Omega-3s: Fish Oil, EPA vs. DHA, and What the Evidence Actually Shows

Fish oil occupies a strange position. It is among the most widely used supplements on earth, recommended for decades on the strength of population studies showing that people who eat more fish have less heart disease. And it is the subject of two large, well-run randomized trials that tested nearly the same dose of nearly the same compound and reached opposite conclusions.

That disagreement is not a footnote. It is the whole story, and understanding it tells you more about how to read evidence than almost any other example in supplement science.

---

EPA and DHA Are Not Interchangeable

Marine omega-3s come as two main long-chain fatty acids. EPAEPA (eicosapentaenoic acid) A long-chain omega-3 fatty acid found in fatty fish and fish oil, best studied for lowering blood triglycerides and reducing inflammation. is the one most studied for triglyceridetriglycerides The main form of fat carried in the blood and stored in fat tissue. High levels travel with insulin resistance and cardiovascular risk, and they respond strongly to diet, alcohol, and omega-3 dose. lowering and cardiovascular outcomes. DHADHA (docosahexaenoic acid) A long-chain omega-3 fatty acid that is the primary structural fat in the brain and retina. Found in fatty fish and algae-based supplements. is the dominant structural fat in the brain and retina, and the one that matters most for neural and visual development.

Most fish oil capsules contain both, in varying and often poorly disclosed ratios. Prescription products differ deliberately: icosapent ethylIcosapent ethyl (Vascepa) A prescription drug that is purified EPA only, with no DHA. It is the form used in the REDUCE-IT trial and is FDA-approved to reduce cardiovascular risk in specific high-risk patients. is purified EPA only, while other formulations combine the two. As you’ll see, that distinction sits at the center of the trial disagreement.

One useful habit: read the back of the bottle, not the front. A “1,000 mg fish oil” softgel typically contains 300 mg or less of actual EPA plus DHA. The rest is other fats.

---

What Omega-3s Reliably Do

Start with the claim that isn’t in dispute. Omega-3 fatty acids lower triglycerides, substantially and in a dose-dependent way. At 4 grams daily, reductions of 20–30% are routine in people with elevated levels. This is a genuine pharmacologic effect and the basis on which prescription omega-3s were first approved.

Here is the trap, and it’s the same one that has caught a dozen other interventions: lowering a risk marker and preventing the disease it marks are different claims requiring different evidence. Triglycerides track with cardiovascular risk. Whether lowering them with omega-3s prevents events is a separate question — which is exactly what the outcome trials were built to answer.

---

The Over-the-Counter Dose: Two Large Null Trials

The 1-gram capsules people actually buy have been tested properly, twice.

VITAL randomized 25,871 US adults to 1 gram daily of marine omega-3s or placebo. Major cardiovascular events occurred in 386 participants on omega-3 versus 419 on placebo — a hazard ratiohazard ratio A number comparing how fast an outcome happens in two groups. Above 1 means more risk in the treated group, below 1 means less, and exactly 1 means no difference. of 0.92 (95% CI 0.80–1.06), not significant. Invasive cancer was likewise unaffected.

There was one intriguing secondary finding: total myocardial infarction fell meaningfully (HR 0.72, 95% CI 0.59–0.90). That is a secondary endpoint in a trial whose primary endpoint was null, which means it is hypothesis-generating rather than practice-changing — but it’s real and worth naming.

ASCEND tested the same 1-gram dose in more than 15,000 people with diabetes and no known cardiovascular disease, followed for a mean of 7.4 years. Serious vascular events occurred in 8.9% of the omega-3 group — no benefit over placebo.

Two large trials, different populations, same answer. If you are taking a drugstore fish oil capsule to prevent heart disease, this is the evidence that applies to you.

---

The Prescription Dose: The Trial That Split the Field

REDUCE-IT enrolled 8,179 statin-treated patients with established cardiovascular disease or diabetes plus risk factors, all with triglycerides of 135–499 mg/dL. They received 4 grams daily of icosapent ethyl — purified EPA — or placebo, for a median 4.9 years.

The result was striking. The primary endpoint occurred in 17.2% on icosapent ethyl versus 22.0% on placebo: a hazard ratio of 0.75, a 25% relative reduction. Cardiovascular death fell from 5.2% to 4.3%. For a supplement-adjacent compound, this was a genuinely impressive drug-level result, and it led to FDA approval for cardiovascular risk reduction.

