The Truth About Seed Oils — the Honest, Balanced Version

If you've spent any time on health-focused social media in the last five years, you've encountered two positions on seed oils that don't agree:

Position A: Seed oils (canola, soybean, sunflower, corn, cottonseed, safflower) are toxic, drive heart disease, and should be avoided entirely.

Position B: Seed oils are fine, the evidence supports them, and the anti-seed-oil movement is anti-science.

Both positions misrepresent the literature. The honest answer is more nuanced and harder to communicate, which is why neither side has won. Here's what the research actually says, and a defensible practical position.

What "seed oils" actually means

The term refers to industrial oils extracted from seeds via solvent extraction (typically hexane), high heat, and chemical refinement. The category typically includes:

• Soybean oil
• Canola oil (rapeseed)
• Sunflower oil (refined)
• Corn oil
• Cottonseed oil
• Safflower oil (refined)
• Grapeseed oil
• Rice bran oil

What unites them: high omega-6 polyunsaturated fatty acid content (mostly linoleic acid), industrial processing, and ubiquity in modern food production. They're cheap, neutral-tasting, and shelf-stable.

What's typically excluded from "seed oils" as criticized: cold-pressed, minimally processed nut and seed oils used in small quantities (e.g., flax, chia), olive oil (a fruit oil, not a seed oil), avocado oil, butter, ghee, lard, tallow, coconut oil.

The contested evidence: Ramsden's recovered trials

The strongest case for caution comes from Christopher Ramsden's reanalysis of two old trials whose original data had been incomplete or unfavorably summarized.

The Sydney Diet Heart Study reanalysis (BMJ 2013): Original 1960s–70s trial replaced saturated fat with linoleic-rich oil (safflower-derived margarine) in 458 men with recent coronary events. Ramsden recovered the missing data and found the intervention group had higher all-cause mortality and CVD mortality than controls — despite lower cholesterol.

The Minnesota Coronary Experiment reanalysis (BMJ 2016): Original 1968–73 trial of >9,400 institutionalized patients tested the same swap. Ramsden's recovery found the intervention group had lower cholesterol but did not show reduced mortality; subgroup analysis suggested possible higher mortality in older participants.

These are not weak signals from observational studies. They're recovered data from RCTs — the gold standard — and they directly challenge the "lower saturated fat, replace with polyunsaturated, see CV benefit" narrative that has dominated US dietary guidelines since the 1977 McGovern report.

The contested evidence: meta-analyses on the other side

But the picture isn't one-sided. Hooper's 2020 Cochrane review remains the canonical reference for the mainstream position. Pooled analysis found cutting saturated fat reduced cardiovascular events ~17%, with the effect concentrated when the replacement food is polyunsaturated fat.

Chowdhury 2014 (Annals of Internal Medicine) ran a different meta-analysis and found no clear support for replacing saturated fat with polyunsaturated to reduce coronary risk. Dehghan and the PURE study group (Lancet 2017) followed 18 countries and found high carbohydrate intake associated with higher mortality, while high fat intake — including saturated — was not.

Astrup 2020 (JACC) assembled an international panel of researchers to reassess the evidence. Their conclusion: total saturated-fat intake, in itself, is not associated with CVD; food matrix (whole foods vs UPF) and the replacement food matter much more than the simple gram count of any single fatty acid.

This is what "contested" means in the literature. Neither side gets to claim the evidence definitively.

What the mechanism people argue

DiNicolantonio and O'Keefe (Open Heart 2018) propose the oxidized linoleic acid metabolite hypothesis: linoleic acid is highly susceptible to oxidation, oxidized linoleic acid metabolites (OXLAMs) accumulate in atherosclerotic plaque, and chronic high linoleic intake may drive inflammation and CVD via these metabolites rather than via cholesterol.

This is a mechanism paper. It's plausible. It's not yet the consensus. But it provides a coherent biological explanation for the Ramsden data: high seed-oil intake doesn't show up in cholesterol numbers (LDL drops on PUFA), but does show up in mortality if the OXLAM hypothesis is right.

