Are seed oils actually bad for you?

On the evidence available, no — and the trials point in the opposite direction. When saturated fat is replaced by the polyunsaturated fat found in seed and vegetable oils, cardiovascular events fall in randomised controlled trials by roughly a quarter to a third. Prospective cohort studies agree. The proposed mechanism for harm — that omega-6 fat drives inflammation — has been directly tested in controlled human feeding studies and does not occur: increasing dietary linoleic acid does not meaningfully raise tissue arachidonic acid or circulating inflammatory markers. That is an unusual situation in nutrition, where a specific biochemical claim has been examined head-on and not confirmed. There are legitimate secondary questions about repeatedly reheated commercial frying oil, but that is a catering practice issue and applies to every oil equally.

What about the omega-6 to omega-3 ratio?

It was a reasonable hypothesis that has largely been abandoned by the field. The problem with a ratio is that it can be improved in two entirely different ways — eating less omega-6 or eating more omega-3 — and only the second reliably corresponds to better outcomes. When absolute intakes are examined separately, omega-3 intake predicts cardiovascular outcomes and omega-6 intake does not predict harm. No major dietary guideline anywhere sets a ratio target. The construct survives in popular discussion because it allows a low omega-6 diet to be presented as the solution, when the actual evidence supports eating more oily fish rather than less sunflower oil.

Why do so many people believe seed oils are harmful?

The claim moved from low-carbohydrate nutrition circles into mainstream discussion through podcasting from around 2020, helped by a memorable framing that named a specific list of oils as culprits. It then attached to something broader: a general and not unreasonable suspicion of industrial food processing, and of institutions perceived as having got dietary advice wrong before — which, on trans fats, they demonstrably did. That history gives the argument emotional traction that the biochemistry does not supply. It is worth understanding, because a belief held for reasons of trust rather than data is not dislodged by data alone, and because the underlying instinct — that heavily processed food deserves scepticism — is actually sound and points towards the ultra-processed food evidence rather than towards the oils themselves.

Is there anything to the concern at all?

Two things, neither of which supports the popular claim. First, oils repeatedly heated to high temperatures over long periods, as in commercial deep-fat frying with inadequate oil turnover, accumulate oxidation products including reactive aldehydes; this is a genuine food safety and catering standards question, it applies to any oil used the same way, and it has nothing to do with linoleic acid being harmful in itself. Second, seed oils travel with ultra-processed foods, so people consuming a lot of them differ systematically in overall diet and circumstances — which makes observational data hard to interpret, and is precisely why the randomised trials matter more here than usual. Neither concern supports avoiding canola or sunflower oil in a home kitchen.

What should I actually cook with?

Extra virgin olive oil is the best-supported choice for general use, with outcome evidence from the PREDIMED trial behind it, and there is no reason to argue against it. Beyond that, rapeseed or canola, sunflower and other seed oils are entirely reasonable, and are a better choice than butter, tallow or coconut oil if the aim is cardiovascular health, because those raise LDL cholesterol. If the goal is to eat better, the intervention that matters is reducing overall ultra-processed food intake, increasing vegetables, legumes, whole grains and oily fish, and cooking more at home — the specific oil in the pan is a minor variable that has attracted attention wildly disproportionate to its effect.

How does GMJ News cover the seed oil question?

GMJ News covers the randomised trial evidence on fat substitution, the tested and unsupported inflammation mechanism, the abandonment of the omega-6 to omega-3 ratio, substitution analysis as the key methodological concept, the recovered Sydney and Minnesota trials and why they are outliers, oil oxidation chemistry and the legitimate commercial frying question, confounding by ultra-processed food intake, what WHO, AHA and EFSA guidelines actually state, and the documented transmission history of the claim from 2020 onwards. Part of PHIG's contested science and nutrition programme.

Home › Topics › Seed Oils

Seed Oils

GMJ News knowledge hub · last reviewed September 2026 · Georgian Medical Journal

Summary Statistics Glossary GMJ news FAQ Documents Organizations Research

Few nutrition claims have travelled further on less evidence than the proposition that seed oils — canola, sunflower, soybean, corn, safflower, grapeseed, cottonseed and rice bran — are a principal cause of modern chronic disease, a narrative that spread rapidly through podcasts and social media from around 2020 and has since acquired substantial political as well as dietary identity (WHO). The scientific position is not ambiguous: randomised trials and their meta-analyses show that replacing saturated fat with polyunsaturated fat lowers cardiovascular events, the proposed mechanism — that dietary linoleic acid drives systemic inflammation via arachidonic acid — is not supported by controlled human feeding studies, and the omega-6 to omega-3 ratio has largely been abandoned as a useful clinical construct because absolute omega-3 intake, not the ratio, predicts outcomes. There are legitimate residual questions worth stating honestly — the oxidation products of repeatedly reheated commercial frying oil, and the fact that seed oils are a marker for ultra-processed food intake — but neither supports the claim that has actually gone viral.

