1. A mechanism paper worth reading before the next claim is written
The review — New insights into mechanisms of action for omega-3 fatty acids in atherothrombotic cardiovascular disease, by R. Preston Mason of Brigham and Women’s Hospital, Harvard Medical School — takes a position that is easy to state and hard to sustain: EPA and DHA produce measurably different effects on membrane structure, lipid dynamics, lipid oxidation rate and cell signalling. Written in 2019, before the STRENGTH result landed, it reads today as a useful map of why the trial record turned out the way it did.
2. EPA and DHA occupy different places in the membrane
The most striking evidence in the paper is physical. Using small-angle X-ray diffraction, Mason’s group showed the two fatty acids adopt different positions in a lipid membrane.
EPA raises electron density across a broad span of the membrane’s hydrocarbon core, indicating an energy-stabilised, extended orientation — its long axis running parallel to the phospholipid acyl chains.
DHA, by contrast, interacts mainly with the phospholipid headgroup region and reduces electron-density in the core, consistent with greater molecular volume and disorder.
The reason is conformational. DHA’s additional carbon atoms and double bonds allow it to isomerise through its full range of conformations within roughly 50 nanoseconds of entering a membrane. That flexibility is not a neutral difference. Cholesterol has a rigid sterol ring and a relatively fixed position in the bilayer, so a rapidly flexing acyl chain interacts with it less tightly. The measurable consequence: DHA promotes cholesterol-rich microdomain formation; EPA does not.
This is where biophysics becomes clinical. Cholesterol is roughly 30–50% of total lipid in mammalian plasma membranes, and when oxidative damage or cholesterol excess drives those immiscible domains to precipitate, they form insoluble extracellular cholesterol crystals — sharp-edged structures documented in the plaques of patients who have suffered myocardial infarction, which mechanically destabilise the fibrous cap. This is not purely a cholesterol-excess phenomenon: membranes exposed to oxidative stress or high glucose form cholesterol domains even at normal cholesterol levels.
Both cholesterol crystals and oxidised LDL are major activators of the NLRP3 inflammasome, which processes pro-IL-1β into the cytokine that initiates atherosclerotic inflammation. EPA, at pharmacologically relevant concentrations, significantly inhibits glucose-induced cholesterol crystalline domain formation in model membrane vesicles.
3. What EPA does that nothing else does
The review is emphatic on a point that matters commercially as much as scientifically: EPA’s antioxidant effect cannot be replicated by other triglyceride-lowering drugs — niacin, gemfibrozil, fenofibrate — or by vitamin E. The failure of vitamin E is attributed to its limited lipophilicity and radical-scavenging capacity; EPA, once embedded, uses its multiple double bonds to interrupt free-radical chain reactions through electron stabilisation.
Two details matter. EPA suppressed oxidation in ApoB-containing particles for longer than DHA did. And EPA’s antioxidant effect on small dense LDL (sdLDL) was enhanced, not merely preserved, when combined with atorvastatin — the two amphiphilic molecules appear to share an action site, and their interaction further stabilises unpaired lipid radicals. sdLDL matters because it oxidises more readily than larger LDL, and is therefore more atherogenic.
4. The statin handshake
The review also describes a synergy that is not about lipids at all. Endothelial dysfunction is causally linked to atherosclerosis; EPA reverses markers of it induced by oxidised LDL or hyperglycaemia — and in isolated human endothelial cells that effect is amplified in the presence of atorvastatin’s active metabolite. The improvement shows up as a favourable shift in the nitric oxide to peroxynitrite (NO/ONOO⁻) release ratio.
The mechanism detail: the NO/ONOO⁻ improvement was independent of eNOS expression levels, pointing to improved eNOS efficiency — a better coupling state — rather than more enzyme. Uncoupled eNOS generates superoxide instead of NO, lowering NO bioavailability and increasing LDL oxidation, so recoupling the enzyme attacks the problem upstream.
The review also reports that EPA-rich HDL isolated from coronary artery disease patients showed enhanced cholesterol efflux capacity and inhibited cytokine-stimulated VCAM-1 expression while increasing resolvin E3 generation — EPA inserting into HDL particles more efficiently than DHA.
A fair caveat: much of this is model-membrane, in-vitro and animal work. It is a mechanistic hypothesis with strong internal consistency, not clinical proof. The review is explicit that confirmatory mechanism trials are still running.
