Three large randomized trials tested omega-3 fatty acids against heart disease. Each enrolled thousands of people and ran for years. They reached three different answers.
One found nothing on either of its primary endpoints. One was stopped early for lack of benefit. The third produced a 25 percent relative risk reduction in major cardiovascular events, and it is the one you have heard about.
The third trial did not test a fish oil supplement. It tested a prescription drug at 4 grams a day, double what FDA says a supplement label should recommend, in people already taking statins, against a comparator that became its own research controversy.
Understanding why those results differ tells you more about omega-3 than any summary of them does.
Three fatty acids, and only one is essential
Omega-3 is a family name, not a compound. Three members matter for human nutrition: alpha-linolenic acid, usually written ALA, which comes from plant sources, and eicosapentaenoic acid and docosahexaenoic acid, EPA and DHA, which come mainly from marine sources.
Your body can convert ALA into EPA and then into DHA. The NIH Office of Dietary Supplements states that the conversion happens primarily in the liver and is very limited, with reported rates of less than 15 percent.
That limit shapes the official numbers in a way most labels obscure. The Office of Dietary Supplements sets Adequate Intakes of 1.6 grams a day for men and 1.1 grams a day for women, in both the 19 to 50 and 51-plus bands. Every one of those figures is expressed as ALA.
The agency explains why in one sentence: “For age 1 and older, the AIs apply only to ALA because ALA is the only omega-3 that is essential.”
An Adequate Intake is the value set when the evidence is not strong enough to set a Recommended Dietary Allowance, a distinction article 322 covers.
The practical consequence: there is no federal intake value for EPA or for DHA. When a product claims to deliver a daily value of omega-3, that value was built around the plant compound, not the marine ones the heart trials tested.
Three large trials, three different answers
The one that found nothing
VITAL randomized 25,871 US adults, men 50 and older and women 55 and older, including 5,106 Black participants. None had prior cardiovascular disease or cancer. It used a two-by-two design, testing a 1 gram daily marine omega-3 capsule and 2,000 IU of vitamin D3 at the same time, over a median of 5.3 years.
Both primary endpoints came back null. Major cardiovascular events gave a hazard ratio of 0.92, with a confidence interval of 0.80 to 1.06 and P = 0.24. Invasive cancer gave 1.03, interval 0.93 to 1.13, P = 0.56.
One secondary endpoint did move. Total myocardial infarction came in at 0.72, interval 0.59 to 0.90. Stroke, cardiovascular death and death from any cause all sat on the null line. The researchers reported no excess risk of bleeding or other serious adverse events.
A secondary endpoint that moves while both primary endpoints do not is a hypothesis for the next trial, not a result.
Funding is worth stating. VITAL was supported by five NIH components, and its ClinicalTrials.gov record states that two manufacturers donated the study agents, matching placebos and packaging. The registration also notes an eligibility rule readers will recognize: to join, you had to not be taking fish oil, or be willing to give it up.
The one that is not a supplement trial
REDUCE-IT is the trial behind most of the enthusiasm, and describing it as a fish oil study is simply wrong. It enrolled 8,179 patients already on statin therapy, with fasting triglycerides between 135 and 499 mg/dL and LDL cholesterol between 41 and 100 mg/dL. Roughly 71 percent were enrolled for secondary prevention, meaning they already had cardiovascular disease.
The agent was icosapent ethyl, a prescription drug consisting of a highly purified ethyl ester of EPA, given as 2 grams twice daily for a total of 4 grams a day. Not a mixed EPA and DHA oil. Not an over-the-counter product.
Over a median 4.9 years, the primary composite endpoint occurred in 17.2 percent of the drug group against 22.0 percent of the comparator group, a hazard ratio of 0.75 with an interval of 0.68 to 0.83 and P below 0.001. Cardiovascular death ran 4.3 percent against 5.2 percent.
The trial also reported harms. Hospitalization for atrial fibrillation or flutter occurred in 3.1 percent of the drug group against 2.1 percent of the comparator group, at P = 0.004. Serious bleeding events ran 2.7 percent against 2.1 percent, at P = 0.06.
The ClinicalTrials.gov record lists a single sponsor, the pharmaceutical company that makes the drug, with no collaborators.
