Seed oils are harmful to health

Cochrane review of 15 randomized controlled trials (16 comparisons, 56,675 participants, minimum 2 years) that reduced saturated fat intake and replaced it with polyunsaturated fat, carbohydrate, monounsaturated fat or protein. Cutting saturated fat reduced combined cardiovascular events by 17% (RR 0.83, 95% CI 0.70-0.98; 12 trials, 53,758 participants), with the effect scaling to the degree of cholesterol lowering, but made little or no difference to cardiovascular mortality (RR 0.95, 95% CI 0.80-1.12) or all-cause mortality (RR 0.96, 95% CI 0.90-1.03). Subgrouping did not show a significant difference between replacing saturated fat with polyunsaturated fat versus with carbohydrate; data for monounsaturated fat and protein replacement were very limited. This review incorporates the recovered Minnesota and Sydney trial data.

Prospective cohort analysis of 221,054 US adults from the Nurses' Health Study, NHS II and the Health Professionals Follow-up Study, with repeated diet assessments over up to 33 years and 50,932 deaths recorded. Highest versus lowest butter intake was associated with 15% higher total mortality (HR 1.15, 95% CI 1.08-1.22), while highest versus lowest total plant-oil intake was associated with 16% lower total mortality (HR 0.84, 95% CI 0.79-0.90). Per 5 g/day, the seed oils specifically were protective, not harmful: canola HR 0.85 (0.78-0.92), soybean HR 0.94 (0.91-0.96), olive HR 0.92 (0.91-0.94). Substituting 10 g/day of butter with plant oils was associated with 17% lower total mortality (HR 0.83, 0.79-0.86) and 17% lower cancer mortality. Observational and reliant on food-frequency questionnaires; plant-oil users differed from butter users on many health behaviours despite adjustment.

Cochrane review of 19 randomized controlled trials (6,461 adults, 1-8 years) comparing higher versus lower intake of omega-6 fats — mainly linoleic acid, the dominant fat in seed oils. Increasing omega-6 intake made little or no difference to all-cause mortality (RR 1.00, 95% CI 0.88-1.12; low-quality evidence), cardiovascular events (RR 0.97, 95% CI 0.81-1.15; low quality) or cardiovascular mortality (RR 1.09, 95% CI 0.76-1.55; very low quality), and may slightly reduce myocardial infarction (RR 0.88, 95% CI 0.76-1.02; low quality, NNT ≈ 53). The review is the most direct randomized test of raising omega-6 intake itself: it finds neither the harm claimed by critics nor a clear mortality benefit, and the authors rate the certainty of evidence as low throughout.

Individual-participant pooled analysis of 30 prospective cohorts across 13 countries — 68,659 participants, 15,198 incident cardiovascular events, follow-up medians ranging 2.5 to 31.9 years — using measured circulating and tissue levels of linoleic acid rather than self-reported diet, which removes food-frequency-questionnaire recall error. Higher linoleic acid was significantly associated with lower risk of total cardiovascular disease (HR 0.93, 95% CI 0.88-0.99), cardiovascular mortality (0.78, 0.70-0.85) and ischemic stroke (0.88, 0.79-0.98) — hazard ratios expressed per interquintile range, i.e. across the range between the midpoints of the first and fifth quintiles — and non-significantly with lower coronary heart disease risk (0.94, 0.88-1.00). Relationships were similar when linoleic acid was evaluated across quintiles. Arachidonic acid, the omega-6 metabolite alleged to drive inflammation, was not associated with higher cardiovascular risk; comparing extreme quintiles, higher arachidonic acid was associated with lower total CVD (0.92, 0.86-0.99). Observational: biomarker levels partly reflect metabolism and overall dietary pattern, so residual confounding remains possible.

Meta-analysis of 8 randomized controlled trials (13,614 participants, 1,042 coronary heart disease events) in which saturated fat was replaced with polyunsaturated fat — predominantly linoleic-acid-rich vegetable oils — for at least one year. Pooled coronary heart disease risk fell 19% (RR 0.81, 95% CI 0.70-0.95, p=0.008), corresponding to roughly a 10% lower risk per 5% of energy swapped from saturated fat to PUFA (RR 0.90, 95% CI 0.83-0.97). Longer trials showed larger benefit. The analysis predates the recovery of the Minnesota and Sydney trial data, and its trial set has been criticized for including multifactorial interventions that changed more than the fat source.

Systematic review of 15 randomized controlled trials (8 parallel, 7 crossover) that raised dietary linoleic acid — the omega-6 fatty acid that dominates soybean, sunflower, corn and safflower oil — in healthy non-infant adults and measured biomarkers of chronic inflammation. Across C-reactive protein, fibrinogen, plasminogen activator inhibitor-1, cytokines, soluble vascular adhesion molecules and TNF-α, none of the trials reported a significant increase in any inflammatory marker. Two studies showed small shifts in prostaglandin and thromboxane metabolite excretion, which the authors judged not to indicate increased inflammation. Their conclusion: "virtually no evidence is available from randomized, controlled intervention studies among healthy, noninfant human beings to show that addition of LA to the diet increases the concentration of inflammatory markers."

Randomized single-blind crossover trial in 19 healthy men comparing a high-fat meal made with a repeatedly heated sunflower oil-palm olein blend (180°C over 10 days of frying; 23.8% polar oxidation compounds) against the same meal made with unheated oil (7.2% polar compounds), with vascular and metabolic measures taken over 4 hours. Despite roughly triple the oxidation products, the heated oil produced comparable postprandial flow-mediated dilation and arterial stiffness to the unheated oil, and comparable glucose and NEFA responses; the only significant difference was a lower postprandial triacylglycerol rise after the heated oil (p<0.001). Acute, single-meal, small and male-only — it says nothing about cumulative effects of years of eating reheated frying oil.