Fish Oil at the Crossroads: What the REDUCE-IT vs. STRENGTH Trials Actually Tell Us

Fish oil is the third most widely used dietary supplement in the United States. Tens of millions of people take it daily, largely on the assumption that it protects the heart. That assumption has been under sustained scientific pressure for several years — not because the science is straightforward, but precisely because it is not. Two major clinical trials asked nearly the same question, in comparable patient populations, with comparable doses of omega-3 fatty acids, and arrived at opposite conclusions. That gap is what this episode is about: what the evidence actually shows on omega-3 fatty acids, cardiovascular risk, and formulation differences — and whether it supports the confidence with which fish oil lines pharmacy shelves. Start with what fish oil is supposed to do. Omega-3 fatty acids, principally EPA (eicosapentaenoic acid) and DHA (docosahexaenoic acid), have a well-documented effect on blood triglycerides. High triglycerides are an independent risk factor for cardiovascular disease, and omega-3s, especially at doses above two grams per day, reliably lower them. This mechanism is not in dispute. The dispute is whether lowering triglycerides with omega-3s actually translates into fewer heart attacks, strokes, or cardiovascular deaths — and whether the answer depends on which specific omega-3 compound is used, at what dose, in which patient population. REDUCE-IT, published in 2018, appeared to settle that question in favour of omega-3s. The trial enrolled 8,179 patients across 473 sites in eleven countries. These were not low-risk individuals. They were statin-treated patients aged 45 or older with established cardiovascular disease, or aged 50 or older with diabetes plus at least one additional cardiovascular risk factor. All had elevated triglycerides — between 135 and 500 mg/dL — despite being on statins. Median follow-up was 4.9 years. The treatment was icosapent ethyl, a purified prescription form of EPA, at 3,840 milligrams per day. The result was striking. The primary composite endpoint — a five-point MACE score covering cardiovascular death, nonfatal myocardial infarction, nonfatal stroke, coronary revascularisation, and hospitalisation for unstable angina — occurred in 17.2% of the treatment group versus 22.0% of the placebo group. That is a hazard ratio of 0.75, with a 95% confidence interval of 0.68 to 0.83. The number needed to treat for the primary endpoint was 21. For the harder secondary endpoint of cardiovascular death, myocardial infarction, or stroke, the hazard ratio was 0.74, with a number needed to treat of 28. These are clinically meaningful numbers. A 25% relative reduction in major cardiovascular events over five years, in a high-risk population already on statins, is not a trivial finding. The methodological controversy surrounding REDUCE-IT is central to interpreting everything that came after, so it needs to be understood clearly. The placebo in the trial was mineral oil. Most cardiovascular trials use a genuinely inert placebo, such as a cellulose capsule. The choice of mineral oil was ostensibly to match the appearance and consistency of the fish oil capsules. Steven Nissen, a prominent cardiologist at the Cleveland Clinic, raised the concern that mineral oil is not biologically inert. In the REDUCE-IT control arm, LDL cholesterol rose by approximately 5 mg/dL relative to baseline, and C-reactive protein — a marker of inflammation — also increased. If the placebo was actively worsening cardiovascular risk in the control group, the apparent benefit in the treatment arm would be inflated: the trial would be measuring the difference between a working drug and a mildly harmful comparator, rather than between the drug and a neutral baseline. This criticism has not been resolved. REDUCE-IT's investigators, including lead author Deepak Bhatt, contest the interpretation and argue that even accounting for the lipid changes in the placebo arm, the magnitude of the cardiovascular benefit cannot be fully explained by mineral-oil effects alone. The PMC analysis attempting to unravel the discrepancies between REDUCE-IT and STRENGTH reaches no definitive conclusion on this point. It is a genuinely open empirical question. STRENGTH was designed in part to address exactly this uncertainty. It used a corn-oil placebo — considered more metabolically neutral than mineral oil — and tested a different formulation: a combination of 2,200 mg of EPA and 800 mg of DHA per day. The primary endpoint was similar to REDUCE-IT's. The result was null. No cardiovascular benefit was detected. Plasma omega-3 levels in STRENGTH participants increased by 269%, a dramatic biochemical signal that patients were taking the drug and absorbing it. Triglycerides fell. But events did not. The obvious interpretation is that the EPA+DHA combination does not reduce cardiovascular events, while purified EPA does. The Lancet published a meta-analysis finding that across omega-3 trials, those using EPA alone showed higher relative reductions in cardiovascular outcomes than those using EPA and DHA together, with statistically significant interaction terms. This aligns with a mechanistic hypothesis: that DHA may counteract some of EPA's cardiovascular effects, or that purified high-dose EPA has actions beyond triglyceride lowering that the combination formulation does not replicate. That is a plausible story. The Cleveland Clinic's secondary analysis of STRENGTH, however, complicates it in an important way. The researchers identified STRENGTH participants who had achieved plasma EPA levels comparable to those seen in REDUCE-IT participants at twelve months — the reasoning being that if high EPA exposure is the active ingredient, this subgroup should show benefit regardless of formulation. They did not. No cardiovascular benefit was found