Oil on a plant-based diet: what the trials say about olive, canola and coconut

Yes. In the largest trial we read, a Mediterranean diet with free extra-virgin olive oil beat advice to cut fat on a combined count of heart attack, stroke and cardiovascular death. FDA gives olive oil only a qualified heart claim, and only if it replaces saturated fat without adding calories. The AHA advises against coconut oil, which raises LDL.

The two studies people reach for on this question point in opposite directions, and they followed different people. In Caldwell Esselstyn’s 2014 follow-up, 198 patients with established cardiovascular disease were counseled onto a plant diet with no added oil. The 177 who stuck with it had one major cardiac event the authors judged to be recurrent disease, a stroke, over a mean of 3.7 years. Nine of them did need stents or bypass surgery and 5 died of non-cardiac causes, none of which the authors counted as recurrence, while 13 of the 21 who didn’t stick with the diet had at least one adverse event. The diet had no meat, fish, fowl or dairy, which puts it close to vegan, and whether vegans start out ahead on the heart is the subject of our piece on whether vegans are healthier than meat eaters. PREDIMED, the Spanish trial, enrolled 7,447 people at high cardiovascular risk who had no cardiovascular disease at the start. It gave households in one group 1 liter of extra-virgin olive oil a week, aiming for 50 g a day or more, about 4 tablespoons, while the control group was told to keep vegetable oils, olive oil included, to 2 tablespoons a day or less. Against that control, the olive oil group’s hazard ratio for a major cardiovascular event was 0.69 (95% confidence interval 0.53 to 0.91). One says no oil, the other says 4 tablespoons a day, and the second also advised fish at least three times a week.

Where the oil-free rule comes from

Oil is fat and close to nothing else. USDA puts olive oil at 884 calories per 100 g, 100 g of it fat, which works out to about 119 calories a tablespoon and roughly 4,000 a pound (our arithmetic). Esselstyn’s paper puts the case against it in stronger terms. It says the western diet of added oils, dairy, meat, fowl, fish and sugary foods injures or impairs the lining of the arteries after each meal.

A 2000 study measured that kind of effect for olive oil. Ten healthy people ate five meals of 900 calories and 50 g of fat each. Three hours after the olive oil meal, flow-mediated dilation of the arm artery, a measure of how well the vessel lining works, had dropped 31%. The remaining four meals, made with canola oil, with salmon, with olive oil plus vitamins C and E, and with olive oil plus balsamic vinegar and salad, did not significantly reduce it. Judging by that effect after a meal, the authors concluded that the helpful parts of the Mediterranean and Lyon Diet Heart Study diets appeared to be antioxidant foods such as vegetables, fruits and vinegar, and fish and canola oils high in omega-3. In that study, then, olive oil eaten with salad and vinegar made no significant difference.

Esselstyn’s own paper is candid about its limits. The authors list self-selected, very determined patients, no control group, which makes causality hard to establish, and no prospective randomization. They close by saying the time is right for a controlled trial, and that meanwhile the data are strong enough for patients to be told about the option. The diet also changed far more than oil. Along with added oils it cut fish, meat, fowl, dairy, avocado, nuts, excess salt and sugary foods, and later caffeine and fructose as well. The paper doesn’t report what dropping the oil did on its own.

What the oil trials found, and where they stop

PREDIMED has its own history. Its first report, in 2013, was withdrawn after the team found protocol deviations in how some participants were assigned; the 2018 version reanalyzed the 7,447 people, aged 55 to 80, and got similar results after leaving out the 1,588 whose assignment was known or suspected to have gone wrong. In raw numbers, 96 of 2,543 people in the olive oil group had a primary event (3.8%), against 109 of 2,450 in the control group (4.4%), over a median 4.8 years. Taken one at a time, only stroke had a confidence interval clear of no effect, at a hazard ratio of 0.65 (0.44 to 0.95); heart attack and cardiovascular death did not. The nut group, 30 g of mixed nuts a day, did about as well overall, with a hazard ratio of 0.72. Nuts were on Esselstyn’s banned list, though PREDIMED’s nut group was also advised to use at least 4 tablespoons of olive oil a day.

