Abstract Objective: Our objective was to determine if early pregnancy urinary metabolomic profiles could predict fetal adiposity and macrosomia. Methods: This is a prospective study of 50 healthy women in their second pregnancy. Fasting urine samples taken during early pregnancy were analyzed using NMR spectroscopy. Maternal glucose and insulin were measured in early pregnancy and at 28 weeks and the HOMA index for insulin resistance calculated. At 34 weeks ultrasound assessed fetal anthropometry including fetal anterior abdominal wall width (AAW). At delivery birth weight was recorded. Probabilistic principal component with covariates analysis (PPCCA), a novel extension of principal component analysis, which facilitates joint modeling of metabolomic data and additional covariate information, was employed to analyze the data. Results: This analysis revealed that maternal HOMA and AAW significantly covaried with the (1)H NMR derived metabolomic profile of the urine. As such, in this cohort of healthy, non-diabetic women, early pregnancy urinary metabolomic profile differed significantly according to both maternal insulin resistance and fetal fat deposition in utero. Conclusion: These findings hold potential for early pregnancy identification of those at risk of fetal macrosomia.
University College Dublin ->