Plasma lipidome and risk of atrial fibrillation: results from the PREDIMED trial

• Authors: Alvaro Alonso, Emory University; E Toledo, University of Navarra; C Wittenbecher, Chalmers University of Technology; C Razquin, University of Navarra; M Ruiz-Canela, University of Navarra; CB Clish, Broad Inst MIT & Harvard; L Liang, Harvard TH Chan Sch Publ Hlth; P Hernandez-Alonso, Univ Rovira & Virgili; N Becerra-Tomas, Univ Rovira & Virgili; F Aros-Borau, Inst Salud Carlos III ISCIII; D Corella, Inst Salud Carlos III ISCIII; E Ros, Inst Salud Carlos III ISCIII; R Estruch, Inst Salud Carlos III ISCIII; A Garcia-Rodriguez, Univ Malaga; M Fito, Inst Salud Carlos III ISCIII; J Lapetra, Inst Salud Carlos III ISCIII; M Fiol, Inst Salud Carlos III ISCIII; AM Alonso-Gomez, Inst Salud Carlos III ISCIII; L Serra-Majem, Inst Salud Carlos III ISCIII; A Deik, Broad Inst MIT & Harvard; J Salas-Salvado, Inst Salud Carlos III ISCIII; FB Hu, Brigham & Womens Hosp; MA Martinez-Gonzalez, University of Navarra
• Journal Title: JOURNAL OF PHYSIOLOGY AND BIOCHEMISTRY
• Volume: Volume 79, Number 2
• Publisher: SPRINGER | 2023-04-01, Pages 355-364
• Type of Work: Article
• Publisher DOI: http://doi.org/10.1007/s13105-023-00958-0
• Keywords: Science & Technology; Life Sciences & Biomedicine; Biochemistry & Molecular Biology; Physiology; Atrial fibrillation; Metabolomics; Lipidomics; Nested case-control study; Dietary intervention

Abstract

The potential role of the lipidome in atrial fibrillation (AF) development is still widely unknown. We aimed to assess the association between lipidome profiles of the Prevención con Dieta Mediterránea (PREDIMED) trial participants and incidence of AF. We conducted a nested case–control study (512 incident centrally adjudicated AF cases and 735 controls matched by age, sex, and center). Baseline plasma lipids were profiled using a Nexera X2 U-HPLC system coupled to an Exactive Plus orbitrap mass spectrometer. We estimated the association between 216 individual lipids and AF using multivariable conditional logistic regression and adjusted the p values for multiple testing. We also examined the joint association of lipid clusters with AF incidence. Hitherto, we estimated the lipidomics network, used machine learning to select important network-clusters and AF-predictive lipid patterns, and summarized the joint association of these lipid patterns weighted scores. Finally, we addressed the possible interaction by the randomized dietary intervention. Forty-one individual lipids were associated with AF at the nominal level (p < 0.05), but no longer after adjustment for multiple-testing. However, the network-based score identified with a robust data-driven lipid network showed a multivariable-adjusted ORper+1SD of 1.32 (95% confidence interval: 1.16–1.51; p < 0.001). The score included PC plasmalogens and PE plasmalogens, palmitoyl-EA, cholesterol, CE 16:0, PC 36:4;O, and TG 53:3. No interaction with the dietary intervention was found. A multilipid score, primarily made up of plasmalogens, was associated with an increased risk of AF. Future studies are needed to get further insights into the lipidome role on AF. Current Controlled Trials number, ISRCTN35739639.