1994-LB: Metabolites Link Infections to Type 2 Diabetes in the Multiethnic Study of Atherosclerosis (MESA)

Introduction & Objective: Prior to T2D, patients experience increased infections, yet no pathways account for this immune/endocrine system cross-talk. We aimed to identify metabolites that link pathogens to T2D.

Methods: Using plasma from 4,424 adults (61 ±9 y; 53% female), taxonomic classification was conducted on reads from sequencing that aligned with a microbial reference panel, and metabolites quantified. Incident T2D (ADA 2003 criteria) was assessed over ~18 y. A cox-proportional hazard model identified baseline microbes as ‘variables of importance’ to incident T2D, with a penalty function to minimize sparse data bias. Metabolome-wide associations identified metabolites shared between T2D-related microbes at baseline, and incident T2D. All models adjusted for age, gender, race/ethnicity, education, and AHA’s Life’s Simple 8 score.

Results: Total microbes differed by race/ethnicity (X²= 30.4, df = 9, P= 3.8*10^-4), but not any other covariates (all P>.05). Incident T2D was associated with 2 microbes: Rahnella (HR=1.90 (1.09-3.30), P=.02) and Plasmodium falciparum (HR=1.34 (1.05-1.73), P=.02). 23 metabolites were associated with incident T2D (P<.05 after an FDR-correction) and at least one of these microbes (P<.05; Fig).

Conclusions: These analyses indicate molecules linking exposure to, and / or continued presence in the plasma of, infectious microbes, and increased risk of T2D.

View largeDownload slide

View largeDownload slide

Close modal