Gestational age and models for predicting gestational diabetes mellitus.

Area of Science:

• Obstetrics and Gynecology
• Metabolomics
• Endocrinology

Background:

• Gestational diabetes mellitus (GDM) is typically diagnosed late in pregnancy via glucose load tests.
• Previous research suggests GDM may be identifiable in early pregnancy using urinary metabolite profiles.
• Metabolomic changes in late pregnancy may differ, necessitating a distinct predictive model.

Purpose of the Study:

• Identify late-pregnancy urinary metabolites associated with GDM.
• Develop a predictive model for GDM using selected metabolites.
• Compare the predictive accuracy of late-pregnancy and early-pregnancy models.
• Investigate the link between GDM urinary markers and declining insulin sensitivity.

Main Methods:

• An observational nested case-control study involving 46 GDM patients and 46 controls.
• Urine samples collected at ≥24 weeks gestation analyzed via global metabolomics.
• Multivariate criteria and a classification tree model used to distinguish GDM from controls.

Main Results:

• Eight metabolites, including markers of mitochondrial dysfunction and oxidative stress, differentiated GDM from controls.
• A classification tree model using four metabolites predicted GDM with 89% accuracy.
• The previously developed early-pregnancy model showed 85.9% accuracy in this cohort.

Conclusions:

• Late-pregnancy urinary metabolites are linked to diminished insulin sensitivity and impaired insulin release in GDM.
• Urinary metabolite profiles show potential for identifying GDM in both early and late gestation.
• Further validation studies are recommended for this noninvasive GDM detection technology.