1. A document written to resolve a contradiction
Most consensus statements summarise. This one reads closer to an intervention. Its abstract names the problem outright: omega-3 fatty acids reduce triglycerides effectively, yet randomised trials of their effect on atherosclerotic cardiovascular disease (ASCVD) “are inconsistent,” and their value in heart failure, arrhythmia, cardiomyopathy, hypertension and sudden cardiac death “remains confusing.”
Published in the Chinese Circulation Journal in February 2023, it sets out the residual-risk rationale — triglyceride (TG) and triglyceride-rich lipoprotein cholesterol (TRL-C) as a causal contributor to the risk that persists after LDL-C is controlled — and then walks through the trial record honestly, including the failures.
2. The contradiction, in one table
Eight major trials. The spread is stark.
- GISSI-P (1999, n=11,324 post-MI): EPA+DHA 1.0 g/day — primary composite endpoint −15%.
- JELIS (2007, n=18,645, all on statin): EPA 1.8 g/day — major coronary events −19% in secondary prevention; no reduction in primary prevention.
- ASCEND (2018, n=15,480 diabetics): EPA+DHA 1.0 g/day — serious vascular events not reduced.
- VITAL (2019, n=25,871): EPA+DHA 1.0 g/day — MACE not reduced.
- REDUCE-IT (2019, n=8,179, >99% on statin): IPE 4.0 g/day — primary endpoint −25%; cardiovascular death −20%.
- STRENGTH (2020, n=13,078, all on statin): omega-3 carboxylic acids 4.0 g/day — no benefit (12.0% vs 12.2%).
- OMEMI (2021, n=1,027 elderly post-MI): EPA+DHA 1.8 g/day — all-cause death unchanged (HR 1.07).
Two trials in the same decade, at the same 4 g/day dose, produced a 25% risk reduction and a null result. That is the puzzle the consensus sets out to solve.
3. Four variables that explain the disagreement
Composition. A meta-analysis cited by the consensus concluded that EPA alone reduces cardiovascular risk more significantly than EPA+DHA. It goes further, citing the INSPIRE study: when EPA is combined with DHA, higher DHA levels attenuate EPA’s benefit — identified as one reason STRENGTH showed nothing.
Dose. Event reduction increases with dose, and high-dose EPA (>1 g/day) outperforms low-dose (≤1 g/day) — notable because ASCEND and VITAL both used 1 g/day.
Achieved blood level. The most striking figure in the document compares serum EPA concentrations actually reached: 89.6 μg/mL in STRENGTH, 135.2 μg/mL in REDUCE-IT, 170.0 μg/mL in JELIS. The inference is that serum EPA must cross a threshold — which means the product has to deliver enough EPA to get there. “EPA purity, content, and the treated patient’s blood EPA level,” the consensus concludes, “may be the main reasons for the divergent cardiovascular benefit across randomised controlled trials.”
Placebo. REDUCE-IT used mineral oil, and a biomarker substudy found that at 12 months the mineral oil arm had risen in markers including oxidised LDL-C (10.9%) and hs-CRP (21.9%), while the fish-oil arm showed no significant change in any of them. The consensus declines to treat this as decisive — an FDA advisory committee judged the effect too small to change the result, and non-mineral-oil trials (CHERRY, JELIS) also showed benefit — but flags the design issue for future trials.
4. EPA and DHA are not interchangeable — and the reason is mechanical
The consensus dedicates a section to why the two fatty acids behave differently, and it is unusually specific.
EPA has a lipophilic, stable extended conformation: it reduces membrane fluidity, maintains uniform cholesterol distribution and — critically — inhibits formation of the cholesterol crystalline domains that destabilise plaque. Its antioxidant activity, the consensus notes, is not replicable by other triglyceride-lowering drugs and is long-lasting. Two imaging trials support the mechanism: CHERRY showed statin plus EPA reduced coronary plaque volume and increased stability; EVAPORATE showed statin plus IPE reduced low-attenuation (unstable) plaque volume by 17% (P=0.0061).
DHA behaves differently: it undergoes rapid conformational change in the cell membrane, which works against stabilisation; its anti-lipid-oxidation activity is comparatively limited; and it cannot reduce cholesterol-rich domain formation.
EPA is also a precursor to 3-series prostanoids — including PGI3, a potent endothelial vasodilator that reduces platelet aggregation — and to 5-series leukotrienes with anti-inflammatory action. By competing with arachidonic acid for membrane phospholipid synthesis, it shifts the balance away from TXA2 toward TXA3, inhibiting platelet activity.
The consensus is careful about limits: EPA acts mainly on blood vessels while DHA is abundant in nervous tissue and the retina, and no head-to-head EPA-versus-EPA+DHA trial exists.
