The dose of omega-3 that lowered heart attacks in the best trial we have was about four grams a day. The dose in the softgel on most bathroom shelves is around 300 milligrams. That is not a rounding error. It is the difference between a drug and a garnish, and it explains why fish oil looks like a miracle in one headline and a waste of money in the next.
Omega-3s are among the most studied supplements on earth, and the research is genuinely messy. But the mess is mostly a dosing and population problem, not a mystery. Once you sort the trials by how much people actually took and who they were, a coherent picture appears.
What these fats actually do inside you
EPA and DHA are long-chain fatty acids your body cannot manufacture in any meaningful quantity. You get them from oily fish, or from fish that ate algae, or from a capsule. Once absorbed, they get physically built into the phospholipid membranes of your cells, including the ones lining your arteries and the ones in your heart’s electrical system.
That structural role matters more than the vague “anti-inflammatory” label suggests. Membranes rich in EPA and DHA behave differently. They change how cells signal, how stable heart rhythm is, and how readily the body switches off inflammation once a threat has passed. EPA in particular is the raw material for a class of molecules called resolvins, which actively tell inflammation to stand down rather than just blocking it from starting. And at gram-level doses, these fats lower blood triglycerides substantially, partly by reducing how much fat the liver packages and ships out.
The catch is that all of this scales with how much you have on board. A trickle of EPA does not meaningfully change your membranes. A flood does.
The trial that made cardiologists pay attention
In REDUCE-IT, patients already on statins, with high triglycerides and either established heart disease or diabetes plus risk factors, took four grams a day of icosapent ethyl, a purified EPA. Major cardiovascular events dropped by about 25 percent over roughly five years NEJM, 2019. That is a large effect for something you swallow, and it landed in a high-risk group where events were common enough to measure.
Now the deflating part. In VITAL, a much larger trial of generally healthy adults taking one gram a day of ordinary fish oil, the composite of major cardiovascular events did not budge NEJM, 2019. Same family of molecules, opposite conclusion.
The two results are not contradictory. They are telling you the same thing from different angles: the dose in VITAL was a quarter of REDUCE-IT’s, and the people were at far lower risk, so there was less to prevent. Give a small dose to low-risk people and you should expect a small, hard-to-detect effect. That is exactly what happened.
One honest caveat keeps REDUCE-IT from being the final word. Its placebo was mineral oil, which appeared to slightly raise LDL and inflammatory markers in the control group. Some of the apparent benefit may reflect the placebo arm getting worse rather than the treatment arm getting better. Reasonable cardiologists still argue about how much this matters. It does not erase the result, but it shaves the edges off.
Reading your own level instead of guessing
You can measure this directly. The omega-3 index reports EPA plus DHA as a percentage of the fatty acids in your red blood cell membranes, which reflects your intake over the previous few months rather than what you ate yesterday.
| Omega-3 index | What it reflects |
|---|---|
| Below 4% | Low tissue levels, associated with higher cardiovascular risk in population data |
| 4% to 8% | Intermediate, where most Western adults land |
| 8% and above | The range seen in populations with high fish intake and lower cardiac event rates |
Most people eating a typical Western diet sit around 4 to 5 percent. Getting to 8 usually takes either real fish several times a week or a supplement in the one-to-two-gram range of actual EPA plus DHA, not one gram of “fish oil” that turns out to be 300 milligrams of the fats that count. Read the back of the bottle, not the front.
What to do on Monday
Start with food. Two or three servings a week of salmon, sardines, mackerel, herring, or trout will move most people toward the middle-to-upper range without a single capsule. Sardines are cheap, shelf-stable, and unfairly maligned.
If you supplement, look at the EPA and DHA numbers on the label and add them up, ignoring the total fish oil figure. For general topping-up, one to two grams combined is reasonable. The four-gram prescription dose is a medical decision for people with high triglycerides and cardiovascular risk, worth an actual conversation with your doctor rather than a self-experiment, because at that dose there is a modest increase in atrial fibrillation and a small bleeding risk.
Take it with a meal that contains fat, since these are fat-soluble and absorb far better alongside other fat than on an empty stomach. If your capsules taste fishy when you burp, they may be oxidizing. Keep them in the fridge and buy from a brand that reports oxidation values. Rancid fish oil is worse than none.
If you care about metabolic risk more broadly, omega-3s fit alongside the fundamentals rather than replacing them, the same way a good blood sugar strategy does more heavy lifting than any single pill.
Where the evidence genuinely thins out
The cardiovascular story is strongest in high-risk people. The story for cognition, mood, and general longevity in healthy adults is far weaker, built more on observational associations and mechanism than on convincing randomized endpoints. Trials of fish oil for preventing dementia have been largely disappointing. The DHA-for-brain-health marketing runs well ahead of the trial data.
So here is the reasonable position. If you eat oily fish a few times a week, you are probably fine and do not need a capsule. If you do not eat fish and your omega-3 index is low, a modest supplement is a cheap, low-risk way to fix a real deficit. And if you have high triglycerides and heart disease, the four-gram conversation is worth having with someone who can weigh the arrhythmia risk against the benefit. Everything in between is marketing dressed as science.