Omega-3 EPA
Eicosapentaenoic acid, a long-chain omega-3 fat
A long-chain omega-3 fat found mainly in oily fish and fish oil, where it comes with DHA. In studies on the mind it has been looked at mostly for mood. This page covers only studies in people on thinking, memory, mood, stress, sleep and energy, sorted from each study's summary.
- Found in: oily fish and fish oil
- Most tested: 2,000 mg a day
- Often taken for 12 weeks
From every paper to the studies that matter
Most omega-3 research is about the heart, blood fats or lab work, so it isn't counted here. This page keeps only studies in people that measured thinking, mood, stress, sleep, energy or quality of life, and most of those gave EPA inside fish oil with DHA.
-
9,081 papers on long-chain omega-3.Every record in PubMed, all years
-
554 are about people.The rest: 3,040 lab and animal studies, 2,359 reviews, 3,128 on chemistry, farming and more
-
27 give EPA without DHA.448 more give it with DHA in fish, krill or cod liver oil, or in a blend
-
21 are randomised against a placebo.Randomised controlled trials In a randomised controlled trial, people are put into groups by chance, one taking the treatment and one a placebo or other comparison. Chance spreads differences such as age and health evenly, so a difference between the groups is more likely to be down to the treatment., the strongest kind of study
Bar lengths use a square-root scale so the small numbers stay visible.
How much research there is
- 27 studies of EPA on thinking, mood, sleep and energy
- 21 of them randomised against a placebo A placebo-controlled trial compares a real treatment with a dummy that looks the same, the placebo. People often feel better when they expect to, so comparing against a placebo shows how much of the change comes from the treatment itself.
- 1,593 people taking part, where studies say
- 487 systematic reviews A systematic review searches for every study on a question in a set, written-down way, judges the quality of each and sums up what they show together. Many include a meta-analysis. The Atlas lists these as pooled reviews. pooling the trials
Studies in people, by year published
EPA given without DHA.
What the studies looked at
Each study can measure several things. Choose a topic to see its studies in the list below.
Studies in people with a diagnosed condition are counted here and listed with their population. We describe what they measured, not what they might treat. Most studies also report safety; see below.
What was tested, and how much
Doses are the total omega-3 a day as each summary gives it, so in fish oil studies they count DHA as well as EPA. Studies of EPA without DHA gave 1,000 or 2,000 mg a day most often, for 8 to 12 weeks.
Daily amount, one dot per study
Studies per extract
Named extracts are the maker's brand. "Not named" means the study didn't say.
All studies in people
Filter the list, then open a study on PubMed. Each row comes from the study's published summary, and fish oil studies are marked as a blend.
| Study | Who took part | People | Extract and daily dose | Length |
|---|---|---|---|---|
| Wang 2024RCT, placebo | Adults with episodic migraine | 70 | Not named, 1,800 mg | 12 weeks |
| Lin 2024RCT, placebo | Maintenance hemodialysis patients with pruritus | 60 | Not named, 2,000 mg | 13 weeks |
| Savard 2023RCT, placebo | Men with prostate cancer after prostatectomy | – | Not named | – |
| Mischoulon 2022RCT, placebo | Overweight adults with depression and inflammation | 61 | Not named | 12 weeks |
| Lamon-Fava 2021RCT, placebo | Adults with depression and chronic inflammation | 61 | Not named | 12 weeks |
| Carney 2019RCT, placebo | Depressed patients with or at risk of heart disease | 144 | Not named, 2,000 mg | 10 weeks |
| Chang 2019RCT, placebo | Youth with ADHD | 92 | Not named, 1,200 mg | 12 weeks |
| Bentsen 2018RCT, placebo | Young adults with acute schizophrenia on antipsychotics | 53 | Not named, 2,000 mg | 16 weeks |
| Huerta 2015RCT, placebo | Overweight or obese women on energy-restricted diet | 73 | Not named, 1,300 mg | 10 weeks |
| Mocking 2012RCT, placebo | Diabetes patients with comorbid major depression | 25 | Not named | 12 weeks |
| Reddy 2011Single-arm or open-label | Early-course schizophrenia patients | 27 | Ethyl-EPA, 2,000 mg | 24 weeks |
