Omega-3 (EPA/DHA)

If you eat oily fish twice a week, you probably don't need fish oil. If you don't, a modest daily dose is a reasonable way to fill the gap, but the evidence that capsules prevent heart attacks or slow aging is much thinner than the marketing. Higher doses also bring a small rise in an irregular heart rhythm called atrial fibrillation.

Eat oily fish twice a week. If you don't, about 1–1.5 g a day of combined EPA+DHA, from a third-party-tested product taken with food, is a sensible personal choice, though no heart guideline recommends it. Going up to 2 g is a deliberate trade, because above about 1.5 g a day the rhythm risk becomes clearer; the 4 g doses of the heart trials are a decision for your doctor.

Fish-oil capsules barely move heart attacks

EPA (eicosapentaenoic acid) and DHA (docosahexaenoic acid) are the two long-chain omega-3 fats in oily fish. The plant omega-3 in flax, chia and walnuts converts to them poorly, typically under 10% and with almost no DHA, so it is not a substitute.

The largest review pooled 86 randomized trials and about 163,000 people. Taking more omega-3 had little or no effect on deaths from any cause, and heart events were about 4% lower, a difference that could be chance. Triglycerides fell reliably, by about 15%; "good" HDL cholesterol barely changed.[1] A more optimistic pooling of 13 trials found about 8% fewer heart attacks and coronary deaths, but no effect on stroke.[2]

The big trials of ordinary doses point the same way:

  • About 26,000 adults without heart disease on 840 mg a day saw no clear drop in their main measure of heart events. Fewer heart attacks and a benefit in people who ate little fish were secondary findings, not the main result.[3]
  • About 15,500 people with diabetes on 1 g a day had serious vascular events at almost the same rate as placebo (8.9% vs 9.2%).[4]
  • About 1,000 older heart-attack survivors on 1.8 g a day saw no benefit, and slightly more new atrial fibrillation that could have been chance.[5]

Heart guidelines have drawn the conclusion. The 2026 US cholesterol guideline rates dietary supplements as "no benefit" for lowering cholesterol or triglycerides in people with abnormal blood lipids. It notes that non-prescription fish oil has not shown clinical benefit for high triglycerides or heart disease, and lists a rise in LDL ("bad") cholesterol and atrial fibrillation as possible harms.[6] Eating oily fish regularly is the well-supported option. A modest supplement if you don't is a personal choice, not a guideline recommendation. For fish, fat types and the seed-oil debate, see Dietary fats.

The one high-dose success is disputed

The only trial that clearly worked used a prescription drug, not a supplement. Over about five years, 8,179 statin-treated adults with high triglycerides took 4 g a day of icosapent ethyl, a purified form of EPA, and had 25% fewer major heart events.[7] A similar trial of 4 g a day of an EPA+DHA mix found nothing.[8]

Neither explanation for the gap is settled. Perhaps pure EPA works and the mix doesn't, but pooling 16 trials, the difference between purified EPA and EPA+DHA mixes was not statistically significant, and most of the EPA trials were open-label studies from Japan without a placebo.[9] Or the placebo flattered the drug: its dummy capsules held mineral oil, which raised LDL cholesterol and an inflammation marker in the comparison group.[10] The US Food and Drug Administration (FDA) estimated that this could explain at most about 3% of the benefit,[11] while the trial's own biomarker analysts called the impact uncertain. A 2024 Japanese trial without mineral oil, of 1.8 g a day of purified EPA in 2,506 statin-treated people with coronary disease and a low blood ratio of EPA to an omega-6 fat, found a similar-sized benefit that just missed statistical significance, in patients who knew which group they were in; new atrial fibrillation roughly doubled (3.1% against 1.6%).[12]

So the only omega-3 with guideline support for preventing heart attacks is prescription icosapent ethyl at 4 g a day. It is for people already on a statin who are at high heart risk and whose triglycerides stay raised. The US guideline says it "may be reasonable";[13] the European one says it "should be considered".[14] That is a decision for your doctor.