The one that was stopped
STRENGTH enrolled 13,078 statin-treated patients at high cardiovascular risk with high triglycerides and low HDL cholesterol, across 675 sites in 22 countries. Thirty-five percent were women and 70 percent had diabetes. It tested 4 grams a day of omega-3 carboxylic acids, a formulation containing both EPA and DHA, against corn oil described in the published report as intended to serve as an inert comparator.
The trial was halted when 1,384 patients had reached a primary endpoint, of a planned 1,600, after an interim analysis indicated a low probability of clinical benefit. The primary endpoint occurred in 12.0 percent on the omega-3 formulation against 12.2 percent on corn oil, a hazard ratio of 0.99 and P = 0.84. Gastrointestinal adverse events ran 24.7 percent against 14.7 percent.
The sponsor was a pharmaceutical company, with a major academic medical center and a contract research organization as collaborators.
And a fourth, for scale
ASCEND randomized 15,480 adults with diabetes but no established atherosclerotic disease to 1 gram a day of omega-3 fatty acids or an olive oil placebo, for a mean of 7.4 years. Serious vascular events occurred in 8.9 percent against 9.2 percent, a rate ratio of 0.97 and P = 0.55. It was funded by the British Heart Foundation and others, not by a manufacturer.
The pattern across all four is clearer than any one of them. The trials testing roughly 1 gram a day in people without established disease found nothing on their primary endpoints. The two testing 4 grams a day in statin-treated high-risk patients split, and they differed in formulation and in comparator.
Why the comparator became the story
REDUCE-IT used pharmaceutical-grade mineral oil as its comparator. STRENGTH used corn oil. That difference is doing more work than it looks.
A biomarker substudy published in Circulation in 2022 measured markers in REDUCE-IT participants at baseline, 12 months, 24 months and the end of the study. In the mineral oil group, markers associated with atherosclerosis rose over time: at 12 months, median increases of 10.9 percent for oxidized LDL cholesterol, 16.2 percent for interleukin-6, 21.9 percent for high-sensitivity C-reactive protein and 28.9 percent for interleukin-1 beta, all at P below 0.001. In the drug group the same markers barely moved.
The authors’ conclusion is the sentence that mattered. At the study’s conclusion, they wrote, the between-group treatment differences largely reflected increases in the mineral oil group. They also stated that the effect of those findings on interpreting REDUCE-IT’s overall risk reductions is uncertain.
There is a published counter-argument, and it should be reported alongside. A 2020 review in a European Heart Journal supplement examined 80 studies using mineral oil placebos and concluded that mineral oil does not meaningfully affect study conclusions at the quantities used in clinical trials.
Read that paper’s author list before weighing it. Six of its ten listed authors give the sponsor of REDUCE-IT as their affiliation. That does not make the analysis wrong, but it does mean the two positions are not independent. Neither has been withdrawn, and we are not resolving this.
One detail is worth adding, because it is easy to check and almost never mentioned. The ClinicalTrials.gov registration for REDUCE-IT lists the comparator only as a placebo, and the words mineral oil appear nowhere in that record. The comparator’s identity comes from the journal publication and the substudies that followed.
The atrial fibrillation signal, and what a 2026 analysis did to it
The NIH Office of Dietary Supplements describes trials finding that taking 4 grams a day of omega-3 supplements for several years slightly increased the risk of atrial fibrillation, an irregular heart rhythm, in people with cardiovascular disease or at high risk of it.
Two 2021 meta-analyses built that position. One, in Circulation, pooled 7 randomized trials covering 81,210 patients with a weighted average follow-up of 4.9 years. It found a hazard ratio for atrial fibrillation of 1.25, interval 1.07 to 1.46. Split by dose, trials above 1 gram a day gave 1.49 and trials at or below 1 gram gave 1.12.
The second, in Cardiovascular Drugs and Therapy, pooled 8 trials covering 83,112 participants. It found a risk ratio of 1.24, interval 1.11 to 1.38, with event rates of 4.0 percent against 3.3 percent. It also found no increase in stroke.
Then the picture changed. On July 28, 2026, Circulation: Arrhythmia and Electrophysiology published a much larger analysis: 35 randomized trials across 37 datasets, covering 114,592 people, including unpublished data the authors obtained by writing to trial investigators.
