Seed oils: the controversy, the chemistry, and what the evidence actually says.

Few topics in modern nutrition produce more heat and less light than the seed oils debate. On one side: a vocal coalition of podcasters, low-carb advocates and a subset of the wellness industry asserting that vegetable and seed oils are the principal driver of chronic disease and should be eliminated. On the other: mainstream cardiology and public health, which has spent fifty years recommending the replacement of saturated fat with polyunsaturated fats including seed oils. Both sides cite studies. Both sides accuse the other of bad faith. The actual literature is more nuanced, less satisfying, and — if read carefully — quite informative about what to do.

§ What seed oils actually are

The term 'seed oils' usually refers to industrially refined oils extracted from soybeans, corn, sunflower, safflower, cottonseed, grapeseed, rice bran, and rapeseed (canola). The unifying chemical feature is a high content of linoleic acid, an 18-carbon omega-6 polyunsaturated fatty acid (PUFA). Linoleic acid is essential — the human body cannot synthesise it — and it is the substrate from which the body builds membrane lipids and signaling molecules.

These oils are produced by mechanical pressing followed by solvent extraction (hexane), then refining, bleaching and deodorising. The processing produces a neutral-tasting, shelf-stable, high-smoke-point oil at low cost. This processing is genuinely industrial in scale, and the resulting oils have been a major source of dietary fat in industrialised diets since roughly the 1960s.

Population-level linoleic acid intake has risen sharply over this period, and several chronic disease epidemics have risen alongside it. The correlation has fuelled a strong intuition that the rise in seed-oil consumption must be causally implicated. Intuition is not evidence, but it is a reasonable starting hypothesis.

§ What the randomised evidence shows

The strongest evidence for any nutritional intervention comes from randomised controlled trials with hard endpoints — heart attacks, strokes, mortality. On the question 'what happens when you replace saturated fat with polyunsaturated fat in the diet', there are several large RCTs going back to the 1960s.

The 2020 Cochrane review by Lee Hooper, pooling fifteen RCTs with nearly 60,000 participants, found that reducing saturated fat intake by replacing it with polyunsaturated fats produced a modest but statistically significant reduction in cardiovascular events (approximately a 17 percent relative risk reduction). The effect on total mortality was smaller and less certain.

The Mozaffarian 2010 meta-analysis came to similar conclusions: replacing 5 percent of energy from saturated fat with PUFA reduced coronary heart disease events by approximately 10 percent.

The Marklund 2019 paper used biomarkers of linoleic acid intake (rather than self-report) in 30 cohorts and found that higher circulating linoleic acid was associated with lower cardiovascular disease and total mortality.

This is a real signal, repeated across study designs, with a coherent biological story (PUFA lowers LDL cholesterol relative to saturated fat). It is not enormous. It is real.

§ What the critics get right

The seed-oil critics raise legitimate points that mainstream summaries sometimes elide.

First, the Minnesota Coronary Experiment — a large 1960s-70s trial buried for decades and reanalysed by Ramsden in 2016 — found that subjects given corn oil in place of saturated fat had lower cholesterol but no improvement in mortality, and in fact a possible increase in death rates in older participants. This is one study, with methodological limitations, but it complicates the simple PUFA-is-good narrative.

Second, polyunsaturated fats are chemically reactive. The double bonds that define them are vulnerable to oxidation, particularly under heat. Repeatedly heated oils, of the sort used in commercial deep-fat fryers, produce aldehydes and other oxidation products with demonstrable cellular toxicity in laboratory work. Whether this matters at the doses produced by home cooking with fresh oil is much less clear.

Third, the dramatic increase in dietary linoleic acid over the twentieth century is a real and unprecedented dietary shift. The default assumption that the body's response to this shift is benign is an assumption, not a tested fact.

§ What the critics get wrong

Where the seed-oil-as-poison thesis falls apart is in the leap from 'oxidised oils in industrial fryers may be harmful' to 'all linoleic acid intake from all sources causes inflammation and chronic disease'. The mechanistic argument relies on the idea that omega-6 fats are pro-inflammatory because they are precursors to arachidonic acid and certain prostaglandins. But human trials measuring inflammatory markers in response to increased linoleic acid intake have repeatedly failed to find a meaningful inflammatory signal at typical or even high dietary intakes. The body regulates the conversion of linoleic to arachidonic acid tightly; eating more linoleic does not flood the system with inflammatory mediators in the way the simple substrate-product model would predict.

The epidemiological data going the other direction — Marklund and others — is hard to reconcile with the inflammation hypothesis. If linoleic acid intake were strongly pro-inflammatory and pro-disease, biomarker studies should pick this up. They consistently do not.

"Almost every food argument that becomes a moral crusade turns out to be an argument about ultra-processed foods, wearing a costume."

§ The real driver, almost certainly

The most honest reading of the literature is that the seed-oil debate is, to a significant extent, a proxy debate for an ultra-processed food debate. The places where seed oils are consumed in enormous quantities — packaged snacks, fast food, deep-fried restaurant items, processed baked goods — are also the places where excess calories, refined carbohydrates, added sugars, sodium, and a general dilution of nutrient density meet the eater. The oil is the carrier, not necessarily the agent.

Repeatedly reused commercial frying oil is a different question, and the evidence here is more concerning. Studies of cardiovascular biomarkers in populations heavily exposed to deep-fried foods do show worsened lipid and oxidative-stress profiles, and this is a reasonable place to be cautious.

§ What to actually do at home

Pragmatic positions consistent with the data:

• →Cooking at home, with any reasonable culinary fat, is unlikely to be a meaningful driver of disease risk. Extra-virgin olive oil for low-to-medium heat, a neutral oil (canola, avocado, refined olive) for higher heat, butter in moderation, are all defensible.
• →The single biggest action a typical adult can take with respect to dietary fat is to reduce intake of deep-fried foods from commercial sources, where repeatedly heated oils dominate.
• →Avoid the framing in which one oil is good and another is poison. The literature does not support either categorical claim.
• →The composition of the rest of the diet — fiber, whole foods, vegetables, fish, intact grains — matters substantially more than whether you sauté in olive or canola.
• →If you want to optimise omega-6 to omega-3 ratio, adding two to three servings of oily fish per week (or an algae-based EPA/DHA supplement for vegetarians) is supported by stronger evidence than any seed-oil avoidance strategy.

§ A closing point

Nutritional science attracts more confident assertion than it deserves. The seed-oils debate is a useful case study in how a real but modest signal can be amplified, in either direction, by people whose business model depends on certainty. The honest position requires holding two ideas at once: industrial deep frying with reused oils is probably worse for you than the public-health establishment admits, and home cooking with a bottle of canola is probably not the cause of the chronic disease epidemic that the loudest critics claim. The interesting argument is about the food. The oil is mostly a passenger.