#380 ‒ The seed oil debate: are they uniquely harmful relative to other dietary fats? | Layne Norton, Ph.D.

Welcome back to The Drive; I’m Peter Attia, and our mission is to turn longevity science into practical tools.

Today’s guest is Lane Norton, a nutrition scientist and strength athlete, and we’re trying a twist on our usual format.

We built a courtroom‑style debate on seed oils with full evidence shared in advance, but the opposing guest withdrew, so I’ll steel‑man the anti–seed oil case and let Lane argue the counterpoint while you act as the jury.

We’ll tackle four lanes of argument: old randomized trials and mortality, LDL oxidation and mechanisms, how oils are processed, and the evolutionary narrative, all under isocaloric conditions.

Everyone carries bias, including me, so I try to name mine and lean on converging lines of evidence rather than single studies or social‑media hot takes.

I came up through a low‑carb lens and was funded by dairy, beef, and eggs, yet I changed my views as the data shifted, which is the point of doing science.

Let’s start with the Minnesota Coronary Experiment, a rare, long, controlled feeding study in institutional settings that cut saturated fat and raised polyunsaturated fat, dropped total cholesterol, but didn’t lower mortality.

That delayed publication made people wonder if swapping butter for oils could be harmful even when calories were held steady.

It’s a landmark, but a major confounder is trans fat, because the intervention used margarine that, at the time, contained roughly one‑quarter to two‑fifths trans fat, which is strongly atherogenic.

Quick primer: saturated fats have no double bonds and tend to be solid; unsaturated fats have double bonds; trans double bonds behave more like saturated fats in structure.

Those geometry differences change membrane fluidity in lipoproteins, which affects receptor recognition and aggregation, and trans fats deliver stiffness plus an oxidizable bond, a bad combination.

Margarine was engineered to mimic butter’s solidity at room temperature, but it ended up worse for the heart, which led to trans fats being effectively banned.

Consider the Sydney Diet Heart Study: men post‑MI raised safflower oil and margarine, and mortality rose in the intervention group within three years.

Again, the margarine carried trans fats, the sample had few total deaths with wide confidence intervals, and very sick cohorts can show reverse causality where low cholesterol tracks with wasting.

If someone downplays trans fats, what else explains the paradox of lower cholesterol without survival benefit?

Short duration blunts power for hard endpoints, and when you pool all RCTs with their mixed confounders the effect is null; pulling omega‑3–containing trials changes direction but still leaves trans fats in play.

The tiny Rose Corn Oil trial reported more cardiac deaths with corn oil and even more with olive oil than control, but the confidence intervals were enormous and diets were free‑living with provided oils.

It wasn’t confounded by trans fats, but it was too small to trust, and even the olive oil arm looked worse than control, which tells you noise dominated signal.

By contrast, meta‑analyses that exclude trans‑fat‑confounded trials show a clear benefit to replacing saturated fat with polyunsaturated fat, on the order of roughly twenty to thirty percent lower risk.

The Veterans study had controlled feeding and long follow‑up, showed an eighteen percent risk reduction that crossed unity, and included omega‑3s, raising the question of power versus confounding.

Single trials chasing mortality need huge samples, so we look at the totality: Oslo showed benefit with omega‑3s in the mix, and the Finnish hospital crossover trial—without trans fats or omega‑3s—found a large, durable drop in events.

In Finland, saturated fat fell from about eighteen to nine percent and polyunsaturated fat rose to around fourteen percent, and events dropped by roughly forty‑one percent with tight confidence intervals; adherence was biologically verified.

The better trials are longer, controlled, and free of trans fats, and their outcomes align with mechanisms and cohort data.

Nutrition is substitution, and the question is net effect; swapping in polyunsaturated fat for saturated fat generally improves risk, which can also happen with fiber‑rich carbs or monounsaturated fat.

Let’s pivot to mechanism and Mendelian randomization as a way to read lifetime LDL exposure, noting MR needs a genetic instrument tied only to the exposure and not directly to other outcome drivers.

MR acts like nature’s randomized trial, with variants that shift LDL over a lifetime and large samples to read risk; pleiotropy is tested, and while this isn’t a direct SFA‑versus‑PUFA trial, it models the LDL changes we expect when we replace one with the other.

