97: Maternal Urine Metabolomics and the Prediction of Fetal Non-Syndromic Congenital Heart Defect (CHD)

Metabolomics identifies and quantifies the downstream small molecules involved in cellular metabolism providing the greatest insight into any given phenotype. Using targeted metabolomics, we profiled maternal urine to identify diagnostic biomarkers of non-syndromic fetal congenital heart defects (CHD). We prospectively collected maternal urine between 14 0/7 -37 0/7 weeks gestation and employed tandem mass spectrometry and nuclear magnetic resonance spectroscopy to analyze our specimens. Logistic regression was used to generate diagnostic algorithms using a urine metabolite concentrations alone or in combination with clinical or ultrasound (4-chamber view) data. We compared 36 cases with 41 controls. The median gestational age at the time of specimen collection was not significantly different for cases versus controls (P= 0.089). A total of 222 metabolites were identified and quantified. In total, we identified 23 metabolites significantly perturbed (P< 0.05) in CHD maternal urine with methionine being the most markedly different. Combining metabolite concentrations (histamine, choline, glucose, formate, methionine, carnitine) with ultrasound 4-chamber view we developed a diagnostic algorithm with an AUC = 0.894 (95% CI: 0.814-0.973) with sensitivity and specificity values equal to 83.8% and 87.8%, respectively (Table 1). Metabolite set enrichment analysis identified several lipid related pathways to be significantly disrupted in CHD including phospholipid biosynthesis, phosphatidylcholine and phosphatidylethanolamine biosynthesis and fatty acid metabolism. We report the use of a maternal urine test by itself or combined with sonographic and demographic markers for CHD screening. Lipid dysfunction is well known to play a pivotal role in postnatal cardiac disorders. Here we found evidence that lipids were extensively dysregulated and could also be important in CHD pathogenesis.