5. Then the paper changes subject — and gets uncomfortable
Midway through, the review turns to fish oil dietary supplements (FODS), and the tone shifts.
The setting is significant. FODS are the most widely used dietary supplement among US adults. They are not regulated by the FDA as drugs; classified as food, they require no pre-market efficacy or safety evaluation, and their active content and chemical integrity are not held to the same standard. Because EPA and DHA content is typically low, a patient may need to swallow ten or more capsules a day to reach a prescription-level dose of up to 4 g.
The supply chain is also different. Bulk FODS oil is a by-product of industrial marine processing — often of animal-feed protein production — and separating protein involves heating fish to around 100 °C. For a molecule whose defining feature is multiple unsaturated double bonds, that is close to a worst-case process. Light and contaminants accelerate the damage further.
6. What the supplement data actually showed
The review collects three datasets, and the numbers are the argument.
- A 2015 USDA-funded study of 47 FODS: only 10 had EPA at or above label; only 12 had DHA matching label; 74% were below label on one or both.
- A New Zealand study of 32 FODS: only 9% matched their labelled omega-3 content; over 80% contained unacceptably high lipid peroxide levels; only three products (8%) met international limits for peroxide and total oxidation value.
- North American FODS were likewise shown to carry excessive lipid peroxides.
Then the comparison that should settle it. A study of leading US-market FODS measured fatty acid composition and oxidative damage against an FDA-approved prescription product. The supplements contained more than 30 fatty acids, including 10 to 14 saturated fatty acids making up over a third of total fatty acids; EPA and DHA content varied several-fold between products; and every widely sold supplement exceeded the recommended upper limit for oxidation products. The prescription product showed no significant oxidation products and no undesirable lipids.
7. Why oxidised oil fails — mechanically
The review closes the loop with a bioactivity experiment. Small dense LDL was treated with either unoxidised omega-3, or oxidised omega-3, or a supplement extract containing both.
Unoxidised omega-3 inhibited sdLDL oxidation by more than 95% (p < 0.001). The oxidised material showed no inhibition at all.
Read that against the mechanism described earlier and the logic is unforgiving. The entire value proposition — interrupting free-radical chain reactions, protecting ApoB particles, suppressing cholesterol domain formation, preserving endothelial NO coupling — depends on the omega-3 molecule arriving with its double bonds intact. Oxidise them in the bottle and the mechanism has nothing to work with. The clinical correlates reported elsewhere are consistent: no lipid benefit, and no improvement in lipid or inflammatory parameters.
8. What this means for the supply chain
Oxidation control is the product, not a quality check. The review’s own evidence shows that oxidised omega-3 is not weaker omega-3; it is inactive omega-3. Peroxide and anisidine values, and total oxidation value, belong on the specification alongside the assay.
Concentration is what makes the dose possible. If the therapeutic range starts at grams per day, then material that requires ten capsules to get there is not a delivery mechanism — it is a tolerance problem. Concentrates at 70–88% EPA+DHA, and monomers above 85% with the counterpart held below 0.1%, are what make a gram-scale dose practical.
The separation is the differentiation. The mechanism review’s central claim is that EPA and DHA are not interchangeable at the level of membrane physics. If that is true, then the ability to separate them — rather than blend them — is not a manufacturing preference. It is a product-definition decision.
Saturated fat is not background noise. Ten to 14 saturated fatty acids making up over a third of total fatty acids is a statement about the refining process that produced the oil — and physical distillation that separates by boiling point leaves a different, inspectable profile than solvent routes.
The takeaway
This review is a mechanism paper, and its most quoted lines will be the ones about EPA’s antioxidant effect being non-replicable and enhanced by statins. But the more consequential section may be the middle one — the quality audit.
It establishes that the cardiovascular mechanism and the commercial failure mode are the same chemistry viewed from two directions. EPA only does what the review describes while it is intact, unoxidised, and present at sufficient concentration. Material that arrives oxidised, diluted with a third saturated fat, or under-labelled cannot produce the effect — and the review’s own data say most of the supplement market arrives that way. That is the case for high-purity concentrate production, made from mechanism rather than marketing.
Industry commentary on a published mechanistic review, intended for a professional non-clinical audience. Not medical advice. The review is Mason RP, “New insights into mechanisms of action for omega-3 fatty acids in atherothrombotic cardiovascular disease,” first published online 12 January 2019. Clinical conclusions should be taken from the source literature and from qualified clinicians.