STRENGTH then tested a very similar idea: 13,078 high-risk statin-treated patients, 4 grams daily of an omega-3 carboxylic acid formulation delivering high EPA plus moderate DHA. The hazard ratio was 0.99 (95% CI 0.90–1.09). The trial was stopped early for futilityfutility A trial stopped early because the accumulating data show the treatment is very unlikely to prove beneficial, making it unethical to continue. Not the same as being stopped for harm..

Two trials. Same dose. Opposite conclusions.

---

The Mineral Oil Problem

REDUCE-IT used mineral oil as its placebo. Over the trial, the placebo group’s LDL cholesterol and C-reactive protein rose. If the comparator actively worsened the control group — possibly by interfering with statin absorption — then the gap between arms would overstate what EPA did.

A 2021 European Heart Journal analysis modeled this directly using a large cohort to mimic both trial designs. The conclusion was measured: differences in the effects of the active and comparator oils on lipids and CRPCRP (C-reactive protein) A blood marker of inflammation made by the liver. The high-sensitivity version (hs-CRP) is used for heart-risk assessment. explain part of the contrast between REDUCE-IT and STRENGTH — with the placebo accounting for roughly 7% of the difference — but not the entire gap.

So the mineral oil critique is legitimate and insufficient. It weakens REDUCE-IT without erasing it. Anyone who tells you the trial was invalidated is overstating; so is anyone who cites the 25% reduction without mentioning the comparator.

---

The Risk Side, Which Rarely Gets Mentioned

Fish oil is usually discussed as though the worst case is a wasted purchase and a fishy aftertaste. The trial data say otherwise.

A meta-analysis of seven cardiovascular outcome trials covering 81,210 participants found omega-3 supplementation associated with an increased risk of atrial fibrillationatrial fibrillation An irregular, often rapid heart rhythm originating in the upper chambers of the heart. It raises the risk of stroke and is a common reason for blood-thinning medication. — hazard ratio 1.25 (95% CI 1.07–1.46) — with the authors noting the disparity across trials appears dose-related. In REDUCE-IT itself, hospitalization for atrial fibrillation or flutter occurred in 3.1% on icosapent ethyl versus 2.1% on placebo. Serious bleeding was numerically higher as well: 2.7% versus 2.1%.

These are small absolute risks. But they matter enormously to how you weigh a supplement whose benefit in healthy people has twice failed to appear. A null benefit with a small real risk is not a neutral proposition.

---

Fish Is a Different Intervention

None of this argues against eating fish. The dietary evidence rests on a broader and more consistent body of research, and whole fish delivers protein, selenium, iodine, and vitamin D while displacing less useful foods.

This gap — food looks good, isolated nutrient doesn’t — has appeared so many times that it should now be a prior rather than a surprise. It happened with beta-carotene, with vitamin E, and largely with vitamin D. The nutrient is rarely doing the work alone.

A word on the omega-3 indexomega-3 index A blood test reporting EPA plus DHA as a percentage of the fatty acids in red blood cell membranes. It reflects intake reasonably well, but no trial has shown that treating to a target index improves health outcomes., the red-blood-cell test marketed as a way to optimize your levels: it is a reasonable measure of intake, but no trial has demonstrated that treating to a target index improves outcomes. Testing to a threshold nobody has validated is the same error the vitamin D field spent a decade making.

---

Where This Leaves Us

For a healthy adult, the case for routine fish oil is weak and getting weaker: two large trials at the over-the-counter dose found nothing, and the pooled safety data show a modest atrial fibrillation signal.

For someone with established cardiovascular disease on a statin whose triglycerides remain elevated, prescription icosapent ethyl is a legitimate conversation to have with a cardiologist — a real trial found a real benefit, tempered by a real methodological objection and a competing null trial.

And for everyone: eat the fish. That recommendation has survived everything the last decade threw at it.

---

A note on what’s still open: The EPA-versus-EPA-plus-DHA question is genuinely unresolved and matters commercially as well as clinically. A trial comparing purified EPA against a non-mineral-oil placebo in the REDUCE-IT population would settle it. None has been run, and the incentives to run one are weak now that the product is approved.