Are seed oils bad for you? The human evidence says no. The largest study followed 221,054 US adults for up to 33 years (Study). People who ate the most plant oil had lower death rates than people who ate the most butter. Swapping two teaspoons of butter a day for plant oil tracked with roughly 17% lower risk of dying early.
The worry is built on real chemistry. That chemistry just does not show up when researchers measure it in people.
What Seed Oils Actually Are
Seed oils are cooking oils pressed from the seeds of plants. The usual list is canola, soybean, sunflower, safflower, corn and grapeseed.
They all contain a lot of linoleic acid, an omega-6 fat. Your body cannot make linoleic acid, so it has to come from food. This one fat is what the whole argument is about.
Getting oil out of a small seed takes more than a press. Seeds are crushed and heated, and big factories rinse the rest out with a solvent. The oil is then degummed, bleached and deodorized. The result is pale and almost tasteless, and it holds up to high heat. That is why restaurants and food factories use so much of it.
US adults eat about 16 to 19 grams of linoleic acid a day, roughly 7 to 8% of their calories (Study). Very little of it comes from a bottle on the counter.
Where the Fear Comes From
Your body turns linoleic acid into arachidonic acid. Arachidonic acid is the raw material for several signaling molecules involved in inflammation. More linoleic acid going in, the argument runs, means more inflammatory signal coming out.
Intake really did rise. Cheap seed oils spread through the food supply over the last century, and omega-6 climbed with them. Omega-3 fats did not keep pace, so the balance shifted. The oils are also refined in factories, and restaurants heat them over and over.
It is a reasonable chain of logic. Researchers have since tested every step of it in people.
Are Seed Oils Bad for Inflammation?
The first step is easy to check. If eating more linoleic acid raised arachidonic acid, blood tests would show it. Researchers pooled 36 human studies that changed how much linoleic acid people on Western diets ate (Review). The changes were dramatic, from a 90% cut up to a six-fold increase. Arachidonic acid in the blood did not follow.
The blood test itself works fine. When people took arachidonic acid directly, their levels rose as expected. So the first step of the argument does not hold up.
The next step fails too. One review gathered 15 trials that flipped a coin to decide who ate extra linoleic acid, all in healthy people (Review). Not one found a rise in any inflammatory marker. That includes C-reactive protein, the blood test doctors use to check for inflammation. An industry-funded nutrition group paid for that review.
A larger pool of trials landed in the same place. Thirty coin-flip trials in 1,377 people found no change in the usual inflammatory markers (Meta-analysis). It does carry one caveat. Among people whose intake rose the most, C-reactive protein may have edged up. So the size of the increase may matter.
The newest trial pushed the markers the other way. Thirty-nine adults with extra weight around the middle took 15 to 20 grams of linoleic acid a day for seven weeks (Trial). Three inflammatory signals in their blood fell. C-reactive protein rose slightly, by an amount chance could easily explain. Inflammatory genes were more active inside fat tissue, so the picture is mixed.
Blood sugar points the same way. Researchers pooled 102 controlled feeding trials in 4,220 adults. Swapping 5% of daily calories from saturated fat or carbohydrate to these fats slightly improved blood sugar control (Meta-analysis). Insulin measures improved only when the swap replaced carbohydrate. The changes were small, and they went the opposite way from the worry.
Butter Versus Plant Oils
Researchers followed 221,054 US nurses and health professionals for up to 33 years and recorded 50,932 deaths (Study). People in the top group for butter had about 15% higher death rates than those in the bottom group. For plant oils it ran the other way. The top group there had about 16% lower death rates.
The swap is easy to picture. Trade 10 grams of butter a day, about two teaspoons, for plant oil. That was linked to roughly 17% lower risk of dying early (Study). Deaths from cancer moved by a similar amount.
The result was not the same for every oil. Canola, soybean and olive oil each tracked with lower death rates on their own (Study). Corn and safflower oil came out too uncertain to call either way.
Nobody assigned these people their diets. Butter eaters differ from oil eaters in many ways. The team adjusted for smoking, exercise, weight, calories and diet quality. Adjusting can never strip out every difference. So this is strong evidence, though it cannot prove cause.
What Blood Levels Show
Diet studies lean on what people remember eating, and memory is unreliable. Blood tests get around that problem.
The largest analysis pooled 30 studies from 13 countries, covering 68,659 people and 15,198 heart events (Pooled analysis). People with more linoleic acid in their blood and tissue had less heart disease. They also had fewer deaths from heart disease and fewer strokes. That is the opposite of what the inflammation story predicts.
