Are seed oils bad for you? Seed oils, butter and beef tallow

The objection

“Seed oils are an industrial product our great-grandparents never ate. Soybean oil consumption in America rose more than a thousandfold over the twentieth century, and the diseases of modern life rose right alongside it. Omega-6 fats fuel inflammation. When scientists actually tested vegetable oil against animal fat, in Sydney and Minnesota, more people died on the oil, and those results went unpublished for decades. People cooked in butter and beef dripping for generations, and the war on saturated fat was a mistake.”

The answer

Seed oils are the oils made from seeds such as rapeseed, sunflower, soybean, corn and safflower, and the evidence from people does not show them harming the heart. People with more of their omega-6 fat in their bodies have had less cardiovascular disease [1], and trials that replaced saturated fat with polyunsaturated oils cut cardiovascular events, most clearly when the oils also supplied omega-3, as soybean oil does [6][2]. Most seed oils are rich in linoleic acid, the main omega-6 fat in food [1][2]. The worry about them starts from real history. Working from food supply data, researchers estimated in 2011 that Americans’ consumption of soybean oil per person rose more than 1,000-fold between 1909 and 1999, lifting the linoleic acid in the diet from 2.79% to 7.21% of energy, and they judged that this had probably lowered the body’s stores of the long-chain omega-3 fats EPA and DHA (Blasbalg et al., 2011) [3]. A trend over 90 years can’t show what caused any disease, though, because so much else about American food and life changed over the same period.

The claim has since moved from social media into government. Harvard Health Publishing reported in 2026 that the US health secretary, Robert F. Kennedy Jr, has called seed oils poisonous, and that the latest federal dietary guidelines refer to beef tallow and butter as healthy [4]. The Dietary Guidelines for Americans for 2025 to 2030 tell people to prioritise oils with essential fatty acids, such as olive oil, list butter or beef tallow as other options and give no guidance on seed oils, while keeping the long-standing limit of saturated fat to no more than 10% of daily calories [5][22][23]. A 2026 review in the journal Lipids says their advice to eat butter, beef tallow and full-fat dairy does not fit that limit (Brenna et al., 2026) [5]. The government’s realfood.gov site, which presents the guidelines, describes real food as prepared without added sugars or “industrial oils”, a term it does not define [21].

In 2019 a global consortium pooled 30 cohort studies from 13 countries, with 68,659 people and 15,198 cardiovascular events, measuring linoleic acid in their blood or body fat rather than asking what they ate (Marklund et al., 2019) [1]. The analysis was partly paid for by a restricted grant from Unilever, which the authors say had no role in its design, analysis or writing [1]. Comparing people in the middle of the top fifth with those in the middle of the bottom fifth, more linoleic acid went with a 7% lower risk of cardiovascular disease in the 21 cohorts that recorded it (hazard ratio 0.93, 95% CI 0.88 to 0.99), a 22% lower risk of dying from it in the 17 that recorded cardiovascular deaths (0.78, 0.70 to 0.85) and a 12% lower risk of ischaemic stroke (0.88, 0.79 to 0.98), while its link with less coronary heart disease fell just short of significance (0.94, 0.88 to 1.00) [1]. These are observational links, so people with more linoleic acid may have differed in other ways, but they run in the opposite direction to the one the case against seed oils predicts.

Trials that cut saturated fat point the same way. A 2020 Cochrane review of 15 randomised trials with 56,675 participants, each lasting at least two years, found that cutting saturated fat lowered combined cardiovascular events by 17% (risk ratio 0.83, 0.70 to 0.98), and judged replacing it with polyunsaturated fat or carbohydrate likely to be useful (Hooper et al., 2020) [6]. The American Heart Association’s 2017 advisory on dietary fats gave a higher estimate than Cochrane’s, saying trials that swapped saturated fat for polyunsaturated vegetable oil cut cardiovascular disease by about 30%, similar to the effect of statins, while swapping it for refined carbohydrates and sugars did not lower it (Sacks et al., 2017) [7]. In the UK, the Scientific Advisory Committee on Nutrition reviewed 47 systematic reviews and meta-analyses in 2019 and advised swapping saturated fat for unsaturated fats and keeping it to no more than about 10% of food energy, and Public Health England’s release on the report suggested cooking with oils instead of butter [8].

Do seed oils cause inflammation?

