The Case Against Seed Oils: 5 Claims, Checked
People say seed oils are bad for five reasons: hexane, heat oxidation, omega-6 inflammation, trans fats and processing. Here is what each one holds up.
The short answer: five claims, not one
People say seed oils are bad for five distinct reasons, and those reasons are not equally well supported: hexane extraction, oxidation products from heating, the omega-6 inflammation hypothesis, trace trans fats created during refining, and seed oils as a marker for ultra-processed food. Two of the five hold up reasonably well — the heating one and the ultra-processed one — while the most repeated claim, that linoleic acid inflames you, is the one where pooled analyses of controlled human trials have generally not found the effect the theory predicts.
"Seed oils" in this argument usually means the refined, neutral-tasting oils extracted from seeds: soybean, canola, corn, sunflower, safflower, cottonseed, grapeseed and rice bran. What counts and what does not is fuzzier than the debate suggests, and it is worth knowing where the line falls before you try to cook around it — see Which Oils Are Not Seed Oils? A Clear List.
Claim 1: They are extracted with a petroleum solvent
Where it came from: the industrial process is genuinely unappetizing to read about. Most commodity seed oil is not pressed. The seed is flaked, heated and washed with hexane, a solvent derived from petroleum, then the oil is degummed, neutralized, bleached and deodorized. That is an accurate description, not a conspiracy theory, and How Cooking Oil Is Made: Seed to Bottle walks through each step.
What the evidence supports: hexane is used, and hexane is not harmless. Long-term inhalation at industrial concentrations has been linked to nerve damage in exposed workers, which is why occupational exposure limits exist.
What it does not support: that the trace amounts left in a bottle on a supermarket shelf do the same thing. The solvent is volatile and mostly recovered during refining; regulators set residue limits for finished oil, and the exposure route that harmed workers — breathing concentrated vapor for years — is not the exposure route of a tablespoon of oil in a pan. If the process still bothers you, cold-pressed and expeller-pressed oils exist and say so on the label.
Confidence in the claim as usually stated: low.
Claim 2: Heating them creates harmful breakdown products
This is one of the two claims with real substance behind it.
Polyunsaturated fats are, structurally, the easiest fats to oxidize — the double bonds that make them liquid at refrigerator temperature are also the sites where oxygen attacks. Heat, oxygen, light, time and especially repeated reuse degrade them into a mixture that includes aldehydes, polymers and what food chemists call total polar compounds. That is not controversial among food scientists; it is also why many countries regulate the polar-compound level in commercial deep fryers and require operators to change the oil.
What it does not support: the leap from "degraded frying oil contains reactive compounds" to "the bottle in your cupboard is harming you." The evidence linking these compounds to disease in people is largely mechanistic and animal-based rather than drawn from human outcomes, and the conditions that generate them in quantity — hours at frying temperature, reused across days, in contact with air — are a commercial fryer, not a home sauté in fresh oil.
Confidence: moderate, but narrowly. The claim earns its keep for repeatedly reused frying oil and the food cooked in it. It does not transfer cleanly to a teaspoon of canola in a stir-fry.
Claim 3: Omega-6 linoleic acid drives chronic inflammation
This is the claim you hear most and the one the evidence supports least, at least in the form it is usually stated.
The mechanism sounds tidy: linoleic acid converts to arachidonic acid, arachidonic acid is the precursor to several pro-inflammatory signalling molecules, seed oils are the main dietary source of linoleic acid, and intake of it rose through the twentieth century alongside chronic disease. Each of those statements is defensible on its own. The chain that connects them is where it breaks down.
Conversion of linoleic acid to arachidonic acid in humans is limited and appears to be regulated — feeding people more linoleic acid does not move tissue arachidonic acid much. And when the question has been tested directly, by pooling controlled trials that raised or lowered linoleic acid intake and measured circulating inflammatory markers, the reviews have generally not found linoleic acid raising those markers. Biomarker cohorts, which measure linoleic acid in blood rather than relying on food questionnaires, have generally associated higher levels with lower cardiovascular risk, not higher.
What is genuinely unsettled: very long-term effects, intakes far above the ordinary range, whether crowding out omega-3 matters independently, and individual variation. "Not demonstrated" is not the same as "disproven." But the confident version of this claim is out ahead of its evidence. What Seed Oil Studies Actually Found goes through the trials in more detail, and What Seed Oils Do in Your Body, Step by Step follows the biochemistry itself.
Confidence: low for the inflammation claim as usually stated.
Claim 4: Refining makes trans fats, and the label hides them
Deodorizing is a high-temperature vacuum step, and high temperature rearranges some double bonds into trans configuration. Refined oils do contain small amounts of trans fat that the seed never had. That part is true.
The labelling half of the complaint is also true, and it is specific to the United States: FDA rules allow a food with less than 0.5 grams of trans fat per serving to declare 0 g. So "0 g trans fat" means "below the rounding threshold," not "none." Small servings and multiple ingredients can add up.
What it does not support: treating these traces as equivalent to the partially hydrogenated oils that were phased out of the food supply. Those delivered trans fat in quantities many times larger, and the outcome evidence against them was strong and human. The traces from deodorizing are far smaller, and no human study has shown harm at those amounts.