| Mantovani 2010RCT, active or other control | Cancer cachexia patients | 332 | Not named | 17 weeks |
| Gustafsson 2010RCT, placebo | Children with ADHD aged 7-12 | 92 | Not named, 500 mg | 15 weeks |
| Jazayeri 2010RCT, active or other control | Adults with major depressive disorder | 42 | Not named, 1,000 mg | 8 weeks |
| Bot 2010RCT, placebo | Diabetes patients with major depression | 25 | Ethyl-EPA, 1,000 mg | 12 weeks |
| Mischoulon 2009RCT, placebo | Adults with major depressive disorder | 57 | Ethyl-EPA, 1,000 mg | 8 weeks |
| Jazayeri 2008RCT, active or other control | Adults with major depression | 60 | Not named, 1,000 mg | 8 weeks |
| Frangou 2007RCT, placebo | Women with bipolar disorder | 14 | Not named, 2,000 mg | 12 weeks |
| Frangou 2006RCT, placebo | Bipolar depression | 75 | Not named | 12 weeks |
| Keck 2006RCT, placebo | Bipolar depression or rapid cycling | – | Not named, 6,000 mg | 17 weeks |
| Osher 2005Single-arm or open-label | Bipolar depression outpatients | 12 | Not named, 1,500 mg | 26 weeks |
| Fux 2004Crossover, placebo | OCD patients on stable SSRI | 11 | Not named, 2,000 mg | 6 weeks |
| Boston 2004Single-arm or open-label | Alzheimer's disease patients | – | Not named | 12 weeks |
| Zanarini 2003RCT, placebo | Women with borderline personality disorder | 30 | Not named, 1,000 mg | 8 weeks |
| Peet 2002RCT, placebo | Patients with persistent depression on antidepressants | 70 | Not named, 1,000 mg | 12 weeks |
| Nemets 2002RCT, placebo | Recurrent unipolar depression on antidepressants | 20 | Not named | 4 weeks |
| Fenton 2001RCT, placebo | Schizophrenia or schizoaffective patients on neuroleptics | 87 | Not named, 3,000 mg | 16 weeks |
What's known about safety
-
In trials
Side effects are usually mild, such as an unpleasant taste, bad breath, headache, heartburn, nausea or diarrhoea. A review of 23 trials found no more side effects with omega-3 than with placebo.
NCCIH, updated Nov 2024; Bafkar 2024, review of 23 trials
-
In the UK
Fish oil, the usual source of EPA, was part of the European diet long before 1997, so it isn't a novel food and is sold as a food supplement. A purified form, icosapent ethyl, is a prescription-only medicine. EFSA set no upper limit but found up to 5 g a day of EPA and DHA together raised no safety concerns in adults.
EU novel food status consultation on fish oil, 2023; Vazkepa product information (emc); EFSA 2012
-
Check first
The prescription form of EPA warns of a higher bleeding risk with blood thinners or antiplatelet medicines such as aspirin, and a higher risk of an irregular heartbeat (atrial fibrillation). It also advises care with a fish or shellfish allergy and has little data in pregnancy. Speak to a pharmacist or doctor first if any of these applies to you.
Vazkepa (icosapent ethyl) product information, emc, read Oct 2026
Where these studies come from
We list every published study in people that gave EPA and measured thinking, memory, mood, stress, sleep, energy or quality of life. EPA without DHA counts as on its own, while fish, krill or cod liver oil and omega-3 not broken down count as a blend. Plant omega-3 (ALA from flaxseed), eating fish and people who chose their own supplements are left out. This first version is built from each study's summary.
-
We search
PubMed, the world's main library of medical research, for every paper on L-theanine.
-
We keep studies in people
Lab and animal work is counted but left out of the list, because it doesn't tell you what happens in people.
-
We note the basics
Who took part, how many, how much and for how long, from each study's published summary. A second check corrects what the first read gets wrong.
-
We read them in full
Results appear once we've read the whole paper and checked every number. The date at the top shows when this entry was last updated.
Download the study list
Every study in people as a spreadsheet, 27 rows, free to reuse with a link back.
Versions
Each update to the data gets a new version number, so you can tell what changed and when.
See every version
1.0 · 24 Oct 2026. First release. Every published study giving EPA to people that measured thinking, memory, mood, stress, sleep, energy or quality of life is listed from its summary. None is read in full yet.
