Associations of urinary microbiome profiles with gestational diabetes mellitus in an Oklahoma case-control study

Abstract

BACKGROUND AND AIM: Previous research has demonstrated associations between the human placental, gut, and vaginal microbiome and gestational diabetes mellitus (GDM). Recent findings suggest urine is a biomarker for integrated profiling of the genitourinary and amniotic microbiome. However, little is known about how urinary microbiome disturbances may relate to GDM. The present study aims to characterize and compare the urinary microbial profiles among pregnant women with and without GDM. METHODS: Bacterial DNA was extracted from mid-pregnancy urine obtained from a prevalence case-control study of 59 GDM cases and 206 controls from the University of Oklahoma Medical Center (August 2009 to May 2010). The taxonomic profile of the urinary microbiome was characterized by high-throughput sequencing of the V4 region of the microbial 16S rRNA gene. Associations between the microbial composition of taxa relative abundance and GDM were evaluated using multivariable logistic regression while controlling for maternal characteristics. Relative abundance (%) was categorized as absent, ≤median or median values. Alpha diversity (richness and evenness) and beta diversity (community structure) were also compared between cases and controls. RESULTS:A total of 27 microbial genera were identified in the cases and controls, which were predominated by the genera Lactobacillus and Gardnerella. The odds of a relative abundance median was decreased for the genus Aerococcus while it was increased for the genus Veillonella among the cases compared to the controls in the crude model. After adjusting for age and race, GDM cases had a decreased odds of higher relative abundance (median vs. absence) of Aerococcus compared to the controls, with the upper bound of the confidence interval (CI) marginally crossing 1.0 (adjusted odds ratio: 0.45, 95% CI 0.18, 1.10). No difference was observed in alpha and beta diversity metrics between cases and controls. CONCLUSIONS:Our hypothesis-generating results suggest the urinary microbiome composition may differ by GDM status. KEYWORDS: Microbiome, Pregnancy outcomes, Molecular epidemiology