Abstract P434: Prediabetes and Diabetes Are Associated With Distinct Gut Microbial Composition in the Framingham Heart Study

Little is known about how the gut microbiome may be associated with impaired glucose regulation (prediabetes and type 2 diabetes). Relative to normoglycemic individuals, prediabetes and diabetes will be associated with distinct microbial composition, which will provide incremental value in classifying glycemic status. We included 1265 Framingham Heart Study participants (mean age 55 yr, 56% female) with 16S rRNA gene sequencing performed on stool samples. Glycemic status was characterized as normoglycemia (=126mg/dL glucose, >=6.5 HbA1c, or self-reported diabetes diagnosis/medication). MANOVA was used to characterize the differential abundance of microbial composition and diversity by glycemic status. Microbial diversity was characterized by α- and β-diversity (Shannon and Bray-Curtis index) and richness. We used elastic net regression and generated c-statistics to determine the utility of gut microbial composition and diversity in classifying glycemic status while adjusting for age, sex, antibiotic use, waist circumference, hypertension/lipid-lowering medication (model 1). The prevalence of prediabetes and diabetes was 368 (29%) and 113 (9%), respectively. Compared to model 1, differential abundance at the family domain was associated with a moderate improvement in classifying participants by glycemic status (c-statistic: 0.78 vs 0.80). We observed a differential abundance of 16 bacteria families between diabetes/normoglycemia, 6 between prediabetes/diabetes, and 3 between normoglycemic/prediabetes, but microbial diversity did not differ. The abundance of Ruminococcacease and Enterobacteriaceae were the top bacterial contributors to glycemic status prediction. Glycemic status was associated with distinct patterns of microbial composition. The presence of diabetes was associated with a lower abundance of butyrate-producing bacteria, which may be linked to glycemic control.