The authors say changes in total fat on their food questionnaire were small, and the biggest differences between groups by the end were the supplied olive oil and nuts, plus fish and legumes. With the control group allowed up to 2 tablespoons of oil a day, the trial never set olive oil against none. Its Mediterranean groups were advised to eat fish at least three times a week and choose white meat over red, and the authors say whether the results apply to people at lower risk needs more research.

The American Heart Association’s 2017 advisory frames the evidence around what replaces saturated fat. It found that trials which cut saturated fat and replaced it with polyunsaturated vegetable oil reduced cardiovascular disease by about 30%, while replacing it with mostly refined carbohydrates and sugars did not. FDA’s olive oil claim says “limited and not conclusive” evidence suggests about 2 tablespoons (23 g) a day may reduce the risk of coronary heart disease, provided the oil replaces a similar amount of saturated fat and doesn’t increase total calories. Its 2018 claim covers oils with at least 70% oleic acid, which the petition said include olive oil, at about 1½ tablespoons (20 g) a day, on “supportive but not conclusive” evidence and with the same calorie condition. Both are qualified claims, short of the significant scientific agreement a full health claim needs. The new 2025 to 2030 Dietary Guidelines tell cooks to prioritize oils with essential fatty acids, such as olive oil, and in the next sentence offer butter or beef tallow as other options. That second sentence settles nothing for a vegan kitchen.

Which oil, and how much

Coconut oil is where I’d draw the line. By USDA’s figures it is about 82% saturated fat, against about 14% for olive oil and 7% for canola. A 2020 meta-analysis pooling 16 studies found coconut oil raised LDL cholesterol by about 10.5 mg/dL compared with nontropical vegetable oils. The AHA advisory advises against using it, because it raises LDL and has no known offsetting benefit, and cites a systematic review whose authors noted that 7 trials found no difference in how much coconut oil and butter, beef fat or palm oil raised LDL. Being a plant doesn’t buy it a pass.

Canola also carries ALA, the plant omega-3. A tablespoon holds 1.28 g, and the adequate intake NIH lists for adults is 1.1 g a day for women and 1.6 g for men. Flaxseed oil carries 7.26 g a tablespoon. The oil-free route has an answer here, and Esselstyn’s program used it, advising flax seed meal for omega-3 and omega-6; whole flaxseed gives 2.35 g of ALA per tablespoon. Whether ALA alone covers you is a separate fight, the one we have in our piece on whether vegans need an algae omega-3.

Some fat on a salad also changes how much of some carotenoids you absorb. In a 2004 study, seven people ate the same salad of spinach, romaine, cherry tomatoes and carrots with dressings made with 0, 6 or 28 g of canola oil. With the fat-free dressing, the alpha-carotene, beta-carotene and lycopene that turned up in chylomicrons, the particles that carry absorbed fat in the blood, was negligible; it rose with 6 g and rose again with 28 g. Seven people is a small study, and it tested oil, not nuts or avocado.

Then the calories, which is where the FDA condition bites. PREDIMED’s goal for the olive oil group was 50 g or more a day, about 440 calories (our arithmetic, from USDA’s 884 calories per 100 g), and the trial gave no calorie limit. FDA’s claim, at 2 tablespoons, asks that the oil not increase total daily calories. A free pour over a pan of vegetables is not a tablespoon.

Where I land

Oil is allowed, and I’d use it. Olive or canola, measured with a spoon rather than a glug, in place of something more saturated rather than in addition to dinner. If you’re following an oil-free plan your cardiologist set for you, that is a choice between you and them, and nothing here argues you out of it. Coconut oil I’d keep for the one recipe that won’t work without it.

The part I’m least sure about is what oil does for someone with little saturated fat left to replace. Both FDA claims are about oil standing in for saturated fat, and the AHA’s 30% is about polyunsaturated vegetable oil doing the same. Take the butter and meat off the plate and I can’t tell you how much is left for the oil to replace. PREDIMED’s participants were advised to eat fish and white meat, and none of the studies we read for this article tested oil inside a vegan diet. If your oil question is really a baking question, our guide to baking without butter gives the oil ratio for a cup of butter, and when applesauce does the job instead.