The same pool checked arachidonic acid, the molecule at the center of the worry (Pooled analysis). Higher levels were not linked to more heart disease. The main results landed at no effect at all. One side analysis even hinted at slightly less of it.
Type 2 diabetes shows a similar pattern. Twenty studies pooled records from 39,740 adults who were free of diabetes at the start (Meta-analysis). Those with the most linoleic acid in their blood were roughly a third less likely to develop it. Arachidonic acid again showed no link either way.
The same holds for dying from any cause. One pooled analysis covered 44 groups of people, 811,069 in total (Meta-analysis). The highest linoleic acid intakes came with about 13% lower risk of dying. The authors call that benefit modest.
Older diet surveys point the same way. Thirteen studies of 310,602 people found about 15% fewer heart events at the highest intakes (Meta-analysis). Nobody assigned those diets either, so the same caution applies.
The Trials Skeptics Cite
Two trials get cited against seed oils, and both deserve a serious look.
The Minnesota Coronary Experiment ran from 1968 to 1973 in six state hospitals and a nursing home, with 9,423 people (Trial). Half of them had corn oil swapped in for saturated fat. Their cholesterol fell by about 14%, and they did not live any longer for it. Researchers later recovered some of the autopsy records. Heart attack scars showed up in 41 of every 100 in the corn oil group, against 22 in every 100 in the control group. Most of that data sat unpublished for roughly 40 years.
The limits are serious. Meals came from the institution, so we know what people ate. But patients came and went constantly. Only about a quarter stayed a year or longer, and follow-up stopped the day they left.
People often blame trans fat for that result, and the paper does not support the simple version of it. Its authors argue the original team most likely picked margarines with as little trans fat as they could find. Ordinary margarines and shortenings were the big trans fat sources back then. Those were staples of the control diet instead.
The Sydney Diet Heart Study tested 458 men aged 30 to 59 (Trial). All of them had recently had a heart attack or severe chest pain. One group switched to safflower oil and safflower margarine. That fat is nearly pure linoleic acid, with no omega-3 at all. Over about three years, 18 in every 100 died in that group. Among the men given no diet advice, 12 in every 100 died.
That result is uncomfortable, and it should not be waved away. The gap was borderline, so chance could still explain it. It also covers one oil, in men who already had heart disease, over a short stretch. There was no omega-3 alongside it.
Cochrane pooled the coin-flip trials in 2018: 19 trials, 6,461 people, followed for one to eight years (Meta-analysis). Raising omega-6 made little or no difference to deaths or heart events. Heart attacks may have dropped slightly. The reviewers rated the evidence weak and called the benefits unproven.
So the fair verdict is narrow. Trials that fed people more omega-6 found no harm and no clear benefit. The long-term and blood-level evidence leans mildly positive. Neither one supports the idea that seed oils drive modern disease.
The ratio argument lands here as well. In 85,425 UK adults, the worst omega-6 to omega-3 balance came with about 26% higher risk of dying (Study). Yet in that same dataset, people with the most omega-6 in their blood were about 23% less likely to die. The ratio looks bad mainly when omega-3 is low. So the fix is to eat more omega-3 fats rather than less omega-6. A 2024 review argues for dropping the ratio as a health measure (Review). Two of its authors part-own a company that sells omega-3 blood tests.
Frying Is a Different Question
Here the critics have a real point. Heat an oil, cool it, then heat it again, and it breaks down. Some of the breakdown products are aldehydes, reactive compounds nobody wants in food. A chemistry review estimated that a 154 gram portion of fast-food fried potatoes carries about 8 milligrams of them (Review). No safety limit exists for those exact compounds. So the authors compared them with the daily limit set for a close chemical relative, and the portion came out far over it.
That evidence is chemistry, plus lab and animal work. The same authors say the human studies linking these compounds to disease have not been done.
The nearest human evidence is fried food itself. Across studies covering 562,445 people, the highest fried food intakes came with about 28% more major heart events (Meta-analysis). Deaths from heart disease and from any cause did not rise. No study of this kind can separate the oil from the batter, the salt and the portion size.
One trial has tested badly abused oil head on. Nineteen healthy men ate a meal cooked in oil that had been used for frying every day for 10 days (Trial). Their blood vessel function afterwards matched the same meal made with fresh oil. The trial was small and short, and one author works for an oil company. On its own it settles little.
The practical advice is dull. Do not reuse frying oil over and over. Keep the bottle capped, away from light and heat. Take the pan off the heat before the oil smokes.