Key messages

SETTLED: replacing saturated fat with polyunsaturated fat lowers cardiovascular events

This is not a matter of interpretation. Randomised controlled trials of dietary fat substitution, synthesised by Cochrane and by the 2020 Mozaffarian and Hooper meta-analyses, show that replacing saturated fat with polyunsaturated fat — which in practice means seed and vegetable oils — reduces cardiovascular events by roughly 25-30%. Prospective cohorts show the same direction. Mendelian randomisation on LDL supports the mechanism. Any claim that seed oils cause heart disease must account for trial evidence showing the opposite, and the popular claim does not attempt to.

UNSUPPORTED: that dietary linoleic acid drives systemic inflammation

The central mechanistic claim is that linoleic acid (omega-6) is converted to arachidonic acid and then to pro-inflammatory eicosanoids. It is biochemically real and quantitatively trivial: controlled human feeding studies show that increasing dietary linoleic acid does NOT meaningfully raise tissue arachidonic acid, because the conversion step is tightly regulated and rate-limited. Systematic reviews of randomised trials find no increase in circulating inflammatory markers (CRP, IL-6, TNF-alpha) with higher linoleic acid intake. The mechanism is invoked constantly and has been tested and found not to operate as claimed.

ABANDONED: the omega-6 to omega-3 ratio as a clinical target

The ratio was a plausible construct in the 1990s and has been largely discarded, because it turns out that absolute omega-3 intake predicts outcomes while the ratio adds nothing independent — one can improve the ratio by eating less omega-6 without any physiological benefit, or by eating more omega-3, which is what actually helps. No major dietary guideline sets a ratio target. It persists in popular discourse precisely because it allows a low omega-6 diet to be presented as the solution when the evidence supports increasing omega-3 instead.

GENUINELY OPEN: repeatedly reheated commercial frying oil

There is a real question here, and it is not the one being argued about. Oils subjected to prolonged repeated high-temperature frying — the commercial deep-fat fryer, not the domestic pan — accumulate oxidation products including aldehydes, and some of these have demonstrated toxicity in laboratory systems. Human outcome evidence remains limited. This is a legitimate food safety and catering regulation question about oil degradation and turnover, entirely separate from whether linoleic acid is harmful, and it applies to any oil heated in the same way, including those the seed oil narrative promotes.

CONFOUNDING: seed oils are a marker for ultra-processed food

Much of the observational signal that appears to implicate seed oils reflects the company they keep. Industrially refined oils are components of ultra-processed foods, fried fast food and commercial baked goods, and people with high intakes differ systematically in overall diet, activity, smoking and socioeconomic position. This is a real interpretive problem in the cohort literature — and it cuts against the seed oil hypothesis rather than for it, because the randomised trials that remove this confounding show benefit rather than harm.

Why the belief spread — the transmission is documentable

The claim moved from the fringes of low-carbohydrate nutrition into mainstream discourse through podcasting from around 2020, acquired a memorable framing in the so-called "hateful eight" oils, and became embedded in a broader wellness and political identity in which industrial food processing stands for institutional untrustworthiness. Understanding this matters clinically: the belief is not primarily held for nutritional reasons and is therefore not primarily dislodged by nutritional argument. Acknowledging the legitimate underlying instinct — that heavily processed food is worth eating less of — is more productive than contesting the biochemistry alone.

Key statistics

25-30%

reduction in cardiovascular events when saturated fat is replaced by polyunsaturated fat in RCTs

Cochrane/Circulation

No increase

in inflammatory markers with higher dietary linoleic acid in systematic reviews of RCTs

Am J Clin Nutr

Rate-limited

linoleic acid does not meaningfully raise tissue arachidonic acid in controlled feeding studies

J Nutr/AJCN

No guideline

sets an omega-6 to omega-3 ratio target — the construct has been abandoned

WHO/EFSA/AHA

~2020

the point at which the claim moved from fringe low-carbohydrate circles into mainstream podcasting

Documented

Reheated oil

the one legitimate open question — commercial frying degradation, applying to all oils equally

EFSA/FSA

Seed oils — where the disagreement actually lies

Source: Bars show strength of supporting evidence. High values indicate well-supported propositions; low values indicate claims tested and not supported.