Higher blood levels go with longer lives, but that isn't proof

People with more omega-3 in their blood tend to live longer, a pattern that is consistent across large cohorts. Across 17 long-running cohorts of about 42,000 adults, those in the top fifth for blood omega-3 had about 15–18% lower death rates than those in the bottom fifth.[15] Adding a large UK cohort, for more than 160,000 people in all, the highest blood DHA levels went with about 17% lower death rates, from heart disease, cancer and other causes.[16] Both analyses come from the same research group. In a long-running US heart study, people in the top fifth for the omega-3 index (above 6.8%) had about a third lower death rates than those in the bottom fifth (below 4.2%).[17]

Two things limit what this means. These are observational links, and people who eat fish tend to be healthier in many other ways. The randomized trials, which remove that problem, show essentially no effect on deaths. And the omega-3 index is promoted by a testing company whose founder co-authors much of this research.

A related measure is the ratio of omega-6 fats (abundant in seed oils and processed food) to omega-3 fats in the blood. In about 85,000 UK adults followed for 13 years, the highest-ratio fifth had roughly 26% higher death rates than the lowest.[18] Don't chase a target ratio, which is hard to measure; raising omega-3 is the easier lever and shifts the ratio anyway.

The aging-clock trial is interesting, small and after the fact

The result that made omega-3 a "longevity supplement" came from a three-year randomized trial in adults over 70. One gram a day of algae-derived omega-3 (330 mg EPA plus 660 mg DHA) slowed several DNA-methylation "aging clocks", blood tests that estimate biological age, by the equivalent of roughly three to four months over three years.[19] Vitamin D and home exercise added to the effect on only one of the clocks, and older-generation clocks did not move.

Read it with its caveats:

  • It was a post-hoc look at a subset, 777 of the trial's 2,157 participants, all from its Swiss sites, and the main trial missed all six of its planned main outcomes, including blood pressure, fractures and cognition.
  • The effects were small. One clock that tracks the pace of aging slowed by about 1%, against 2–3% in a calorie-restriction trial.
  • The comparison was sunflower oil, not an inert pill, and the heart results were mixed. Against it, omega-3 nudged LDL cholesterol up slightly and did not reduce major heart events, though the trial was too small to test that; a supplement maker was among the trial's funders, as the paper discloses.[20]
  • Clocks are stand-ins, not outcomes. A slower clock over three years is not the same as a longer life.
  • The paper did not report an omega-3 index. Over a third of participants (38%) started with low blood EPA+DHA, and the effect looked somewhat larger in people who started lower.

The useful point is the dose: 1 g a day, far below the 4 g of the heart trials.

How it might work is plausible rather than proven. EPA and DHA build into cell membranes, damp the low-grade inflammation of aging, and shift the body's signalling fats from inflammation toward its resolution.

The brain and mood: the hype has outrun the evidence

Memory and thinking. At most a small benefit. An overview of 14 trials in about 27,000 adults aged 40 and over found a slight improvement on a standard memory-and-thinking test, with higher doses or longer use doing no better.[21] A larger analysis of 58 trials reported a bigger gain, but its certainty was low, the trials disagreed sharply, and in cognitively healthy adults and in those over 60 the effect was not significant.[22]

People at genetic risk of Alzheimer's. In 323 healthy middle-aged adults at high risk, higher red-blood-cell DHA went with slightly better thinking scores over three years, an observational link that did not differ between carriers and non-carriers of APOE-ε4, the main inherited risk gene.[23] Trials have not shown a benefit. In a two-year trial, 365 adults aged 55–80 with low DHA intake, nearly half carrying APOE-ε4, took 2 g of DHA a day or placebo. The DHA reached the brain, but thinking and brain volume did not differ from placebo; the trial was small for those outcomes and lost 38% of participants.[24] A three-year trial in 102 adults aged 75 and over also missed its main brain-scan goal, with only a hint of benefit on a secondary measure in gene carriers.[25] See Dementia prevention.

Depression. A small, uncertain effect. Across 36 trials, omega-3 lowered symptom scores slightly but did not raise the share of people who responded or recovered, and there were signs of publication bias. The best results, in a narrow subgroup of Asian patients with mild-to-moderate depression not on other medication, came from 1–1.5 g a day at an EPA:DHA ratio between 1:1 and 2:1.[26] A specialist psychiatry guideline instead favours EPA-rich products, at 1–2 g of EPA a day, with a clinician confirming the diagnosis first.[27] No single ratio is established. Small trials hint that high-dose EPA may help depressed people with raised inflammation, but in one such trial of 61 adults the difference from placebo was not statistically significant.[28] If you try it for low mood, use it alongside treatment, not instead of it.

