Abstract
BACKGROUND: Dietary fat quality and carbohydrate processing shape lipid and lipoprotein profiles involved in skin barrier integrity and cutaneous inflammation relevant to atopic dermatitis (AD). We assessed the causal relevance of dietary patterns and food items to AD and mapped lipid-lipoprotein mediators. METHODS: We conducted a two-sample Mendelian randomization using genome-wide significant instruments for 83 UK Biobank diet traits and 241 serum lipid/lipoprotein measures, with AD cases from FinnGen R10 (European ancestry). Primary analyses used inverse-variance weighted MR with extensive sensitivity analyses, false discovery rate control, reverse MR, and multivariable MR. Mediation was assessed using the product-of-coefficients approach. Instrument strength was adequate (median F > 10). RESULTS: Using two-step Mendelian randomization, we identified specific dietary items with causal effects on AD risk. Notably, a dietary pattern characterized by higher unsaturated fats-exemplified by the protective effect of "other oil‑based spreads"-was associated with lower AD risk (OR = 0.56, 95% CI 0.34-0.93, P = 0.023). Conversely, a pattern reflecting refined-grain intake, represented by the risk-increasing effect of "brown bread", was associated with higher AD risk (OR = 1.78, 95% CI 1.10-2.89, P = 0.01). Mediation analyses mapped the underlying lipid pathways: sphingomyelin SM C20:2 mediated 15.9% of the protective effect of oil-based spreads (β (mediation) = -0.09, P = 0.003), and VLDL particle measures mediated 8.9% of the risk associated with brown bread (β (mediation) = 0.05, P = 0.003). A complex antagonistic mediation was observed for muesli via phosphatidylcholine PC aa C36:0 (proportion mediated: -13.6%, P < 0.001). Reverse MR analyses supported the proposed direction of causality (all P > 0.05), and findings were robust across sensitivity analyses. CONCLUSION: Dietary patterns high in unsaturated fats, particularly oil-based spreads, appear protective against AD, while refined-grain intake, especially brown bread and black bread, increases AD risk. These effects are mediated through lipid pathways involving sphingomyelins and VLDL metabolism, highlighting modifiable nutritional targets for AD prevention and adjunctive management.