An Inventory of Fish Oil: Four Grams, One Gram, and a Capsule of Corn Oil

Consider what the fish-oil literature actually asks people to swallow. Four grams a day of purified icosapent ethyl — prescription-grade EPA — handed to statin-treated patients whose triglycerides were still too high. One gram a day of combined EPA and DHA, distributed to some twenty-five thousand primary-prevention volunteers. High-dose EPA+DHA carboxylic acids measured, with a straight face, against daily capsules of corn oil. Smaller capsules for aching joints and chronically painful backs. And, in the United Kingdom, a half-million-person databank quietly recording who took fish oil and whose heart later slipped into an irregular rhythm.

Out of this strange pantry come some of the most reliable numbers in supplement science — and some of its most famous disappointments.

Omega-3s are long-chain polyunsaturated fats — principally EPA and DHA from oily fish, with a plant form, ALA, in walnuts and flaxseed — that the body cannot synthesize. They build into cell membranes and serve as raw material for resolvins, protectins and maresins, the mediators that help resolve inflammation. Their single most consistent metabolic effect is a dose-dependent lowering of blood triglycerides. The question the giant trials were built to answer is whether that biochemistry translates into fewer heart attacks and longer lives. The defensible answer: sometimes, in specific people, at specific doses — and not without cost.

What the giant trials settled — and what they didn't

Three results frame the whole debate. In REDUCE-IT, high-dose purified EPA (icosapent ethyl) reduced ischemic events by 25% in statin-treated patients with elevated triglycerides — a genuine landmark. In VITAL, 1 g/day of EPA+DHA showed little overall benefit in primary prevention. In STRENGTH, high-dose EPA+DHA carboxylic acids were no better than corn oil. The meta-analyses explain the divergence: both a 2020 updated meta-analysis in Mayo Clinic Proceedings and a 2021 systematic review in EClinicalMedicine found the effect on cardiovascular outcomes depends on dose, formulation and population. Cochrane's 2020 review delivered the coldest water: supplements have little or no effect on all-cause or cardiovascular mortality.

What is not in dispute: triglycerides fall, roughly 15% at higher doses, and blood pressure and resting heart rate dip modestly. Beyond the heart, the same molecules have been pooled for chronic pain in a 2025 systematic review and meta-analysis in Frontiers in Medicine, and for rheumatoid arthritis in a 2012 meta-analysis in Archives of Medical Research — literatures a reader can inspect directly.

The complication: the rhythm problem

Doses above 1 g/day raise the risk of atrial fibrillation. In a large UK Biobank cohort, regular fish-oil use was associated with a 13% higher risk of incident atrial fibrillation in people without cardiovascular disease — though possibly with benefit once disease is established. That single sentence contains the whole tension: the same habit reads as hazard in the healthy and as plausible help in the sick. Associations from a cohort cannot prove cause, but the signal appears in trial meta-analyses too, and it grows with dose.

Practical meaning

The case is strongest where the trials were strongest: triglyceride lowering, and prescription-dose EPA in high-risk patients already on statins — a conversation for a clinician, not a cart. The case is weakest for a healthy person buying a 1 g blend for insurance against a heart attack; that trial has been run, and it largely failed. This evidence describes EPA and DHA themselves, not any specific commercial product.

So the inventory ends where it began: four grams that worked, one gram that didn't, and a capsule of corn oil that — on the day it counted most — beat the famous molecule outright. Fish oil is not a talisman. It is a dose, a formulation, and a population, and the label that omits those three things is telling you less than the trial did.

Educational, not medical advice.