Are seed oils actually bad for you? What the evidence separates

Thirty trials raised people's intake of linoleic acid, the main fat in sunflower, soybean and corn oil, and then measured inflammation in their blood.¹ Nothing moved. The central claim against seed oils is specific enough to test, it has been tested, and it did not hold up.

The claim, and why it can be checked

The argument runs in three steps. Linoleic acid converts in the body to arachidonic acid. Arachidonic acid feeds the signaling that produces inflammation. So more seed oil means more inflammation.

Every step is real chemistry. That is what makes the claim worth taking seriously, and also what makes it checkable. The question is only whether the chain produces anything measurable in a living person.

What happened when the trials looked

Thirty controlled studies in 1,377 people fed extra linoleic acid and measured the standard inflammation markers in blood.¹

Those markers use different units, so results get put on one shared scale. On that scale 0.2 counts as a small effect, 0.5 as moderate and 0.8 as large.

C-reactive protein, the marker doctors actually order, came in at 0.09 across 16 studies — and the range around it ran from -0.05 to 0.24, which includes no effect at all.¹ Tumor necrosis factor sat at -0.01. Interleukin-6 at 0.11. Adiponectin at 0.17. Monocyte chemoattractant protein 1 at 0.14. Every one of those ranges crossed zero too.¹ Fibrinogen, plasminogen activator inhibitor-1 and the adhesion molecules did not budge either. The studies agreed with each other, and there was no sign that unflattering results had gone missing.¹

One thread deserves pulling, because it is the only one there. Looking within that null result, the people whose linoleic acid intake rose most steeply may have seen C-reactive protein rise.¹ That is a hypothesis found inside a negative finding, at intakes above what most of the trials tested. It is not nothing. It is also not what the popular claim says.

What happens to people who eat a lot of it

Asking people to recall what they ate is unreliable, so the strongest observational work measures the fatty acid sitting in blood or fat tissue instead. The largest such effort reanalyzed the raw data from 30 long-running studies across 13 countries, covering 15,198 heart events among 68,659 people.²

More linoleic acid in the body went with less disease, not more. Comparing people near the top of the range with people near the bottom, total cardiovascular disease ran 7% lower, deaths from heart disease 22% lower, and clot-type strokes 12% lower.² Arachidonic acid — the very metabolite the mechanism blames — was not linked to higher risk of anything, and the people with the most of it had 8% less cardiovascular disease.²

A 2025 survey of 150 sets of cohort data found the same direction, for both what people ate and what was in their blood: lower rates of heart disease, of cancer, and of death from any cause.³ It also found what an honest summary has to carry. Higher omega-6 intake tracked with higher risk of ovarian and endometrial cancer, and the protective pattern disappeared in people who already had a health condition.³

These studies follow people who chose their own diets. They establish that people with more linoleic acid in their tissue are not the people having more heart attacks. They cannot establish that the oil is why.

The real question is: instead of what?

Nobody eats fat on its own. Cutting seed oil means eating something else in its place, and that is the comparison the trials were built to make — though they were built around saturated fat rather than around seed oils.

The Cochrane review gathered 15 trials and 56,675 people. Cutting saturated fat lowered combined heart events by 17%, across 12 trials and 53,758 people.⁵ In people without existing heart disease, 56 had to make that change for one to avoid an event; in people who already had heart disease, 53.⁵ Deaths from any cause barely shifted.⁵ And it could not tell apart replacing saturated fat with polyunsaturated fat from replacing it with carbohydrate.⁵

A 2026 review of 17 trials and 66,337 people sharpened the picture by sorting people by their starting risk. For adults at low risk, the benefits over five years fell below the reviewers' own thresholds for what counts as worth having. For adults at high risk, they cleared them.⁴ And the replacement that stood out was seed-oil fat: swapping saturated fat for polyunsaturated fat left about a quarter fewer nonfatal heart attacks.⁴

So the trial evidence puts seed oils on the useful side of a swap — in people already at risk, for one outcome, in a comparison the reviewers themselves rated only moderately credible.⁴ That is a far narrower claim than either side of the argument tends to make.

Bottle against bottle

The choice people actually face is one bottle against another, and that has been studied mostly through cholesterol rather than through heart attacks.

A review of nine earlier analyses found that using palm oil in place of polyunsaturated oils raised LDL cholesterol by 3.43 mg/dL, and in place of monounsaturated oils by 9.18 mg/dL.⁶ Lard raised LDL by 8.39 and 9.85 mg/dL against those same two.⁶ Coconut oil replacing polyunsaturated oils raised total and HDL cholesterol but left LDL alone.⁶

How swapping fats changed LDL cholesterolmg/dL change in LDL

Palm oil replacing polyunsaturated oil+3.43 mg/dL

Palm oil replacing monounsaturated oil+9.18 mg/dL

Lard replacing polyunsaturated oil+8.39 mg/dL

Lard replacing monounsaturated oil+9.85 mg/dL

Soybean oil replacing other polyunsaturated oilsno change

The quiet result is the one nobody quotes. Soybean oil substituted for other polyunsaturated oils changed blood fats not at all, and rice bran oil lowered LDL.⁶ Swapping among the seed oils does close to nothing. Swapping out of them, toward tropical or animal fat, pushes LDL up.

What none of this tested

Heating. Not one study here compared a heated oil with an unheated one, or measured the compounds that form in a hot pan. The frying argument sits entirely outside this evidence, which makes it an open question rather than a refuted one.

Very high intakes. The trials raised linoleic acid within a range a person could reach by eating, and the only hint of an inflammatory signal came from the steepest increases.¹

Food, as opposed to fat. Seed oils reach most people inside fried and packaged food. A finding about linoleic acid in blood is not a finding about what those foods do as foods.

What is still unknown

Whether the ovarian and endometrial cancer link means anything causal. It comes from cohorts, and it points the opposite way from every other outcome in the same review.³

Whether the polyunsaturated replacement advantage is real. It rests on one comparison, in one review, for one nonfatal outcome.⁴

Whether any of it matters to someone at low risk. Over five years, the answer there was little or no benefit.⁴

And whether the oil or the food it arrives in carries the effect. Nothing here separates the two.