“Seed oils are toxic”

"Toxic" is a strong word, and the wrong one. The claim still deserves taking apart rather than waving away, because it rests on three separate ideas that are not equally wrong.

The first is inflammation. Linoleic acid, the main omega-6 fat - one of the two families of polyunsaturated fat in food - is supposed to inflame the body. Thirty randomised trials in 1,377 adults tested that directly, feeding people more of it and then measuring the markers doctors use to detect inflammation [2]:

• C-reactive protein, the standard blood marker of inflammation
• tumour necrosis factor and interleukin-6, two signalling molecules the immune system releases
• fibrinogen and plasminogen activator inhibitor, two proteins involved in blood clotting
• the adhesion molecules that let immune cells stick to blood vessel walls

None of them changed [2]. The one hint of an exception came from the people whose intake rose furthest, where C-reactive protein edged upwards, and the authors treated that as a question for later work rather than a result [2].

Then there is the idea that eating a lot of it shortens life. That has been examined in 44 groups of adults totalling 811,069 people, and the association runs the other way. Set against those eating the least linoleic acid, the people eating the most were about 13% less likely to die of any cause, about 13% less likely to die of heart disease and about 11% less likely to die of cancer [1]. Blood and fat-tissue samples from a further 65,411 people pointed the same way, and those measurements do not depend on anyone remembering what they ate [1].

The third is oxidation, the idea that heating the oil turns it into something harmful. Here the human evidence barely exists. Repeatedly reheated frying oil does change chemically, but the research on it runs on hamsters, rats and cell cultures. We found no trial that followed people eating food fried in reused oil and counted what happened to them. That is not a clean bill of health for reused frying oil, but it is an open question rather than the settled one the claim assumes.

There is a real weakness in the seed-oil case, and it sits on the other side of the argument. The Minnesota Coronary Experiment randomly assigned 9,423 adults in six state hospitals and a nursing home to a corn-oil diet or a saturated-fat diet. Cholesterol dropped 13.8% on corn oil against 1.0% on the control diet, and the death rate did not improve - not in the group as a whole, and not in any subgroup the original investigators had planned to examine [3]. The trial itself ran in 1968-73, and its results were not published until 2016.

So the swap lowers cholesterol without a demonstrated effect on how long people live, and anyone selling seed oil as proven life extension is going past what the trials found. That is a real limit on what the oil can be sold as. It is a separate question from whether the oil harms you.

Who this evidence does not describe:

• anyone eating food fried in oil that has been reused repeatedly, which none of this research followed
• people eating far outside the ordinary dietary range, since the inflammation trials fed normal amounts of linoleic acid [2]
• anyone after a seed-oil-versus-olive-oil answer, which none of these studies were built to give

One caution on reading the mortality figures. They come from watching populations rather than assigning people to diets, so those eating more linoleic acid differ from those eating less in ways beyond the oil, and the studies could only adjust for what they measured [1]. The randomised evidence does not have that problem, and pooling the Minnesota trial with four others covering 10,808 people between them, replacing saturated fat with linoleic acid moved the death rate in neither direction [3].