meta-analysis. Lancet Psychiatry, published online 10 November 2026. PubMed 43008763, doi:10.1016/S2215-0366(26)00326-300326-3)
Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. Lancet, 2018;391:1357–1366. PubMed 29477251, doi:10.1016/S0140-6736(17)32802-732802-7)
What this piece relies on
Cards
Articles in this section are educational and are not medical advice, a diagnosis, or a prescription. Consult a qualified professional before acting on anything you read here.
Read next
ResearchMedicine and treatmentA tau antibody missed its main Alzheimer's score
For anyone following Alzheimer's research: what a phase 2 trial found on the score clinicians use to track daily function, and which questions that result leaves open.
ResearchMedicine and treatmentEarly TB treatment and an esophageal cancer drug — plus 2 stories
4 stories this week: severe pneumonia in infants with HIV, esophageal cancer after a first treatment line, blood pressure and memory in type 2 diabetes, and a heart-rhythm ablation. Each note names who was studied and where the result stops.
Digest · 4 stories
ResearchMedicine and treatmentThe busiest minute of the day lined up with fewer deaths
For a reader who already moves on ordinary days and still cannot tell which part of that day is worth a conversation with a doctor. Inside is a comparison of people as they already were, not a plan for the next week.