Preventing depression. Not supported. When 18,353 adults aged 50 and over without depression took 1 g of fish oil or placebo for about five years, mood scores did not differ, and depression was about 13% more common in the fish-oil group.[29]

How much to take, and how to check

  • Oily fish twice a week or more (sardines, mackerel, herring, trout, salmon) makes a supplement unnecessary. Don't rely only on farmed salmon: eight fillets bought in US shops in 2023 held about 0.9 g EPA+DHA per 100 g, less than half the 2.0 g in older food tables. Two 150 g portions a week then supply only about 0.4 g a day.[30]
  • No fish: 1–1.5 g a day of combined EPA+DHA. One gram matches the aging-clock trial. Even this band is not completely free of rhythm risk, and in people at high heart risk going above about 1.5 g a day raises it more clearly (see below); if you have atrial fibrillation or are at high risk of it, agree the dose with your doctor.
  • Athletes: the same band. A sports-nutrition position stand found athletes often run low, and estimated that about 1.4 g a day extra lifts a typical athlete's omega-3 index to 8%. It does not set a higher dose for athletes; supplements may ease soreness and markers of muscle damage.[31]
  • High triglycerides or heart disease: higher doses only with a doctor. Prescription omega-3 at 4 g a day lowers triglycerides by 20–30%, more when they start higher.[32]
  • Vegetarians and vegans: algae oils provide DHA and increasingly EPA, the same molecules fish get by eating algae. In one industry-funded trial it raised blood-plasma levels as well as fish oil, dose for dose.[33]

The omega-3 index is the share of EPA+DHA in your red-blood-cell membranes, measured with one blood test. Under 4% is low, 4–8% intermediate, and 8–12% the usual target. Above 12%, reconsider the dose, especially before surgery or if you take blood thinners. The 8% target comes from heart research, not from a proven longevity threshold. From a five-month dosing trial, researchers estimated that about 1.2 g a day would lift a 75 kg adult from a typical 4.3% to 8%, with heavier people needing more.[34] Athletes and people who skip fish often test at 4–5%.

Safe upper limits. European food-safety authorities consider up to 5 g a day safe for healthy adults, and the FDA concluded in 2019 that supplements up to 5 g a day, as labelled, are safe; foods with added omega-3 are held to a lower 3 g a day.[35] That is a safety ceiling, not a target.

Buying oil that isn't rancid

Product quality is where omega-3 most often goes wrong. The same chemical bonds that make EPA and DHA active also make them easy to oxidise during manufacture, shipping and storage.

Half of 171 North American products exceeded at least one voluntary industry oxidation limit, with flavoured, bulk and children's products the worst.[36] In a later test of 72 US products, about two in three flavoured supplements went over the limit, against about one in eight unflavoured ones.[37]

Oxidised oil seems mainly to stop working rather than to do harm. In a small seven-week trial of 54 healthy adults, fresh fish oil lowered LDL and related particles more than a deliberately oxidised oil or a sunflower-oil control did.[38] There is no point paying for a product that may not work.

  • Look for an independent test mark, such as International Fish Oil Standards (IFOS), Labdoor, US Pharmacopeia (USP) or NSF. IFOS publishes oxidation and heavy-metal results per batch.
  • Choose the form on price. Both common forms absorb well with a meal. The natural triglyceride form has an edge without much fat; the cheaper ethyl-ester form does nearly as well with a fatty meal.
  • Refrigerate after opening and finish the bottle within a few months.
  • Trust your nose. Good oil smells nearly neutral. A strong fishy smell, or fishy burps after a dose, signals oxidation.

The heart-rhythm catch, and who should be careful

Omega-3 can nudge up atrial fibrillation, an irregular heart rhythm. The largest trial analysis, of 35 trials and co-funded by an omega-3 industry body, found no clear rise at 1.5 g a day or less, the range most supplement users take. Above 1.5 g a day in people already at high heart risk, the odds rose by about 40%, about 1 extra case per 125 people treated. Across all trials the odds were about 14% higher.[39] An earlier pooling of 7 heart trials in 81,210 patients saw a 25% rise overall, and a slight one (about 12%) even at 1 g a day or less.[40]

In the UK cohort, people in the top fifth for blood omega-3 had about 14% less atrial fibrillation than those in the bottom fifth, an observational link that does not show supplementing protects you.[41] Whether taking fish oil does the same is disputed. One analysis found healthy regular fish-oil users had about 13% more atrial fibrillation and borderline more stroke;[42] a later one, by researchers with commercial ties to omega-3 products, found no link once age was modelled more finely. Both are observational. At 4 g a day, the extra atrial fibrillation is a trade worth making only where that dose has shown benefit: prescription purified EPA in high-risk patients.

