Omega-3 fatty acids are fats the body cannot make, found in oily fish and, in a shorter form, in some plant oils. EPA and DHA, the marine forms, are what supplements and prescription products contain. Populations that eat more fish have less heart disease, and that observation launched decades of trials.

The result is a literature that genuinely disagrees with itself, and this guide tries to show why rather than pick a side. The dose, the formulation, the comparator oil and the patients differed between the trials that found a benefit and the trials that did not, and each of those differences has its defenders.

For a reader deciding about a shelf supplement, the most relevant trials are the ones that used shelf-like doses in ordinary people. Those were null.

EPA and DHA are incorporated into cell membranes, where they change membrane behavior and are converted into signaling molecules that are less inflammatory than those made from omega-6 fats. At high doses they lower blood triglycerides substantially, which is the effect the prescription products are licensed for.

The proposed routes to fewer heart attacks include that triglyceride lowering, reduced inflammation, stabilization of plaques and effects on heart rhythm. The rhythm story cuts both ways: an anti-arrhythmic effect was long proposed, but the trials with the highest doses found more atrial fibrillation, not less.

Ordinary doses in ordinary people: null. VITAL gave nearly 26,000 American adults a gram a day of marine omega-3 for a median of five years and found no reduction in major cardiovascular events or cancer. A 2020 Cochrane review of 86 trials in over 160,000 people found little or no effect of increasing long-chain omega-3 on all-cause mortality, cardiovascular events or stroke, with high-certainty evidence, and a small reduction in cardiovascular mortality of moderate certainty.

High-dose prescription EPA in high-risk patients: positive, then contested. REDUCE-IT randomized 8,179 patients with raised triglycerides on statins to a high dose of icosapent ethyl, purified EPA, or placebo. Ischemic events fell by a quarter and cardiovascular death fell too. Two years later STRENGTH tested a high dose of combined EPA and DHA against corn oil in 13,078 similar patients and was stopped early for no benefit at all. The main difference is the comparator: REDUCE-IT's placebo was mineral oil, which may have worsened outcomes in the control arm, and STRENGTH's was corn oil, which may not. Nobody has run the trial that settles it.

Meta-analyses. A 2019 pooled analysis of 13 trials in 127,477 people found marine omega-3 associated with modestly lower rates of heart attack, coronary death and cardiovascular death, with a dose-response, even after excluding REDUCE-IT. The Cochrane review, with more trials and stricter inclusion, is less impressed. Both are defensible readings of the same evidence.

At supplement doses omega-3 is well tolerated; fishy aftertaste and loose stools are the usual complaints. At the high doses used in the cardiovascular trials two signals appeared. Atrial fibrillation needing hospital admission was more common on icosapent ethyl than placebo in REDUCE-IT, and the same signal has appeared in other high-dose trials. Serious bleeding was numerically higher and not statistically significant.

VITAL, at a gram a day, found no excess of bleeding or other serious adverse events. The safety profile, like the efficacy profile, depends on the dose, and the shelf product and the prescription product are a fourfold step apart.

Omega-3 exists as both prescription medicine and food supplement. Icosapent ethyl and omega-3 ethyl esters are licensed in the United States, the European Union and the United Kingdom for severely raised triglycerides, and icosapent ethyl additionally for cardiovascular risk reduction in the REDUCE-IT population. Those are drug approvals, with a reviewed label.

Fish oil capsules on a shelf are food supplements. No regulator has reviewed their claims, and content and oxidation state vary by product. A supplement is not a lower dose of the licensed drug; it is a different product that nobody has had to prove anything about.