It stratified by dose and by baseline cardiovascular risk at the same time. Only one of the four resulting groups showed a statistically significant increase: people at high cardiovascular risk taking more than 1,500 mg a day, with an odds ratio of 1.43 and an absolute risk difference of 0.8 percent. The other three came in at 1.07, 1.06 and 1.03, none significant.
Report the whole result, though. Across all studies combined, the same paper still found atrial fibrillation risk increased by 14 percent. The 2026 analysis narrows where the signal lives. It does not erase it.
Now the disclosures, which are printed in the paper itself. Its first author reports a research award from the omega-3 industry’s trade association. A second author is chief medical officer of a company that sells omega-3 products. The senior author holds stock in a laboratory that sells blood fatty acid testing.
That is three industry ties on one paper, declared openly. It does not invalidate 35 trials. It is information you are entitled to have while reading the most favorable current analysis of a safety signal.
One thing the trials make plain is that atrial fibrillation is measured, not felt. In the 2026 analysis it was ascertained through electrocardiograms, medical records, insurance claims and reported adverse events. Whether a high-dose product belongs in the regimen of someone with cardiovascular disease is a decision for the clinician managing that disease, working from those records.
Bleeding risk and interactions with blood-thinning medication are covered in article 677.
What a qualified health claim means
On June 19, 2019, FDA announced new qualified health claims for EPA and DHA relating to blood pressure and hypertension. The word qualified is doing specific regulatory work in that sentence.
For an authorized health claim, FDA requires evidence meeting a standard called significant scientific agreement. FDA found that the omega-3 blood pressure evidence did not meet that standard. It did meet a lower standard, credible evidence, which permits a qualified claim only if the claim carries required qualifying language.
That required language is a single sentence: “However, FDA has concluded that the evidence is inconsistent and inconclusive.”
Each permitted claim must be followed by that sentence and by a statement of how much EPA and DHA a serving provides. Products using the claim must contain at least 0.8 grams of EPA and DHA combined per serving.
So a label may say that consuming EPA and DHA combined may help lower blood pressure and reduce the risk of hypertension. The next sentence, by rule, is FDA saying the evidence is inconsistent and inconclusive. That disclaimer is not fine print somebody added. It is the condition on which the claim is allowed to exist.
The upper limits, and the one nobody set
Three authorities have published different kinds of number here, and the Office of Dietary Supplements reports all of them.
The Institute of Medicine did not establish a Tolerable Upper Intake Level for any omega-3. There is no UL to exceed, which is unusual in this subject and is not the same as saying no amount is too much.
The European Food Safety Authority concluded that long-term consumption of EPA and DHA supplements at combined doses of up to about 5 grams a day appears to be safe. FDA concluded that dietary supplements providing no more than 5 grams a day of EPA and DHA are safe when used as recommended, and set a separate labeling rule: supplement labels should not recommend a daily intake higher than 2 grams.
Those two FDA figures answer different questions, one about safety and one about what a label may advise. Note where the trial doses fall: the 4 grams a day used in REDUCE-IT and STRENGTH sits above FDA’s labeling ceiling and below its safety ceiling. That gap is part of why a prescription product exists at all.
The agency also notes that high doses of EPA or DHA might reduce immune function and increase bleeding time.
Rancidity, mercury, and what you cannot taste
Oxidation
Omega-3 oils go rancid. How often that happens in products on shelves depends, unhelpfully, on who ran the survey.
A 2015 analysis in Scientific Reports tested 32 fish oil products on the New Zealand market. Only 3 met or exceeded their labeled EPA and DHA content, and 69 percent contained less than 67 percent of label. On oxidation, 83 percent exceeded recommended peroxide values, 25 percent exceeded anisidine thresholds and 50 percent exceeded recommended total oxidation values. Only 8 percent exceeded none of the three. Best-before date, cost and country of origin were all poor markers of quality.
A 2017 analysis in the same journal tested 47 products on the same market and found the opposite. It reported 72 percent compliance on peroxide value, 86 percent on anisidine value, 77 percent on total oxidation, and 91 percent compliance with EPA and DHA content claims. The paper says its results stand in stark contrast to the earlier New Zealand assessment.
