Objectives: We tested the application of bioinformatic algorithms in studying the metabolomic profiles of neonatal urine samples with clinical evidence of severe asphyxia at birth and subsequent neurodevelopmental handicap.
Design and methods: The clinical outcomes of 256 newborns that required direct admission to neonatal intensive care unit for respiratory support or did not require direct admission were studied. Urinary metabolite profiles were measured by high throughput mass spectrometry and analyzed by bioinformatic methods.
Results: We found a positive relationship between suppressed biochemical networks involved in macromolecular synthesis and birth asphyxia associated with significant neonatal oxidative stress and morbidity. The metabolomic discriminators between good neonatal outcome and poor neonatal outcome were established using hierarchical clustering analysis. Concentrations of eight urinary organic acids in distinct biochemical pathways were elevated and significantly associated with the prognosis of neurodevelopmental handicap with high sensitivity and specificity: ethylmalonate, 3-hydroxy-3-methylglutarate, 2-hydroxy-glutarate and 2-oxo-glutarate were associated with good neonatal outcome, whereas glutarate, methylmalonate, 3-hydroxy-butyrate and orotate were associated with poor outcome.
Conclusions: The data demonstrated the potential application of bioinformatics methods in this metabolomic study and proved its clinical relevance.