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 multivariableadjusted
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.
; Fitó, Montserrat; Lapetra, José; Fiol, Miquel; Alonso-Gómez, Ángel M.; Serra-Majem, Lluís; Deik, Amy; Salas-Salvadó, Jordi; Hu, Frank B.; Martínez González, Miguel Ángel
