Omega-3 (EPA/DHA, Fish Oil): It Really Lowers Triglycerides, but the Heart Story Is Still Contested

Fish oil is one of the most widely taken supplements among people over 40, often because they believe it prevents heart attacks, nourishes the brain, and reduces inflammation. But when you lay the large trials side by side, the real picture is far more complex. Some things are clearly true, some come out as zero, and some even carry harm.

The long-chain omega-3s that matter for health are EPA and DHA, found mostly in oily fish such as salmon, mackerel, and sardines. The body makes very little of them from plant fatty acids (ALA), so they have to come from food or supplements. This article separates what the evidence supports strongly from what is still being debated.

A Three-Line Summary

1. The only effect proven to be causal is lowering triglycerides, by 20 to 30 percent at a dose of 4 grams per day, with bigger drops the higher the starting level.
2. For preventing heart attacks, ordinary fish oil shows no effect in high-quality evidence. The single positive trial, REDUCE-IT, is questioned over its placebo and is contradicted by STRENGTH.
3. High-dose omega-3, from 4 grams upward, raises the risk of atrial fibrillation (AFib) in a confirmed way, while effects on the brain and depression are weak to nonexistent.

Topic 1: Lowering Triglycerides Is the Most Solid Effect

If there is one thing omega-3 reliably does, it is lowering triglycerides, a type of fat in the blood, and this comes from randomized trials (RCTs), so it is a causal relationship, not just a correlation.

At a combined EPA and DHA dose of 4 grams per day, triglycerides drop by roughly 20 to 30 percent in people with moderately high levels (200 to 499 mg/dL), and the higher the starting level, the bigger the drop. In people with very high levels of 500 and above, the MARINE trial found that 4 grams of EPA lowered them by as much as 33 percent. This direction is confirmed by the 2019 AHA Science Advisory pooling 17 RCTs.

There is a caveat about LDL: DHA tends to raise LDL while EPA is neutral toward it. This direction is confirmed by several sources, while the exact figure of about a 7 mg/dL increase comes from a single meta-analysis, so it remains only a direction.

Topic 2: Preventing Heart Disease, the Overall Result Is Zero, and Here Is Why It Is Debated

This is the heart of all the confusion. The large trials point in opposite directions.

Trial	Formula/dose	placebo	Main result (MACE)
REDUCE-IT (positive)	EPA only, 4 g	mineral oil	HR 0.75, a 25 percent reduction
STRENGTH (zero, stopped early)	EPA+DHA, 4 g	corn oil	HR 0.99
VITAL (zero)	1 g	placebo	HR 0.92
ASCEND (zero)	1 g	olive oil	RR 0.97

The picture from large pooled analyses is no effect. Aung 2018 (n=77,917) found a major vascular event RR of 0.97, and Cochrane 2018 to 2020 (n=162,796) concluded there was little or no effect with high confidence.

So why is REDUCE-IT the only positive trial? There are several hypotheses, and each has two sides:

• The placebo was an unstable mineral oil. In the REDUCE-IT placebo group, LDL rose 10 percent and CRP rose 22 percent. If the placebo makes people worse, the difference looks better than it really is. STRENGTH, which used the more stable corn oil and came out zero, supports this concern. The counterargument is that an LDL difference of only about 5 mg/dL cannot explain the full 25 percent effect. This question is not settled, because no EPA-only RCT with a genuinely stable placebo exists for comparison.
• EPA only versus EPA+DHA. The positive trials used EPA only (REDUCE-IT, JELIS), while the EPA+DHA trials came out zero. But the idea that DHA causes harm is still mechanistic and observational, and not clear.
• Dose. The 1-gram trials all came out zero, but STRENGTH used 4 grams and was still zero, so high dose alone is not the answer.

Two claims have already been refuted. First, that high blood EPA levels drive the benefit: the prospective analysis of STRENGTH found that people whose EPA rose the most, matching REDUCE-IT levels, got no benefit. Second, that triglyceride lowering is the mechanism: both trials lowered triglycerides equally, by about 18 percent, yet the heart results differed.

Topic 3: Atrial Fibrillation (AFib), the Harm People Overlook

This is the downside with strong evidence. High-dose omega-3 raises the risk of atrial fibrillation, confirmed by two or more independent meta-analyses.

