Objective

Prenatal detection of congenital heart defect (CHD) can reduce pediatric morbidity and mortality. Currently half of the CHD cases fail to be detected however, highlighting the need for population screening tools. Metabolites are small molecules and byproducts of cellular metabolism. They traffic easily between the fetus and maternal vascular compartments and are detectable in other body fluids. We investigated the diagnostic utility of salivary metabolomics for isolated non-syndromic CHD detection. Study Design

In this prospective study, maternal saliva was obtained in 40 fetal CHD pregnancies (confirmed by newborn echocardiogram) and 40 gestational age (GA) matched controls. Saliva was collected and subsequently underwent metabolomic analysis using direct injection liquid chromatography coupled with mass spectrometry (DI-LC-MS). Logistic regression models were developed in a training group (26 CHD/ 26 controls) and validated in an independent test group (14 CHD/ 14 controls). Area under the ROC (95% CI), sensitivity, and specificity were calculated. Metabolite pathway analysis was used to elucidate CHD pathogenesis. Results

The median GA at testing was not different between groups: CHD (24.470 ± 5w2d) vs controls (23w3d ± 3w5d) (p=0.8). A total of 269 metabolites were identified in saliva. The concentration of 20 metabolites differed significantly (p< 0.05) between groups. Predictive algorithms (Table 1) achieved good diagnostic performance (AUC > 0.75) and metabolites discriminated the patient groups (Fig 1a). Alanine, aspartate and glutamate metabolism was dysregulated in CHD. and were recently shown to be disrupted in childhood cyanotic CHD. Tryptophan metabolism, also altered, is known to affect myocardial remodeling and cardiac function (Fig 1b). Conclusion

We report for the first time the potential of salivary analysis to detect CHD and elucidate its pathogenesis. Expanded metabolomic analysis with an addition of Nuclear Magnetic Resonance Spectroscopy could further boost the accuracy.