Check the affiliations. Two of the 2017 paper’s authors list the Global Organization for EPA and DHA Omega-3s, the industry trade body whose voluntary limits are the benchmark being tested against. Others list a nutritional ingredients manufacturer and a fish oil company. A 2018 analysis in Food Chemistry of more than 1,900 globally sourced samples found 13.9 percent exceeding a peroxide value of 5; two of its authors list the same trade body.
The split is clean. Independent academic surveys find high oxidation. Industry-affiliated surveys find compliance. Both sets of numbers are published, neither has been retracted, and we are not picking one.
The only US-market survey here found something both camps can use. A 2023 study in the Journal of Dietary Supplements tested 72 marine and algal omega-3 products sampled between 2014 and 2020 and found flavoring was the dominant variable. Among flavored products, 68 percent exceeded the trade body’s total oxidation limit, against 13 percent of unflavored ones. On peroxide value it was 65 percent against 32 percent. Two of that paper’s four authors list a commercial supplement-testing company.
The takeaway there is uncomfortable but useful. You cannot detect oxidation by taste, and the flavoring that masks a fishy taste tracked with the worst oxidation numbers in the only American survey available.
Mercury and the fish comparison
The Office of Dietary Supplements states that seafood contains varying levels of methylmercury, and that omega-3 supplements have not been found to contain it because it is removed during processing and purification.
That is the strongest sourced claim available. We looked for a peer-reviewed survey of mercury or PCB content in US-market fish oil supplements and did not find one, so treat any stronger statement as unverified. Contamination testing in powders is covered in article 340.
On food rather than capsules, the American Heart Association issued a science advisory in 2018 concluding that 1 to 2 seafood meals a week should be included to reduce the risk of congestive heart failure, coronary heart disease, ischemic stroke and sudden cardiac death, especially when seafood replaces less healthy foods. A separate 2017 advisory addressed supplements. They were written as different questions.
What to check before you believe an omega-3 claim
Five questions do most of the work, and they come straight out of the trials above.
Which compound? ALA, or EPA and DHA? They are not interchangeable, conversion between them is under 15 percent, and only ALA has a federal intake value.
Which dose? One gram a day and 4 grams a day were tested in different people and produced different answers, including on atrial fibrillation.
Which comparator? Mineral oil, corn oil and olive oil are not equivalent, and in one case the comparator generated its own literature.
Which endpoint? VITAL’s myocardial infarction result was secondary. Both of its primary endpoints were null. Headlines routinely swap the two.
Who paid, and who declared what? Sponsorship appears in the trial registration and conflicts appear in a paper’s disclosure section. Both are public, and both took about a minute to look up for this article.
The single most useful thing to carry away is that omega-3 names at least three different compounds, tested at two very different doses, in populations that barely overlap. Trials that agree with each other are mostly the ones that were run the same way. When two omega-3 results conflict, the design usually explains it before the biology does.
Sources
- NIH Office of Dietary Supplements. “Omega-3 Fatty Acids – Health Professional Fact Sheet.” ods.od.nih.gov
- US Food and Drug Administration. “FDA Announces New Qualified Health Claims for EPA and DHA Omega-3 Consumption and the Risk of Hypertension and Coronary Heart Disease (June 19, 2019).” fda.gov
- ClinicalTrials.gov (NIH National Library of Medicine). “Vitamin D and Omega-3 Trial (VITAL), NCT01169259.” clinicaltrials.gov