5. The regulatory consequence has already landed
IPE is the only omega-3 fatty acid approved by the FDA, Canada and the EU for a cardiovascular risk-reduction indication. EPA+DHA preparations once held an EU cardiovascular indication too — but after review, the EU concluded that EPA+DHA 1 g/day was ineffective for secondary prevention, offered no favourable benefit-risk balance post-myocardial-infarction, and withdrew that indication in 2019. Elsewhere, EPA+DHA mixtures now hold only a severe hypertriglyceridaemia indication.
The 2022 American Diabetes Association guideline reinforces the boundary: REDUCE-IT results should not be extrapolated to other products. China’s 2020 post-CABG consensus recommends high-purity EPA rather than a DHA-containing mixture.
6. Where the evidence is solid
Triglycerides. Definitive and dose-dependent. Omega-3 at 4 g/day lowers TG by 20–30% at TG 2.3–5.6 mmol/L and by ≥30% above that, and different compositions perform similarly on this endpoint. In severe hypertriglyceridaemia the effect is markedly larger: MARINE showed IPE 4 g/day cut TG by 33.1% versus 19.7% at 2 g/day. On that basis the FDA approved EPA+DHA in 2004 and IPE in 2012 for severe hypertriglyceridaemia.
Heart failure. GISSI-HF cut all-cause death by 9% (P=0.041) over 3.9 years. In 2022 the AHA/ACC/HFSA heart failure guideline recommended omega-3 as add-on therapy for NYHA class II–IV patients — the first time omega-3 has been recommended from a heart failure treatment perspective.
Blood pressure. Modest but real, and dose-responsive: a meta-analysis of 71 trials put the optimal dose at 2–3 g/day, reducing blood pressure by roughly 2.6/1.6–1.8 mmHg, with a stronger dose-response in hypertensive and elderly populations.
7. Where the evidence is not
Atrial fibrillation. The consensus is blunt. Trials do not support reducing AF recurrence, and the 2017 AHA advisory recommends against omega-3 in AF patients. A recent meta-analysis found omega-3 increased atrial fibrillation relative risk by 25%. There is also a bleeding signal: JELIS showed total bleeding events rising from 0.6% to 1.1% (P=0.0006) and REDUCE-IT a trend (2.7% vs 2.1%) — neither reached haemorrhagic stroke or fatal bleeding, but co-administration with anticoagulants warrants monitoring.
Sudden cardiac death. Early claims of 45–81% reductions did not survive; a pooled analysis of 20 studies (n=68,680) found no significant reduction. The one persistent signal is in patients not receiving guideline-adjusted therapy, where risk fell 36% — reinforcing that omega-3’s measurable benefit shrinks as background treatment improves.
8. What the consensus actually asks for
The operative recommendation is numerical: in high- or very-high-risk ASCVD patients already on strict lifestyle intervention and statins, if TG remains above 1.5 mmol/L (135 mg/dL), high-dose IPE at 4 g/day is recommended. Note the thresholds — not “consider omega-3,” but a specific molecule, a specific dose, a specific residual-risk marker.
9. What this means for the supply chain
Purity and monomer separation move from differentiator to specification. The consensus’s own explanation for the trial failures implicates achieved serum EPA, which is a function of how much EPA the product contains. A portfolio running from 70–88% EPA+DHA concentrates to single-monomer grades above 85% with the counterpart held below 0.1% is what the evidence now asks for.
DHA is not a free rider. If DHA attenuates EPA’s benefit, and DHA-containing preparations lost their cardiovascular indication in the EU, then blending EPA and DHA and labelling it omega-3 is a formulation decision with clinical and regulatory consequences — not a default.
Dose changes the sensory requirement. At 4 g/day, the consensus’s adverse-event section leads with gastrointestinal complaints — fishy taste, belching, diarrhoea, nausea — noting that taking the dose with food reduces them and improves absorption. Colour and odour stop being cosmetic at this dose; they become adherence variables. Gardner ≤ 2 and an essentially odourless concentrate are the supply-side answer to a documented compliance problem.
One number to watch. The consensus records that EPA+DHA preparations are reported to raise LDL-C by roughly 15–36% while ApoB does not change, suggesting larger LDL particles rather than more of them.
The takeaway
This consensus does not read like a marketing document. Its most useful contribution is negative: it dismantles the idea that “omega-3” names a therapeutic class. What it names is a family of molecules whose clinical behaviour depends on which one you use, how pure it is, how much is delivered, and what blood level is reached.
For manufacturers, that is a demanding specification. It is also, precisely, the specification that separates a concentrate producer from a fish oil seller.
Industry commentary on a published expert consensus, intended for a professional non-clinical audience. Not medical advice. Clinical recommendations, dosing and treatment decisions should be taken from the source document — Chinese Expert Consensus on the Role and Application of Omega-3 Fatty Acids in the Prevention and Treatment of Cardiovascular Diseases, Chinese Circulation Journal, February 2023, Vol. 38, No. 2 — and from qualified clinicians.
