Seed oils: what they are, what the linoleic acid evidence shows and where the claims run ahead

A bottle of sunflower or canola oil used to be the unremarkable item in the shopping basket, the thing a recipe meant when it said “vegetable oil”. In the past few years it has acquired a reputation. Podcasts and social media accounts describe seed oils as industrial, inflammatory and a hidden driver of heart disease, obesity and much else, and some restaurants now advertise that they cook without them.

Mainstream heart advice says the opposite. The American Heart Association’s position is that replacing saturated fats such as butter and lard with unsaturated oils lowers cardiovascular risk. The research behind that position is large, and it is also less tidy than either side tends to admit. The biggest review of long-term studies, from the Harvard T.H. Chan School of Public Health, found that people who ate more of the main fat in seed oils died at lower rates. Randomized trials, which are a stronger test, show lower cholesterol and have not clearly shown fewer deaths.

What are seed oils?

The term covers oils pressed or extracted from seeds: soybean, corn, sunflower, safflower, cottonseed and canola, which is known as rapeseed oil in Britain. Olive, coconut and palm oil come from fruit and are not usually included.

What these oils share is a high content of linoleic acid. This is a polyunsaturated fat, a description of its chemical structure that also explains why the oils are liquid at room temperature, and it belongs to the omega-6 family. The body cannot make linoleic acid and needs a small amount from food. A review from the University of Tennessee calls it the most highly consumed polyunsaturated fatty acid in the Western diet and notes that it is found in virtually all commonly consumed foods, because the oils are used throughout food manufacturing as well as in home kitchens.

How much linoleic acid do people eat, and how has that changed?

Far more than a century ago. Researchers at the US National Institutes of Health reconstructed American food supplies year by year and found that the estimated per capita consumption of soybean oil increased more than a thousandfold between 1909 and 1999. Over the same period, the availability of linoleic acid (LA) increased from 2.79% to 7.21% of energy, LA being the standard abbreviation. In other words, the share of calories coming from this one fat rose about two and a half times.

That rise is the starting point for the case against seed oils: a food that was scarce for most of human history now supplies a meaningful slice of the diet, over the same decades in which obesity and diabetes became common. Timing alone does not show cause, since sugar, portion sizes, cars and screens changed over those decades too. The question is what studies of the fat itself show.

Do seed oils drive inflammation?

The inflammation argument rests on body chemistry. Linoleic acid can be converted into another fat, arachidonic acid (AA), and arachidonic acid is the raw material for signaling molecules that promote inflammation. A 2012 review summarized the worry: concern has been expressed that a high intake of dietary n-6 polyunsaturated fatty acid contributes to excess chronic inflammation, n-6 being another way of writing omega-6.

The chain is real in a test tube. Whether it operates in people eating ordinary diets has been checked directly, and the answer so far is no.

The first link fails. The Tennessee reviewers, Brian Rett and Jay Whelan, gathered human trials in which linoleic acid intake was deliberately raised or lowered and blood fats were measured. Linoleic acid’s role as the metabolic precursor of arachidonic acid, meaning the substance it is made from, did not translate into higher levels: when dietary linoleic acid levels were increased up to six fold, no significant correlations with arachidonic acid levels were observed. The body appears to regulate the conversion tightly.

The end of the chain was tested as well. Guy Johnson and Kevin Fritsche reviewed randomized trials that fed healthy people more linoleic acid and measured markers of inflammation in their blood. Fifteen trials qualified. The authors concluded that virtually no evidence is available from randomized, controlled intervention studies to show that adding the fat raises those markers.

These trials were short and measured blood markers, so they do not rule out every possible effect. They do mean the central mechanism claimed for seed-oil harm has been looked for in people and not found.

What do long-term studies of linoleic acid and death show?

The largest test is a 2020 meta-analysis by Jun Li, Frank Hu and colleagues at Harvard. Before it, the authors wrote, evidence on linoleic acid and death was inconsistent and has not been summarized by a systematic review. They found that thirty-eight studies reporting 44 prospective cohorts were identified, in which people’s diets were recorded and deaths were then counted over years. Together these included 811,069 participants with dietary intake assessment.

People who ate the most linoleic acid had a 13% lower rate of death from any cause than those who ate the least, a 13% lower rate of death from cardiovascular disease (CVD) and an 11% lower rate of death from cancer. The authors’ summary is restrained: higher intake was associated with a modestly lower risk of mortality from all causes.