Seed oils have not raised inflammation in the trials that tested their main omega-6 fat, linoleic acid, in healthy people. The concern is that the body turns linoleic acid into arachidonic acid, from which it makes some inflammatory signalling molecules [1][9]. A 2012 systematic review of 15 randomised trials in healthy people found that none reported a significant rise in any of a wide range of inflammatory markers, C-reactive protein and tumour necrosis factor among them, when people ate more linoleic acid (Johnson and Fritsche, 2012) [9]. In the 2019 pooled cohorts, people with more arachidonic acid in their blood or tissue did not have more cardiovascular disease, and those in the top fifth had less of it than those in the bottom fifth (hazard ratio 0.92, 0.86 to 0.99) [1]. The NHS says most people already get enough omega-6 from their diet, and counts polyunsaturated fats among those that can help lower LDL cholesterol [10].

Did the Sydney and Minnesota trials find vegetable oil raised deaths?

The Sydney trial found more deaths on vegetable oil and the Minnesota trial found no fewer, in data recovered decades later. These are the strongest evidence the critics have, though the wider body of trials found fewer cardiovascular events from cutting saturated fat and no clear effect on deaths either way. The Sydney Diet Heart Study, run from 1966 to 1973, gave 458 men who had recently had a coronary event safflower oil and safflower margarine in place of animal fats, common margarines and shortenings, while a control group had no specific dietary advice [2]. When Ramsden and colleagues recovered and reanalysed the data in 2013, 17.6% of the oil group had died against 11.8% of the controls (hazard ratio 1.62, 1.00 to 2.64) [2]. The Minnesota Coronary Experiment, from 1968 to 1973, put 9,423 people in six state mental hospitals and a nursing home on corn oil and corn oil margarine or on a diet high in animal fat [11]. Cholesterol fell by 13.8% on the oil diet against 1.0% on the control diet, but deaths did not fall, and in the recovered data each 30 mg/dL (0.78 mmol/L) drop in cholesterol went with a 22% higher risk of death (hazard ratio 1.22, 1.14 to 1.32; Ramsden et al., 2016) [11].

The reanalysers set out the limits themselves. Minnesota followed patients only while they were in hospital, only about a quarter stayed in the study for a year or more, and its link between falling cholesterol and death is observational in nature even though it comes from a trial [11]. In Sydney the extra deaths were most evident in smokers and heavy drinkers, a pattern the incomplete Minnesota data can’t test [11]. Safflower oil carries linoleic acid and no omega-3, and when Ramsden’s team pooled the Sydney trial with two others that swapped animal fat for corn oil, the rise in coronary deaths fell just short of significance (1.33, 0.99 to 1.79), while four trials that raised omega-3 along with omega-6 cut cardiovascular deaths by 21% (0.79, 0.63 to 0.99) (Ramsden et al., 2013) [2]. Their 2016 meta-analysis of five trials of oil rich in linoleic acid found no benefit for coronary deaths (1.13, 0.83 to 1.54) [11], but the 2020 Cochrane review, which counts the Sydney trial among its 15, still found 17% fewer cardiovascular events when saturated fat was cut [6].

Is saturated fat actually bad for you?

Too much saturated fat is bad for the heart, because it raises LDL cholesterol, and LDL cholesterol causes heart disease. A 2017 consensus statement from the European Atherosclerosis Society, drawing on meta-analyses of more than 200 cohort studies, Mendelian randomisation studies and trials with over 2 million participants, concluded that the evidence “unequivocally establishes” that LDL causes atherosclerotic cardiovascular disease, with risk rising with both the level of LDL and how long it lasts (Ference et al., 2017) [12]. A 2018 network meta-analysis of 54 trials found that, for each 10% of calories swapped, safflower, sunflower, rapeseed, flaxseed, corn, olive and soybean oil all lowered LDL cholesterol by 0.23 to 0.42 mmol/L more than butter did, and so did palm oil, coconut oil and beef fat (Schwingshackl et al., 2018) [13]. Safflower, sunflower, rapeseed, corn and soybean oil also lowered it by 0.20 to 0.33 mmol/L more than lard, though the authors rated the evidence for most comparisons low or moderate [13].

Beef tallow comes out better than butter in that analysis, which fits a 1994 review: beef fat is 19% stearic acid, a saturated fat that raises cholesterol less than the others, so it raises cholesterol less than its total saturated fat would suggest. The same review still found tallow raises cholesterol compared with fats that have less of the cholesterol-raising kinds, and advised cutting it down in a cholesterol-lowering diet (Denke, 1994) [14]. Harvard Health Publishing (2026) puts tallow at more than 50% saturated fat and says several randomised trials have found LDL rising after people ate it [4]. Whether you need any animal fat at all is covered in Do vegans need collagen or animal fat?.