Confidence: the chemistry is solid; the health significance at these levels is unestablished. Low to moderate.
Claim 5: Seed oils are a marker for ultra-processed food
This is the second claim with substance, and it is the one that best explains why people feel better when they quit.
Refined seed oil is in most packaged snacks, commercial baked goods, sauces, dressings and fried restaurant food, because it is cheap, neutral and stable. Anyone who adopts a strict no-seed-oil rule stops eating nearly all of that at once, and starts cooking and reading labels. Controlled feeding research has found that people eat more calories when the same nutrients are delivered in ultra-processed form, so that exclusion has a plausible independent effect.
So "I cut seed oils and felt better" is usually a true observation with a confounded explanation. The oil changed, but so did roughly everything else on the plate.
Confidence: moderate that seed oil content marks ultra-processed food; low that the oil itself is the operative ingredient.
The five claims side by side
| Claim | Where it came from | What the evidence supports | What it does not support | Confidence |
|---|---|---|---|---|
| Hexane extraction | Real industrial refining process | Hexane is used; high chronic inhalation harmed workers | Harm from trace residues in bottled refined oil | Low |
| Oxidation from heating | Food chemistry of degraded frying oil | Reused frying oil forms aldehydes and polar compounds; regulated commercially | Home cooking in fresh oil; human outcome data is thin | Moderate, for reused frying oil |
| Omega-6 inflammation | Linoleic acid to arachidonic acid pathway | The pathway exists and intake rose historically | Pooled human trials have generally not found raised inflammatory markers | Low |
| Trans fats from refining | High-temperature deodorizing step | Trace trans isomers form; US labels round under 0.5 g to zero | Equivalence with partially hydrogenated oils | Low to moderate |
| Marker for ultra-processed food | Seed oil is in most packaged and fried food | Cutting seed oils removes a lot of ultra-processed food | That the oil is the harmful part rather than the pattern | Moderate |
What a fair reading leaves you with
If you want to act on the parts that hold up, the priorities are unglamorous: eat less deep-fried restaurant food, do not reuse frying oil at home more than lightly, store oil away from heat and light, and cook more of your own meals. None of that requires deciding who won the argument. If you prefer olive oil, butter or avocado oil on taste, tradition or smoke point, that is a perfectly good reason and needs no inflammation theory attached.
Where NutriAI fits, honestly: the app estimates a meal from a photo and gives it a letter grade for estimated inflammatory potential, and oil is exactly where a photo estimate is weakest — you cannot see how much oil is in a restaurant dish, or which oil it was. So it does not try to police oil brands. What it can do is let you log meals and symptoms and, over time, surface possible patterns with stated confidence — a correlation in your own log, not a diagnosis, and not a clinical measurement. If fried meals really do track with how you feel, that is something a log can show you and an argument on the internet cannot.
What this is based on
- Systematic reviews pooling controlled human feeding trials of linoleic acid intake and circulating inflammatory markers
- Observational cohort studies using blood linoleic acid biomarkers rather than dietary questionnaires
- Food-chemistry literature on thermal degradation of frying oils, aldehyde formation and total polar compounds
- Regulatory frameworks for solvent residues in edible oils and for trans fat labelling thresholds in the United States
- Controlled feeding research comparing energy intake on ultra-processed versus minimally processed diets
Frequently asked questions
- Which seed oil claim is the strongest?
- The heating one, narrowly defined. Repeatedly reused frying oil forms aldehydes and polar compounds, and that degradation is measurable enough that many countries regulate oil quality in commercial fryers. The claim is much weaker for fresh oil used once at home.
- Do seed oils cause inflammation?
- The usual version of that claim is not well supported. Pooled analyses of controlled human trials that raised or lowered linoleic acid have generally not found it raising circulating inflammatory markers, and blood-biomarker cohorts have generally associated higher linoleic acid with lower cardiovascular risk. Long-term effects remain less certain.
- Is the hexane in seed oils dangerous?
- Hexane is genuinely used as an extraction solvent, and chronic industrial inhalation has harmed workers. That is a different exposure from the trace residue left in refined bottled oil, which is regulated and has not been shown to cause harm in people at those levels. Cold-pressed oils avoid the step entirely.
- Why do people feel better after cutting seed oils?
- Usually because they also cut most packaged snacks, fried restaurant food, sauces and commercial baked goods at the same time, and started cooking. The change is real; attributing all of it to the oil rather than the whole dietary pattern is the part that outruns the evidence.
- Does "0 g trans fat" on a label mean zero?
- In the United States, no. FDA rules let a food with less than 0.5 grams per serving declare 0 g, so the label means "below the rounding threshold." Refining does create trace trans fats, though in amounts far below what partially hydrogenated oils once delivered.
- Should I stop cooking with seed oils?
- The evidence does not demand it. If you want to change something, reducing deep-fried restaurant food and avoiding heavily reused frying oil targets the part of the case that holds up. Choosing another fat for taste or cooking style is fine on its own terms.