Food questionnaires are imprecise, and people who cook with vegetable oil may differ from people who cook with butter in many other ways. A second line of evidence gets around the first problem. Linoleic acid cannot be made by the body, so the amount in someone’s blood or body fat reflects what they have eaten. An international consortium led by Matti Marklund at Uppsala University pooled such measurements from 30 prospective observational studies from 13 countries. Over the years of follow-up, more than 15,000 cardiovascular events occurred among 68 659 participants, and those with more linoleic acid in their blood or tissue had fewer of them, with cardiovascular deaths about a fifth lower across the range. On the fat at the center of the inflammation theory, the result was the reverse of the prediction: AA levels were not associated with higher risk of cardiovascular outcomes.

The consortium had set out to address this uncertainty and inform international recommendations, and it concluded that these results support a favorable role for LA in CVD prevention.

What have seed oil trials found?

Cohorts can mislead, and trials are the check. The trial record is where the critics of seed oils have their strongest material.

The fullest summary is a Cochrane review led by Lee Hooper at the University of East Anglia. It covered 19 randomized trials, with 6461 participants who were followed for one to eight years. In most of them the omega-6 fats usually displaced dietary saturated or monounsaturated fats. The reviewers found that increased intake of omega-6 fats may make little or no difference to all-cause mortality or to cardiovascular events overall. There was a hint of fewer heart attacks, too uncertain to rely on. The one solid finding was on a blood test: high-quality evidence that increasing omega-6 fats reduces total serum cholesterol a little in the long term. The authors’ verdict was that, although benefits of omega-6 fats remain to be proven, people at high risk of heart attack might gain from them.

One trial in that record has become famous. The Minnesota Coronary Experiment ran from 1968 to 1973 in one nursing home and six state mental hospitals in Minnesota, where 9,423 residents were served either a standard diet or one in which saturated fat was swapped for linoleic acid from corn oil and corn oil polyunsaturated margarine. Much of its data went unpublished for decades. Christopher Ramsden of the National Institutes of Health recovered and analyzed it in 2016. The corn oil diet lowered cholesterol by about 14%, as intended. Yet the survival curves showed no mortality benefit for the intervention group, and among participants whose cholesterol fell furthest, deaths were higher.

Ramsden’s team argued that incomplete publication has contributed to overestimation of the benefits of replacing saturated fat with vegetable oils rich in linoleic acid. The trial has a clear weakness: only 2,355 of its participants stayed on the study diets for a year or more, as patients moved in and out of hospital. It still stands as a caution that lower cholesterol from diet did not, in this case, mean longer life.

What heart guidelines say about seed oils

The American Heart Association reviewed the same trial record in a 2017 presidential advisory and came to a firmer conclusion than Cochrane. The advisory reported that randomized controlled trials that lowered intake of dietary saturated fat and replaced it with polyunsaturated vegetable oil reduced CVD by about 30%, which it compared to the effect of statin drugs. Its authors wrote: “we conclude strongly that lowering intake of saturated fat and replacing it with unsaturated fats, especially polyunsaturated fats, will lower the incidence of CVD”.

Two expert groups reading overlapping evidence reached different strengths of conclusion. Neither found that seed oils raise cardiovascular risk.

Guidelines of this kind are also about swaps. The advisory’s claim concerns oil replacing butter, lard or fatty meat. It makes no claim that adding oil to an unchanged diet helps. How much fat of any kind suits a person with heart disease or high cholesterol depends on their overall risk, which their doctor or dietitian assesses.

Where the seed oil evidence runs out

The cohorts are observational, and people who eat more linoleic acid differ from those who eat less in ways no analysis fully captures. The trials that could settle the matter are mostly fifty years old, small by modern standards, and were run in institutions on diets and margarines that no longer exist. A large modern trial of seed oils with deaths as its outcome has not been done.

The studies also measure the fat, and people eat foods. A large share of the linoleic acid in a modern diet arrives in fried and packaged products, so someone with a high intake may be cooking at home with sunflower oil or eating a great deal of fast food. Research on linoleic acid in blood does not distinguish between them, and the health records of those two diets are unlikely to be the same. Questions about oils that are heated repeatedly for deep frying, or about how oils are refined, sit outside what these studies examined.

The evidence that exists does not show seed oils harming the heart or raising inflammation, the long-term studies lean toward benefit, and the trials have yet to prove that benefit in lives saved.