Mendelian randomization helped settle one worry for me: LDL doesn’t seem to drive cancer, but it does track tightly with atherosclerotic disease.

Those MR data flipped my view on LDL; lifelong, randomized differences in LDL show a near-linear link to cardiovascular risk, and swapping saturated fat for polyunsaturated fat reliably lowers LDL and ApoB, with CETP being the notable discordance that highlights ApoB as the true signal.

Here’s the puzzle: MR suggests around a 50 to 55 percent event drop per roughly forty milligrams per deciliter LDL reduction, yet statins deliver about 22 percent—does that imply statins help but also harm?

It’s the clock, not a hidden toxin; starting a statin at midlife can’t erase decades of exposure, while being born with lower LDL compounds benefit across a lifetime, and the consistency holds across drugs, diet, and mechanisms.

Some argue statins work by anti-inflammatory effects rather than LDL lowering; do the total data refute that?

Dose-response and trial-to-trial consistency make the LDL-causality case strong, much like the robust, graded benefits seen with dietary fiber, whereas claims built on “what else could it be” don’t fit the pattern.

Shift to seed oils: linoleic acid feeds arachidonic acid, inflammation drives plaque, and oxidized LDL is dangerous—shouldn’t high linoleic diets raise risk?

Big-picture data show higher linoleic intake and tissue levels link to lower cardiovascular events, conversion to arachidonic acid doesn’t surge with intake, and the lipid model centers on ApoB particles entering, being retained, then oxidizing primarily inside the artery wall.

Even if only a little LDL is oxidized in blood, could that small fraction disproportionately seed the plaque?

Per particle, oxidized LDL is worse, but plasma levels are tiny and short-lived; oxidation ramps up after retention in the intima, so the bigger lever is fewer ApoB particles getting in and less aggregation, which polyunsaturated-rich particles favor through better fluidity, less ApoB modification, and less ceramide-driven clumping—think fewer sparks from a smaller fire.

Do peripheral oxidized LDL spikes come from plaques?

After heart attacks, circulating oxidized LDL jumps, likely from lesions, but tracing origins precisely is hard; the practical point remains that saturated-fat–enriched LDL aggregates more, which accelerates lesions.

What about industrial processing—hexane, heat, and residues in seed oils?

Hexane is used because it’s nonpolar and easy to strip with low-heat steam; residues are typically under one part per million or not detectable, toxicity is mainly from inhalation, and refining actually lowers peroxides and aldehydes, with only tiny trans fat formation.

Intake of linoleic acid has exploded, and tissue levels climbed; by first principles, how could that be good?

Appeals to what’s “natural” don’t predict longevity; we live longer, our foods and animals have changed, the Hadza’s low LDL suggests ancestral exposure wasn’t high, and modern trials show replacing saturated fat with polyunsaturated fat improves LDL and often insulin sensitivity, liver fat, and inflammation markers.

Given the social backlash against seed oils, what’s a cautious listener supposed to do?

If you skip seed oils, still displace saturated fat with lean proteins and monounsaturated fats like olive or avocado oil, and remember the true frying risk is thin, repeatedly heated oils that build reactive byproducts.

If I’m going to eat fries, is lard better than seed oil, and if I medicate LDL, should I care more about oxidation from cooking?

We lack head-to-head human trials on frying media; saturated fats oxidize less with heat but worsen lipids, so enjoy fries sparingly and don’t crown lard-fried food as healthy.

Beware health halos; spend your mindshare on calorie control, activity, and fiber, keep LDL and ApoB in check, and remember blood pressure, fitness, and insulin sensitivity carry major weight—risk is probabilistic, but lower LDL all else equal is safer.

Bottom line, we’re majoring in the minor; seed oils often tag along with ultra-processed foods, so cutting them can help by improving overall diet quality, but there’s no need to be dogmatic—use the dressing you like and please, restaurants, retire the “no seed oils” badge.

Thanks for the thoughtful deep dive; I enjoyed laying out the evidence.

Thanks for listening; show notes are on the site, you can find me on the usual channels, and please remember none of this is medical advice—work with your clinician.