A study of butter against plant oils that followed 221,054 American nurses and health professionals for up to 33 years found that those who ate the most butter had a 15% higher death rate than those who ate the least (hazard ratio 1.15, 1.08 to 1.22), and those who used the most plant oil, counting soybean, canola, olive, corn and safflower oil together, a 16% lower one (0.84, 0.79 to 0.90) (Zhang et al., 2025) [15]. Swapping 10 g of butter a day for plant oils went with 17% lower death rates overall (0.83, 0.79 to 0.86) and from cancer (0.83, 0.76 to 0.90), though the 6% lower rate of cardiovascular death was not statistically significant (0.94, 0.86 to 1.03) [15]. The authors note that their mostly white, health-professional volunteers limit how far the findings generalise and that some confounding may remain [15].

Sceptics are right that deaths have been harder to shift. The Cochrane review found little or no effect of cutting saturated fat on deaths from any cause (risk ratio 0.96, 0.90 to 1.03) or from cardiovascular disease (0.95, 0.80 to 1.12), and estimated that among people without heart disease, 56 would need to cut their saturated fat for about four years for one to avoid a cardiovascular event [6]. Saturated fat from plants gets no exemption either, which matters on a vegan keto diet heavy in coconut: in a 2020 meta-analysis of 16 trials, coconut oil raised LDL cholesterol by 10.47 mg/dL compared with non-tropical vegetable oils (Neelakantan et al., 2020) [16]. If you take a statin or another medicine for your cholesterol or heart, changing the fat you cook with is no reason to stop it, and you should speak to your GP before changing or stopping any medicine.

Is rapeseed oil unhealthy?

Rapeseed oil does better than most cooking fats in the studies above, and it has less omega-6 and more omega-3 than sunflower, corn or soybean oil. In US food composition figures tabulated by Ramsden and colleagues in 2013, 100 g of canola oil had 18.6 g of linoleic acid and 9.1 g of the plant omega-3 ALA, against 65.7 g of linoleic acid and no ALA in sunflower oil and 50.3 g and 7.0 g in soybean oil [2]. Canola is the name Canada registered in 1979 for rapeseed bred to be low in erucic acid and glucosinolates, according to a 2013 review funded by the canola industry [17]. In the 2018 network meta-analysis, rapeseed oil ranked second of the 13 fats for lowering LDL cholesterol, behind safflower oil [13]. In the same US cohorts that compared butter with plant oils, each extra 5 g of canola oil a day went with a 15% lower death rate (hazard ratio 0.85, 0.78 to 0.92), and soybean and olive oil were also linked to lower death rates, while safflower oil showed no clear link and corn oil a borderline trend towards higher risk (1.07, 0.99 to 1.15) [15]. The NHS lists rapeseed oil among the sources of unsaturated fat [10].

Heat is one place where the worry about processed oils has a real basis. A 2007 review of the chemistry of deep-fat frying found that frying breaks oil down through oxidation, hydrolysis and polymerisation, and that high temperatures, repeated frying and a high content of unsaturated fat all make oil less stable, so the more unsaturated the oil and the more often it is reused, the faster it degrades (Choe and Min, 2007) [24]. The review dealt with the oil’s chemistry and flavour and gave no figures for effects on health [24].

Should you cut out oil altogether?

There is no good evidence that you need to, and a large Spanish trial of a diet rich in added olive oil found fewer heart attacks, strokes and cardiovascular deaths, counted together, than on low-fat advice. PREDIMED assigned 7,447 Spanish adults aged 55 to 80 at high cardiovascular risk to a Mediterranean diet with extra-virgin olive oil, a Mediterranean diet with nuts, or advice to cut down on fat, and over a median of 4.8 years major cardiovascular events struck 3.8% of the olive oil group against 4.4% of the group given low-fat advice, a hazard ratio of 0.69 (0.53 to 0.91) (Estruch et al., 2018) [18]. Its original 2013 report was withdrawn after the authors found that some participants had not been properly randomised, and the 2018 reanalysis found similar results with the 1,588 people affected left out [18]. PREDIMED tested olive oil, so for seed oils the evidence is the trials and cohorts above.

Caldwell Esselstyn’s oil-free programme counselled 198 volunteers with cardiovascular disease to cut out meat, fish, dairy and added oil; of the 177 who kept to it for an average of 3.7 years, one had a recurrent major cardiac event, a stroke (Esselstyn et al., 2014) [19]. The study had no randomised control group and its patients changed far more than their oil, so it can’t show what oil itself does, as the answer on reversing heart disease explains. Oil is fat, which carries 9 kcal a gram against 4 for carbohydrate and protein, so it adds up quickly in cooking [10].