- ClinicalTrials.gov (NIH National Library of Medicine). “A Study of AMR101 to Evaluate Its Ability to Reduce Cardiovascular Events in High-Risk Patients With Hypertriglyceridemia and on Statin (REDUCE-IT), NCT01492361.” clinicaltrials.gov
- ClinicalTrials.gov (NIH National Library of Medicine). “Outcomes Study to Assess STatin Residual Risk Reduction With EpaNova in HiGh CV Risk PatienTs With Hypertriglyceridemia (STRENGTH), NCT02104817.” clinicaltrials.gov
- PubMed. “Manson JE et al. "Marine n-3 Fatty Acids and Prevention of Cardiovascular Disease and Cancer." New England Journal of Medicine, 2019. PubMed. DOI: 10.1056/NEJMoa1811403.” doi.org
- PubMed. “Bhatt DL et al. "Cardiovascular Risk Reduction with Icosapent Ethyl for Hypertriglyceridemia." New England Journal of Medicine, 2019. PubMed. DOI: 10.1056/NEJMoa1812792.” doi.org
- PubMed. “Nicholls SJ et al. "Effect of High-Dose Omega-3 Fatty Acids vs Corn Oil on Major Adverse Cardiovascular Events in Patients at High Cardiovascular Risk: The STRENGTH Randomized Clinical Trial." JAMA, 2020. PubMed. DOI: 10.1001/jama.2020.22258.” doi.org
- PubMed. “ASCEND Study Collaborative Group. "Effects of n-3 Fatty Acid Supplements in Diabetes Mellitus." New England Journal of Medicine, 2018. PubMed. DOI: 10.1056/NEJMoa1804989.” doi.org
- PubMed. “Gencer B et al. "Effect of Long-Term Marine Omega-3 Fatty Acids Supplementation on the Risk of Atrial Fibrillation in Randomized Controlled Trials of Cardiovascular Outcomes: A Systematic Review and Meta-Analysis." Circulation, 2021. PubMed. DOI: 10.1161/CIRCULATIONAHA.121.055654.” doi.org
- PubMed. “Jia X et al. "Association Between Omega-3 Fatty Acid Treatment and Atrial Fibrillation in Cardiovascular Outcome Trials: A Systematic Review and Meta-Analysis." Cardiovascular Drugs and Therapy, 2021. PubMed. DOI: 10.1007/s10557-021-07204-z.” doi.org
- PubMed. “Abuknesha NR et al. "Effects of Omega-3 Fatty Acid Treatment on Risk for Atrial Fibrillation: An Updated Meta-Analysis of 35 Trials including 114 592 Individuals." Circulation: Arrhythmia and Electrophysiology, 2026. PubMed. DOI: 10.1161/CIRCEP.125.014785.” doi.org
- PubMed. “Ridker PM et al. "Effects of Randomized Treatment With Icosapent Ethyl and a Mineral Oil Comparator on Interleukin-1beta, Interleukin-6, C-Reactive Protein … A REDUCE-IT Biomarker Substudy." Circulation, 2022. PubMed. DOI: 10.1161/CIRCULATIONAHA.122.059410.” doi.org
- PubMed. “Olshansky B et al. "Mineral oil: safety and use as placebo in REDUCE-IT and other clinical studies." European Heart Journal Supplements, 2020. PubMed. DOI: 10.1093/eurheartj/suaa117.” doi.org
- PubMed. “Albert BB et al. "Fish oil supplements in New Zealand are highly oxidised and do not meet label content of n-3 PUFA." Scientific Reports, 2015. PubMed. DOI: 10.1038/srep07928.” doi.org
- PubMed. “Bannenberg G et al. "Omega-3 Long-Chain Polyunsaturated Fatty Acid Content and Oxidation State of Fish Oil Supplements in New Zealand." Scientific Reports, 2017. PubMed. DOI: 10.1038/s41598-017-01470-4.” doi.org
- PubMed. “Hands JM, Anderson ML, Cooperman T, Frame LA. "A Multi-Year Rancidity Analysis of 72 Marine and Microalgal Oil Omega-3 Supplements." Journal of Dietary Supplements, 2023. PubMed. DOI: 10.1080/19390211.2023.2252064.” doi.org
- PubMed. “De Boer AA et al. "Examination of marine and vegetable oil oxidation data from a multi-year, third-party database." Food Chemistry, 2018. PubMed. DOI: 10.1016/j.foodchem.2018.01.180.” doi.org
- PubMed. “Rimm EB et al. "Seafood Long-Chain n-3 Polyunsaturated Fatty Acids and Cardiovascular Disease: A Science Advisory From the American Heart Association." Circulation, 2018. PubMed. DOI: 10.1161/CIR.0000000000000574.” doi.org
- PubMed. “Siscovick DS et al. "Omega-3 Polyunsaturated Fatty Acid (Fish Oil) Supplementation and the Prevention of Clinical Cardiovascular Disease: A Science Advisory From the American Heart Association." Circulation, 2017. PubMed. DOI: 10.1161/CIR.0000000000000482.” doi.org