A multi-metabolite signature robustly predicts long-term mortality in the PREDIMED trial and several US cohorts

Gonzalo Fernández-Duval*, Cristina Razquin, Fenglei Wang, Huan Yun, Jie Hu, Marta Guasch-Ferré, Kathryn Rexrode, Raji Balasubramanian, Jesús García-Gavilán, Miguel Ruiz-Canela, Clary B. Clish, Dolores Corella, Enrique Gómez-Gracia, Miquel Fiol, Ramón Estruch, José Lapetra, Montse Fitó, Luis Serra-Majem, Emilio Ros, Liming LiangCourtney Dennis, Eva M. Asensio, Olga Castañer, Francisco J. Planes, Jordi Salas-Salvadó, Frank B. Hu, Estefanía Toledo, Miguel A. Martínez-González

*Corresponding author af dette arbejde

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

1 Citationer (Scopus)

Abstract

Metabolome-based biomarkers contribute to identify mechanisms of disease and to a better understanding of overall mortality. In a long-term follow-up subsample (n = 1878) of the PREDIMED trial, among 337 candidate baseline plasma metabolites repeatedly assessed at baseline and after 1 year, 38 plasma metabolites were identified as predictors of all-cause mortality. Gamma-amino-butyric acid (GABA), homoarginine, serine, creatine, 1-methylnicotinamide and a set of sphingomyelins, plasmalogens, phosphatidylethanolamines and cholesterol esters were inversely associated with all-cause mortality, whereas plasma dimethylguanidino valeric acid (DMGV), choline, short and long-chain acylcarnitines, 4-acetamidobutanoate, pseudouridine, 7-methylguanine, N6-acetyllysine, phenylacetylglutamine and creatinine were associated with higher mortality. The multi-metabolite signature created as a linear combination of these selected metabolites, also showed a strong association with all-cause mortality using plasma samples collected at 1-year follow-up in PREDIMED. This association was subsequently confirmed in 4 independent American cohorts, validating the signature as a consistent predictor of all-cause mortality across diverse populations.

OriginalsprogEngelsk
Artikelnummer156195
TidsskriftMetabolism: Clinical and Experimental
Vol/bind170
Antal sider15
ISSN0026-0495
DOI
StatusUdgivet - 2025

Bibliografisk note

Publisher Copyright:
© 2025 Elsevier Inc.

Citationsformater