Association of the EAT-Lancet diet, serial measures of serum proteome and gut microbiome, and cardiometabolic health: a prospective study of Chinese middle-aged and elderly adults.

BACKGROUND

The EAT-Lancet diet was reported to be mutually beneficial for the human cardiometabolic system and planetary health. However, mechanistic evidence linking the EAT-Lancet diet and human cardiometabolic health is lacking.

OBJECTIVES

We aimed to investigate the role of blood proteins in the association between the EAT-Lancet diet and cardiometabolic health and explore the underlying gut microbiota-blood protein interplay.

METHODS

Our study was based on a prospective cohort including 3742 Chinese participants enrolled from 2008-2013 with serum proteome data repeatedly measured ≤3 times (N = 7514) and 1195 with gut metagenomic data measured ≤2 times over 9 y (N = 1695). Least absolute shrinkage and selection operator and multivariable linear regression were used to explore the associations of the EAT-Lancet diet (assessed by semi-quantitative food frequency questionnaire) with serum proteins and gut microbes. Linear mixed-effect model and logistic regression were used to examine the associations of selected proteins with 11 cardiometabolic risk factors and 4 cardiometabolic diseases, respectively. Mediation analysis was used to identify potential mediation effects. Multiple comparisons were adjusted using the Benjamini-Hochberg method.

RESULTS

The mean (standard deviation) age of enrolled participants was 58.4 (6.1) y (31.6% men). The EAT-Lancet diet was prospectively associated with 4 core proteins, including α-2-macroglobulin (A2M) (pooled β: 0.12; 95% confidence interval [CI]: 0.05, 0.2), retinol-binding protein 4 (pooled β: -0.14; 95% CI: -0.24, -0.04), TBC1 domain family member 31 (pooled β: -0.11; 95% CI: -0.22, 0), and adenylate kinase 4 (pooled β: -0.19; 95% CI: -0.3, -0.08). The identified proteins were prospectively associated with cardiometabolic diseases (pooled odds ratio ranged from 0.8-1.18) and risk factors (pooled β ranged from -0.1 to 0.12), mediating the association between the EAT-Lancet diet and blood triglycerides. We then identified 5 gut microbial biomarkers of the EAT-Lancet diet, and discovered a potential gut microbiota-blood protein interplay (EAT-Lancet diet→Rothia mucilaginosa→A2M) underlying the EAT-Lancet diet-cardiometabolic health association.

CONCLUSIONS

Our study presents key molecular evidence to support the role of EAT-Lancet diet adherence in promoting cardiometabolic health.