What This Means in Your Kitchen
Most linoleic acid does not reach people from a bottle at all.
In a survey of 3,815 US adults, added fats and oils supplied only about 1 in every 10 linoleic acid calories (Study). Grains supplied about a quarter and meat about a fifth. Sweets and snacks took much of the rest. Nine of every 10 of those calories arrive already inside a food. Plenty of that is still seed oil, used in a factory rather than at your stove.
That matters, because the packaged food is what keeps tracking with poor health. Researchers fed 20 adults two menus matched for fat, sugar, salt, fiber and calories offered (Trial). On the ultra-processed menu people ate about 500 calories a day more. They gained about a kilo in two weeks on it.
A longer trial repeated the result. Fifty-five adults in England ate two eight-week menus, both built to the same national healthy-eating guidance (Trial). Weight fell about twice as much on the minimally processed menu. The ultra-processed meals packed more calories into the same volume. The gap was small in the end: about a kilo over eight weeks.
So the question worth asking is what the oil is in. Our piece on ultra-processed foods lays out that case in depth.
Butter is not poison either. Olive oil has the deepest trial record behind it, so it stays an excellent default. It is also the main fat in the Mediterranean diet. Cooking dinner in canola oil is not the same as eating fried snacks every day. Both involve a seed oil, but the exposure is very different.
Frequently Asked Questions
Are Seed Oils Bad for You?
The human evidence says no. Trials that added linoleic acid to healthy people’s diets found no rise in inflammatory markers (Review). People with more of it in their blood have less heart disease (Pooled analysis) and less type 2 diabetes (Meta-analysis). The risk that does show up sits with fried and packaged food, not with the bottle in your cupboard.
Is Olive Oil Better Than Canola Oil?
Olive oil has more trial evidence behind it, mostly as part of a whole diet. In the 33-year study, canola, soybean and olive oil each tracked with lower death rates (Study). Gram for gram, canola looked at least as good as olive oil there. Either one is a sensible everyday choice.
Should I Cook With Butter or Tallow Instead?
There is no need to fear them, and no reason to treat them as an upgrade. In that same long study, more butter tracked with higher death rates and more plant oil with lower ones (Study). Butter on toast is fine. Making it your main cooking fat is the part the data argues against.
Does Frying at Home Ruin the Oil?
Frying once in fresh oil is not the problem. The aldehyde estimates that worry people come from oil reused for days in commercial fryers (Review). One small trial found no short-term effect on blood vessel function after a meal cooked in heavily reused oil (Trial). Replace the oil instead of topping it up, and do not let it smoke.
Do I Need to Fix My Omega-6 to Omega-3 Ratio?
Probably not directly. In UK data the ratio looked worse in people who died sooner, yet the same data showed higher omega-6 tracking with lower risk (Study). The signal comes from low omega-3, not high omega-6. Some fatty-acid researchers now suggest measuring omega-3 on its own and retiring the ratio (Review).
Key Takeaways
- The mechanism does not show up in people. Eating more linoleic acid does not raise arachidonic acid in the blood, so the first step fails (Review). Coin-flip trials in healthy adults found no rise in inflammatory markers either (Review).
- Blood levels point the other way. Pooled data on 68,659 people found less heart disease and fewer heart deaths in those with more linoleic acid in their blood (Pooled analysis).
- The butter swap matters most. Over 33 years, trading two teaspoons of butter a day for plant oil was linked to roughly 17% lower risk of dying early (Study).
- The skeptics’ trials are real but narrow. The recovered Minnesota data showed no survival benefit, in people who mostly left within a year (Trial). Pooling every trial shows little difference either way (Meta-analysis).
- Reheated frying oil is the real caveat. Repeatedly used oil builds up aldehydes, though the human disease evidence is still missing (Review).
- Most of it arrives inside packaged food. About 9 in every 10 linoleic acid calories are already in a packaged or prepared food when you buy it (Study).
Judge the Meal, Not the Oil
Across 33 years of follow-up, tens of thousands of blood samples and dozens of trials, this fat keeps coming out neutral or mildly helpful. The fried and packaged food it arrives in is the part that looks bad.
So cook with what you enjoy. Use olive oil for flavor and canola or sunflower for high heat, and keep butter for the jobs you want butter for.
Then look at how much of your week comes out of a packet.
Seed oils are the loudest argument in nutrition right now. They are not the part of your diet most worth worrying about, and that is worth knowing before you throw out a bottle.
This article is for educational purposes and is not medical advice. Talk to a qualified clinician before changing your health regimen.
Leave a comment