
Evidence explainer · Heart & Circulation
How to Read Heart Research When the Headlines Disagree
Use two omega-3 trials to see why the population, formulation, comparison group and measured outcome matter more than an ingredient headline.
- Question
- How can two studies of omega-3 lead to different conclusions?
- Sources
- 5 cited
- Quick answers
- 3 questions
- Clinician review
- None; sources cited inline
In short
Key takeaways
- Start with who was studied, what was given, what it was compared with and which outcome changed. One study rarely settles every question.1
- REDUCE-IT studied a prescription EPA formulation in statin-treated people at increased cardiovascular risk; STRENGTH studied a different EPA/DHA formulation and comparator.23
- Those findings do not establish that any retail fish-oil product produces the same result.4
Chapter
Four questions before you accept the headline
Who took part? What exact intervention did they receive? What did the comparison group receive? What outcome did researchers count? These details connect a headline to the actual experiment. A short change in a laboratory marker and years of follow-up for clinical events are not the same test.1
Read the limitations alongside the result. Statistical significance does not by itself tell you how useful an effect is, whether it applies to another population or whether the balance of benefit and harm favors an intervention. A single study belongs within a wider evidence base.1
Chapter
Two trials, different formulations and comparisons
REDUCE-IT enrolled 8,179 statin-treated people with elevated triglycerides and established cardiovascular disease or diabetes with additional risk factors. Its prescription EPA formulation, icosapent ethyl, was compared with placebo. Over a median 4.9 years, a primary composite event occurred in 17.2% of the treatment group and 22.0% of the placebo group. Those percentages describe that trial population and follow-up, not a personal prediction.2
STRENGTH enrolled 13,078 statin-treated people at high cardiovascular risk. It compared a carboxylic-acid formulation containing EPA and DHA with corn oil. It found no statistically significant difference in its composite cardiovascular outcome. This was not the same formulation or comparator as REDUCE-IT, so placing both under a single “fish oil works” or “fish oil fails” label loses important information.3
The comparator also matters: REDUCE-IT used mineral oil, which was not biologically neutral and may have enlarged the apparent between-group benefit. NIH also notes an increased risk of atrial fibrillation in higher-dose omega-3 trials involving people with cardiovascular disease or high risk. Those uncertainties and harms belong beside the benefit figures, not only in a footnote.5
Chapter
Read the size and meaning of an effect
In REDUCE-IT, subtracting the reported event proportions gives a 4.8-percentage-point difference over the trial’s follow-up. That is an absolute difference; it is not a promise that any individual’s risk will fall by that amount. The outcome was also a composite containing several kinds of cardiovascular events rather than only heart attacks.2
Ask whether the result concerns symptoms, a blood marker or clinical events, and look for harms as well as benefits. Our omega-3 evidence guide includes formulation-specific findings, comparator concerns and heart-rhythm risks. The measurement guide helps distinguish the outcomes before comparing them.
Chapter
Keep the evidence attached to what was actually tested
A retail supplement may differ in formulation, amount, quality controls and intended use. NHLBI specifically distinguishes prescription omega-3 medicines from dietary supplements when discussing high triglycerides. It is not scientifically sound to transfer a prescription-trial result to a bottle solely because both contain omega-3.4
Start Follow oxygen to reconnect the study question with normal physiology, or compare omega-3 labels after reading the evidence. Label comparison can clarify product facts; it cannot turn an untested retail product into the intervention used in a trial.
Questions
Quick answers
Why not declare one trial right and the other wrong?
Does the 4.8-percentage-point difference predict my benefit?
No. It is the difference between the trial’s reported group event proportions over its follow-up. Personal risk, treatment suitability and harms require clinical context.2
Does an ingredient mechanism prove a benefit?
No. A plausible explanation helps generate a research question. Human outcomes and safety still need to be tested, and one study rarely provides a final answer for every use.1
Keep reading
More in Heart & Circulation
Researched and written by Kymata Health Editorial; sources cited inline. Drawn from the studies and documents cited below. Educational only, not medical advice.
Sources
References
Every study and document cited in this guide, numbered as in the text.
- 1How To Make Sense of a Scientific Journal Article: Results (accessed 2026). NCCIH, 2026.
- 2Cardiovascular Risk Reduction with Icosapent Ethyl for Hypertriglyceridemia. New England Journal of Medicine; abstract in LVHN Scholarly Works, 2019.
- 3Effect of High-Dose Omega-3 Fatty Acids vs Corn Oil on Major Adverse Cardiovascular Events in Patients at High Cardiovascular Risk: The STRENGTH Randomized Clinical Trial. JAMA, 2020. PMID 33190147
- 4High Blood Triglycerides (accessed 2026). NHLBI, 2026.
- 5Omega-3 Fatty Acids: Health Professional Fact Sheet (accessed 2026). NIH Office of Dietary Supplements, 2026.
Dietary supplements are not intended to diagnose, treat, cure or prevent any disease, and these statements have not been evaluated by the FDA. Prescription options need a licensed clinician. Talk to your clinician or pharmacist before starting anything new, especially if you take prescription medicines, are pregnant or breastfeeding.