even among those STRENGTH patients with the highest plasma EPA levels, matching what REDUCE-IT achieved. This is a direct challenge to the EPA-mechanism hypothesis. If high plasma EPA is what drives the benefit in REDUCE-IT, a trend toward benefit would be expected in STRENGTH patients who reached similar plasma levels. The absence of that signal supports Nissen's view that REDUCE-IT's result may be an artefact of the placebo rather than a pharmacological effect of EPA. The Lancet meta-analysis and the Methodist DeBakey Cardiovascular Journal analysis lean toward EPA superiority being real. Neither side has conclusive evidence. The trial that would resolve this — a head-to-head comparison using a genuinely inert placebo — has not been done. There is a separate signal that is more consistent across the evidence, and it runs in the other direction: atrial fibrillation risk. In REDUCE-IT, hospitalisation for atrial fibrillation or flutter occurred in 3.1% of the icosapent ethyl group versus 2.1% of the placebo group — a difference that was statistically significant at p equals 0.004. Across the broader literature, high-dose omega-3 supplementation is associated with an elevated risk of atrial fibrillation. Standard over-the-counter doses appear to carry only a small bleeding and atrial fibrillation signal, but at the prescription-grade doses used in REDUCE-IT, the risk is measurable. Whether the cardiovascular benefit outweighs the atrial fibrillation risk depends on the individual patient's baseline risk profile. A-fib is itself a serious condition — it raises stroke risk substantially — so a therapy that reduces myocardial infarction while increasing A-fib is not straightforwardly beneficial for every patient in the eligible population. The evidence does not clearly establish whether the A-fib risk is smoothly dose-dependent across the full range from over-the-counter to prescription levels, and it does not identify which patient subgroups are most vulnerable. This brings us to the over-the-counter question, because that is where most people actually engage with fish oil. Retail supplement doses typically range from 250 mg to 1,000 mg of combined EPA and DHA per capsule, and most people take one to two capsules daily — a fraction of the doses used in either REDUCE-IT or STRENGTH. Deepak Bhatt, REDUCE-IT's lead investigator, was explicit on this point: the trial results should not be extrapolated to retail fish oil supplements. The supplements are not regulated for purity, concentration, or bioavailability, and there is no demonstrated cardiovascular benefit at standard over-the-counter doses in clinical trials. The clinical recommendations that emerged from REDUCE-IT specifically reference 4 grams per day of EPA in statin-treated patients with elevated triglycerides — a very different context from someone buying fish oil at a pharmacy because they heard it is good for the heart. This is where cardiology and the supplement aisle diverge sharply. The American Heart Association and cardiologists interpreting STRENGTH concluded that the evidence for over-the-counter fish oil as a cardiovascular preventive is weak. The Methodist DeBakey Journal supports a targeted recommendation: prescription EPA for a specific high-risk subgroup, not general supplementation. Supplement sales, meanwhile, continued largely undisturbed. The marketing language on fish oil bottles is not written by cardiologists reviewing the STRENGTH data. One further limitation is worth naming. Essentially all the major trial evidence on omega-3s and cardiovascular outcomes is in patients with elevated triglycerides — which is the clinical rationale for using them in the first place. What the evidence shows for patients without elevated triglycerides at baseline, who represent a large fraction of the people actually buying fish oil, is not addressed by these trials. REDUCE-IT required a minimum triglyceride of 135 mg/dL. Whether omega-3s at any dose reduce cardiovascular events in people with normal triglycerides is a genuinely open question. So here is what the evidence actually shows. Prescription-grade purified EPA at 4 grams per day produced a substantial reduction in cardiovascular events in REDUCE-IT, a large, well-conducted trial in statin-treated high-risk patients with elevated triglycerides. That result is real in the sense that the outcome data are real — but whether the mechanism is genuine EPA pharmacology or partially an artefact of a harmful placebo remains empirically unresolved. The EPA+DHA combination tested in STRENGTH showed no cardiovascular benefit, even at doses that produced a 269% increase in plasma omega-3 levels. A Lancet meta-analysis suggests EPA-only formulations perform better than EPA+DHA formulations across multiple trials. The Cleveland Clinic's secondary analysis of STRENGTH challenges the EPA-specific mechanism by showing that STRENGTH participants who matched REDUCE-IT's plasma EPA levels still showed no benefit. Both interpretations remain on the table. High-dose omega-3s at clinically studied levels consistently raise atrial fibrillation risk — a finding that has to be weighed against any cardiovascular benefit. And none of this translates to the standard over-the-counter fish oil supplement, which has not demonstrated cardiovascular benefit in rigorous trials and operates at a fraction of the studied dose. The most important unresolved question in this space is whether REDUCE-IT's 25% event reduction reflects EPA's pharmacology or the mineral-oil placebo's harm. Answering it would require a trial that does not yet exist — one testing purified EPA against a genuinely inert control. Until that trial is done, the two possible explanations for the REDUCE-IT result remain structurally incompatible with each other, and the clinical recommendations built on that trial rest on a foundation that has not been fully secured.