Glossary of key terms

Linoleic acid and the arachidonic acid pathway
BiochemistryLinoleic acid is an essential omega-6 fatty acid that humans cannot synthesise. It can be elongated and desaturated to arachidonic acid, which is the precursor for prostaglandins, thromboxanes and leukotrienes — some pro-inflammatory, some anti-inflammatory or resolving. The popular argument stops at the first step of that chain. Two facts complete it: the conversion is tightly rate-limited by delta-6 desaturase, so large increases in dietary linoleic acid produce minimal increases in tissue arachidonic acid; and arachidonic acid is also the precursor for lipoxins and other specialised pro-resolving mediators that terminate inflammation. The pathway is a regulatory network, not a one-way inflammation switch, and this is why controlled feeding studies show no inflammatory effect.

Substitution analysis
Nutritional epidemiologyThe single most important methodological concept in dietary fat research, and the one most often ignored in popular argument. Because total energy intake is broadly fixed, no nutrient can be reduced without something replacing it, so the meaningful question is never "is X harmful?" but "is X harmful compared with what?" Reducing saturated fat produces benefit when polyunsaturated fat replaces it, no clear benefit when refined carbohydrate replaces it, and intermediate results with monounsaturated fat. Meta-analyses that pool trials without regard to the replacement nutrient dilute genuine effects towards null — which is exactly how "saturated fat is exonerated" and, inversely, several seed oil arguments are constructed from real data.

The Sydney Diet Heart Study and Minnesota Coronary Experiment
Evidence/HistoryThese recovered trials are the strongest evidence the seed oil position has, and they deserve a fair hearing rather than dismissal. Both replaced saturated fat with omega-6-rich oils and, on reanalysis of recovered original data, reported unfavourable mortality results. The counterarguments are substantial: the Sydney study used safflower oil and margarines containing high levels of industrial trans fat, a genuinely harmful exposure not present in modern oils; the Minnesota trial had high participant turnover, short exposure and institutionalised populations; and both are outliers against the larger body of trials. They are legitimately cited in the scientific literature as reasons for caution about overconfidence, and they do not establish the harm claimed for them.

Oil oxidation, smoke point and aldehydes
Food chemistryPolyunsaturated fats have more double bonds and are therefore more susceptible to oxidation than saturated or monounsaturated fats, generating hydroperoxides and secondary products including aldehydes such as 4-hydroxynonenal. The rate depends on temperature, duration, oxygen exposure, light, prior oil use and the presence of antioxidants including tocopherols. In practice: domestic cooking at normal temperatures for normal durations produces minimal degradation; repeated commercial deep frying without adequate oil turnover produces substantially more. Refined oils generally have higher smoke points than unrefined ones, and smoke point is a poor proxy for oxidative stability in any case. This is a real food safety topic and a poor argument against dietary linoleic acid.

What the guidelines actually say
PolicyWHO recommends replacing saturated and trans fats with unsaturated fats, specifically including polyunsaturated sources. The American Heart Association issued a dedicated advisory reaffirming that omega-6 polyunsaturated fat intake is beneficial and specifically addressing the inflammation claim. EFSA sets adequate intake levels for linoleic acid as an essential nutrient. Dietary guidelines in the EU, UK, US, Canada and Australia are aligned. No national or international body recommends restricting seed oils, and this near-total consensus is itself sometimes offered as evidence of institutional capture — a reasoning pattern that cannot be falsified and should be named as such.

Olive oil, tallow and the alternatives proposed
ComparativeThe alternatives promoted in place of seed oils deserve the same scrutiny they demand. Extra virgin olive oil has the strongest outcome evidence of any culinary fat, supported by the PREDIMED trial, and is a reasonable primary choice on its own merits — this is not in dispute. Beef tallow, butter and coconut oil are high in saturated fat and raise LDL cholesterol; coconut oil in particular has been marketed on the basis of medium-chain triglyceride research that does not apply to it, since most of its fat is lauric acid, which behaves as a conventional LDL-raising saturated fat. Recommending tallow over canola is therefore a recommendation that moves in the direction the trial evidence advises against.