Nutrients · Nov 16, 2026
Seed Oils, Dietary Linoleic Acid, and Omega-6 Biomarkers in Adult Health: An Umbrella Review of Systematic Reviews and Meta-Analyses
By Clemente-Suárez VJ, Yáñez-Sepúlveda R, López-Gil JF et al.
Background/Objectives: Seed oils are widely consumed, yet claims linking linoleic acid (LA) and omega-6 polyunsaturated fatty acids (n-6 PUFAs) to inflammation and chronic disease remain contentious. We evaluated the evidence in adults while separating named oils, dietary fatty acids, and circulating/tissue biomarkers. Methods: PubMed/MEDLINE and Europe PMC were searched to 1 October 2026, with supplementary and citation searches. Adult systematic reviews/meta-analyses were eligible; pregnancy and pediatric populations were excluded. Every synthesis was appraised item by item with AMSTAR 2 (A MeaSurement Tool to Assess systematic Reviews 2) and ROBIS (Risk Of Bias In Systematic reviews) and assigned a confidence tier by a rule defined post hoc, during peer review. Overlap was quantified with the corrected covered area; sensitivity and subgroup analyses examined the Sydney Diet Heart Study, the Minnesota Coronary Experiment, and LA-specific trials. No meta-meta-analysis was performed. The review was not prospectively registered. Results: Thirty-three syntheses were included; the six of higher confidence were not independent (four from one World Health Organization-commissioned program sharing trials). Clinical-endpoint evidence concerned LA/n-6 exposures, largely biomarker-based; named-oil evidence came from short-term trials of intermediate outcomes. Randomized evidence increasing omega-6 showed little or no effect on all-cause mortality (risk ratio [RR] 1.00, 95% confidence interval [CI] 0.88–1.12; robust to both trials) or cardiovascular events (RR 0.97, 0.81–1.15; heterogeneity mainly from the Sydney trial), with low certainty; in a hypothesis-generating subgroup, LA-specific trials showed more cardiovascular events than mixed n-6/omega-3 (n-3) trials (RR 1.45 vs. 0.80; two trials each). Prospective dietary/circulating evidence generally showed inverse associations, mostly from overlapping cohorts. LA-rich oils lowered low-density lipoprotein cholesterol when replacing butter (sunflower −0.34 mmol/L per 10% of energy) but differed little from other unsaturated oils, without increasing major inflammatory markers. Cancer findings varied by site and exposure metric, including an ungraded positive dietary LA–colorectal cancer association with an imprecise tissue LA estimate. Conclusions: In adults, the evidence does not support a uniform harmful or beneficial effect of seed oils as a class. Interpretation depends on the oil, replacement nutrient, background n-3 intake, exposure metric, outcome, and design; most evidence is of low certainty, and long-term clinical-outcome evidence for individual oils remains limited.