The signal that's strongest: UPF, not seed oils per se

Here's where the literature actually converges, and where we land our position.

The Hall 2019 inpatient trial demonstrated that ad-libitum ultra-processed food intake drove ~500 kcal/day surplus over matched whole-food eating. UPF is heavily seed-oil-laden — restaurant fryers run on canola/soy, packaged snacks use high-oleic sunflower or palm, cooking sprays are seed-oil-based. So adults eating high-UPF diets are also eating high-seed-oil diets, and untangling which variable drives the bad outcomes is genuinely hard.

The Lane 2024 BMJ umbrella review of UPF outcomes found convincing evidence for cardiovascular mortality, T2D, depression, and certain cancers across 32 health outcomes. The strongest dietary signal in modern epidemiology is UPF — not seed oils specifically, not saturated fat, not any single nutrient.

This matters because it suggests the practical move is the same regardless of which side of the seed-oil debate you find more convincing: eat whole foods, drop UPF, and the seed-oil question largely answers itself.

Our editorial position

After reading the literature, this is where we land — clearly flagging it as a position rather than settled science:

1. The Ramsden data is real and significant. Strong RCT-derived evidence that swapping saturated fat for linoleic-rich oils didn't reduce mortality, and may have increased it in some populations. That's a meaningful signal that the simple "saturated fat bad, polyunsaturated good" model deserves substantial doubt.

2. The mechanism case (OXLAMs) is plausible but not yet established. We treat it as worth taking seriously without treating it as proven.

3. The strongest move is UPF reduction, not seed-oil obsession. UPF reduction captures most of the practical benefit because UPF is where most of your seed-oil intake comes from anyway. Cooking at home in olive oil, butter, ghee, lard, or tallow handles 80%+ of the question.

4. We don't claim seed oils are "toxic." That language overstates the evidence. We do think the precautionary principle favors avoiding them when reasonable alternatives exist — especially for high-frequency cooking fats and for adults with cardiovascular risk.

5. We don't recommend olive oil as a replacement for industrial seed oils based on saturated-fat fear. The Mediterranean evidence (PREDIMED, Lyon Diet Heart) is for the whole pattern — vegetables, legumes, fish, olive oil, modest wine — not for olive oil substituted into otherwise-poor diets.

Practical guidance

For home cooking, prefer (in our order):

• Olive oil (extra virgin) for low-medium heat
• Butter or ghee for medium heat
• Tallow, lard, or duck fat for high heat (these are the best high-heat fats by oxidation stability)
• Coconut oil for medium-high heat (high-MCT, good stability, distinctive flavor)
• Avocado oil if you want a neutral mid-high heat option

For restaurant eating:

• Most restaurants use seed oils for cost reasons. Selecting "less seed-oil" restaurant patterns means: protein cooked plain, vegetables steamed or grilled, salads dressed with oil + vinegar (request EVOO), avoid deep-fried items, avoid sauces with embedded oils.

• The 80/20 framework helps: control cooking fat at home, accept that occasional restaurant exposure isn't the lever that determines your metabolic outcome.

For packaged food:

• Read labels. "Vegetable oil," "soybean oil," "canola oil," "sunflower oil" (refined), "corn oil," "cottonseed oil" — these are seed oils. They're in nearly all UPF.

• Replacing UPF with whole foods does this work for you. You don't need to learn a label-reading discipline if you're not eating packaged products.

What we won't tell you

We won't tell you seed oils are "the cause" of obesity or chronic disease. The literature doesn't support a single-cause story, and the cultural framing that singles them out distracts from the much-stronger UPF and food-environment signals.

We also won't dismiss the people raising the seed-oil concern. Ramsden's data is real. The mechanism work is plausible. The precautionary stance is defensible. Mainstream medicine's full-throated endorsement of polyunsaturated fat replacement looks more dated each year.

The honest position is that we don't know with certainty, and reasonable people can read this evidence differently. What we can say with high confidence: eat whole foods, cook at home in stable fats, drop UPF, and most of the disagreement stops mattering for your day-to-day metabolism.