In October 2023, researchers from the University of Colorado, Purdue and the USDA published something genuinely novel in The Journal of Nutrition: they tracked compounds unique to salmon from the plate into human blood.
It made the rounds as a cholesterol story. That is not quite what it is, and the real finding is more interesting than the headline.
What the researchers actually did
Using a technique called nutrimetabolomics, the team catalogued 508 compounds in salmon, of which 237 were distinctive to salmon compared with the other foods in the study diet. Then they looked for those compounds in the blood of 41 adults who ate salmon twice a week for five weeks as part of a Mediterranean-style eating pattern.
They found 143 of them circulating. After five weeks, 48 had measurably increased, along with 30 predicted metabolites โ breakdown products formed as the body processes those compounds. A handful of them tracked alongside better scores on a combined measure of cardiometabolic health that included LDL cholesterol and apolipoprotein B.
The honest read on what that means
We want to be straight about the limits here, because this study is easy to overstate and we have overstated it ourselves in the past.
Everyone in the study ate salmon. The trial was comparing two versions of a Mediterranean diet that differed in how much lean red meat they contained โ not in fish. Salmon was served twice weekly in both arms. So the cholesterol improvements participants saw came from adopting the whole eating pattern in place of their usual diet, and this study was not designed to separate out salmon's contribution.
The associations were few and exploratory. Out of hundreds of compounds detected, four showed an association with the cardiometabolic score. The study involved 41 people over five weeks, and the authors describe their findings as exploratory and call for follow-up trials in independent groups.
And triglycerides did not change. The parent trial found no significant change in triglycerides, glucose or insulin. Triglycerides appear here only as one component of the composite score.
So no: this study does not show that compounds in salmon lower your cholesterol. It was not built to.
What it does show, which is arguably better
The real achievement is a measurement breakthrough. For the first time, researchers could verify what someone had eaten from a blood sample rather than a food diary.
That matters more than it sounds. Nearly everything we know about diet and long-term health rests on people accurately remembering and reporting what they ate, which they are famously bad at. A biomarker that says "this person has been eating salmon" turns a self-report into a measurement. It means future studies of fish and heart health can be far more precise about who actually ate what.
Where the case for eating fish actually rests
Not on this paper. It rests on a much larger body of evidence: decades of cohort studies consistently associating regular fish consumption with lower cardiovascular mortality, the well-established triglyceride-lowering effect of EPA and DHA, and the recommendations that follow from them. The American Heart Association recommends two servings of fish a week, preferably fatty fish, at about 3 ounces cooked each. We go through that case properly in heart-healthy seafood: the best fish for a healthy heart.
Think of this 2023 paper as a new tool for studying that question, not a new answer to it.
And a note on personalised advice
If you are managing cholesterol specifically, a registered dietitian nutritionist or your doctor can give you advice built around your own numbers, medications and family history. Food is one variable among many, and no single fish is a treatment.
Source: Hill EB et al., "Salmon Food-Specific Compounds and Their Metabolites Increase in Human Plasma and Are Associated with Cardiometabolic Health Indicators Following a Mediterranean-Style Diet Intervention," The Journal of Nutrition, October 2023. Parent trial: O'Connor LE, Campbell WW, et al., American Journal of Clinical Nutrition, 2018.