  • Atrial fibrillation. If you have it or are at high risk of it, agree any dose with your doctor.
  • Bleeding. Across trials of over 120,000 people, mostly older adults with or at risk of heart disease, in an analysis funded by the maker of icosapent ethyl, omega-3 did not significantly raise overall bleeding, bleeding strokes or gut bleeding. The exception was prescription high-dose purified EPA, which raised bleeding by about half in relative terms, about 1 extra bleed for every 166 people treated, with no clear rise in serious bleeding.[43] Stopping high doses about a week before planned surgery is a common precaution, not a tested rule.
  • Blood thinners. Across trials, extra bleeding tracked the EPA dose rather than the share of people on aspirin, a comparison between trials rather than a test in individual aspirin users, which is mildly reassuring for ordinary doses with aspirin. If you take warfarin, apixaban or a similar drug, check with your prescriber before high doses; those drugs were not analysed separately.
  • Pregnancy. DHA needs rise. A 2024 international guideline advises at least 250 mg a day of EPA+DHA for women of childbearing age, plus an extra 100–200 mg a day of DHA in pregnancy. Women with low DHA intake or blood levels should take about 600–1,000 mg a day from the second trimester (no later than about 20 weeks) until about 37 weeks, to lower the risk of preterm birth.[44] Pooled trials show about 17% lower odds of preterm birth.[45] Choose products with verified low contaminant levels.

What to skip

  • Flaxseed or chia oil as a substitute. Good food, but the body makes too little EPA and DHA from them.
  • Krill oil at premium prices. Its small absorption edge doesn't justify the cost per gram of EPA+DHA.
  • Flavoured, gummy or chewable omega-3 for adults. They have the worst freshness record and little EPA+DHA per dose.

Further reading

  • Abdelhamid AS et al. Omega-3 fatty acids for the primary and secondary prevention of cardiovascular disease (Cochrane). Cochrane Database Syst Rev 2020.[46]
  • Bhatt DL et al. Cardiovascular Risk Reduction with Icosapent Ethyl for Hypertriglyceridemia (REDUCE-IT). NEJM 2019.[47]
  • Nicholls SJ et al. Effect of High-Dose Omega-3 Fatty Acids vs Corn Oil on Major Adverse Cardiovascular Events (STRENGTH). JAMA 2020.[48]
  • Manson JE et al. Marine n-3 Fatty Acids and Prevention of Cardiovascular Disease and Cancer (VITAL). NEJM 2019.[49]
  • Harris WS et al. Blood n-3 fatty acid levels and total and cause-specific mortality from 17 prospective studies (FORCE consortium). Nature Communications 2021.[50]
  • Bischoff-Ferrari HA et al. Individual and additive effects of vitamin D, omega-3 and exercise on DNA methylation clocks (DO-HEALTH). Nature Aging 2025.[51]
  • Gencer B et al. Effect of Long-Term Marine n-3 Fatty Acids Supplementation on the Risk of Atrial Fibrillation in RCTs of Cardiovascular Outcomes. Circulation 2021.[52]
  • O'Keefe E et al. Associations Between Plasma Omega-3 and Fish Oil Use With Risk of Atrial Fibrillation in the UK Biobank. J Am Heart Assoc 2025.[53]
  • Skulas-Ray AC et al. Omega-3 Fatty Acids for the Management of Hypertriglyceridemia: A Science Advisory From the AHA. Circulation 2019.[54]
  • Jackowski SA et al. Oxidation levels of North American over-the-counter n-3 (omega-3) supplements and the influence of supplement formulation and delivery form on evaluating oxidative safety. J Nutr Sci 2015.[55]
  • Jäger R et al. International Society of Sports Nutrition Position Stand: Long-Chain Omega-3 Polyunsaturated Fatty Acids. J Int Soc Sports Nutr 2025.[56]
  • Lázaro I et al. Red blood cell ω-3 status and longitudinal cognition in individuals at risk of Alzheimer disease (ALFA+). J Nutr 2025.[57]
— § —