• Gencer 2021 (Circulation, n=81,210) found HR 1.25, in a dose-response pattern: above 1 gram per day HR 1.49, versus HR 1.12 at 1 gram or less.
• Lombardi 2021 (EHJ-CVP) found IRR 1.37.
• Per-trial reports show AFib HR 1.50 in REDUCE-IT and 1.69 in STRENGTH, while low-dose VITAL at 840 mg was not significant.

The relative risk is real, but the absolute risk is around 1 percentage point, and the FDA label for Vascepa carries a clear AFib warning. Anyone taking high-dose fish oil should know this.

Topic 4: Brain, Memory, Depression, Weak to Zero

• Memory and dementia in healthy older adults: zero. Cochrane found no difference in MMSE, and AREDS2 (n=3,073) agreed.
• Depression: Cochrane 2021 found a small effect (SMD -0.40, about 2.5 HDRS points, below the 3-point clinical significance threshold), with very low confidence, and any signal came from EPA, not DHA.
• DHA in pregnancy making children smarter is not supported by evidence. Cochrane 2018 found neurodevelopment barely differed. The real benefit of DHA in pregnancy is reducing preterm birth, not IQ. Do not confuse the two.

Topic 5: Eating Fish Versus Taking Supplements, Not the Same

Observational data find that eating more fish is associated with lower death from heart disease (Jayedi 2021, pooling about 2 million people, CVD-mortality RR 0.91), with a stronger benefit in people who already have heart disease (PURE 2021, the vascular-disease group HR 0.84).

But this benefit is a correlation, with a healthy-user bias, meaning fish eaters tend to take better care of themselves in other ways too. And crucially, supplement RCTs cannot reproduce it. The AHA recommendation is to eat oily fish 1 to 2 times per week (about 250 mg EPA+DHA per day), eating real food, not assuming a pill gives the same result.

Topic 6: Supplement Quality, Rancidity, and Contaminants

• Rancidity (oxidation): some supplements exceed oxidation thresholds, which is a real problem, but its scale is still debated. A 2015 New Zealand study found over 83 percent exceeded limits, a 2020 UAE study found 41 percent, and industry-funded work found most passed. The numbers vary with the funding source.
• Contaminants: large fish accumulate mercury and PCBs, so limit shark, swordfish, king mackerel, and tilefish, especially in pregnancy. Supplements tend to be lower in contaminants, and overall the benefit of eating fish still outweighs the risk.
• Omega-3 Index is the share of EPA+DHA in red blood cells, with a target of around 8 percent, and it is associated with lower sudden cardiac death. But it is a correlation and a marker, not a cause, because pushing the value up with supplements does not reliably reduce events. It may simply mark an overall fish-eating, healthy lifestyle.

Points to Watch: Beliefs the Evidence Does Not Support

“Fish oil prevents heart attacks for everyone”

Not true for ordinary fish oil. Cochrane and Aung, high-quality evidence, show zero. Only high-dose prescription EPA is positive, and that is still debated.

“Taking fish oil equals eating fish”

No. The benefit of real fish is not reproduced by a pill.

“Lower triglycerides with omega-3 and you will protect your heart”

Refuted. STRENGTH lowered triglycerides as much as REDUCE-IT but came out zero.

“The higher the blood EPA, the better”

Refuted by the STRENGTH analysis: those with the highest EPA got no benefit.

“High-dose fish oil has no harm”

Not true. Doses from 4 grams raise the risk of AFib (REDUCE-IT HR 1.50, STRENGTH HR 1.69).

Still Unknown, Still Unsettled

• Whether REDUCE-IT would still be positive with a genuinely stable placebo cannot be answered, because no such RCT exists. This knot is not resolved.
• The AFib risk at low doses of 1 gram or less is still conflicting: low-dose VITAL was zero, while Lombardi suggests risk at every dose.
• The best EPA-to-DHA ratio for real outcomes has no answer yet.

A Small Step You Can Take

If your goal is heart health, eating oily fish 1 to 2 times a week is reasonable and fits the AHA recommendation, but do not expect a fish oil pill to prevent heart attacks. If your triglycerides are high and you are considering a high dose, talk to your doctor first, because doses from 4 grams carry an AFib warning. This is eating with understanding, not eating to follow a trend.