METABOLOMICS APPLIED TO BIOMARKER DISCOVERY IN RESPIRATORY OUTCOME OF PRETERM INFANTS

Background: Bronchopulmonary dysplasia (BPD) is the most common chronic lung disease in preterm infants. Despite significant improvements in neonatal intensive care, its incidence is still increasing. The long-term respiratory and neurological consequences of BPD have a major impact on survivors' quality of life. Currently, there is still no specific treatment once BPD is established, so it is important to identify children at risk in the first days of life. Metabolomics is the most recent "omics" science and allows the identification and quantification of all metabolites present in a complex biological system. The application of the metabolomic approach to diseases permits the identification of characterizing metabolites, even unknown ones, which help to understand the pathogenetic mechanisms underlying the disease and potentially can become a diagnostic tool. Aim of the study: The aim of the study is to apply the metabolomic approach in order to establish whether a biomarker profile exists at birth, capable of predicting the development of BPD in preterm infants. The secondary aim is to characterized the urinary metabolome of BPD patients at birth. Materials and Methods: This is a monocentric prospective observational case-control study. It was conducted at the Neonatal Intensive Care Unit (NICU), Hospital-University of Padova, Padova, Italy. Premature neonates admitted to the NICU, born at less than 32 weeks' gestational age (GA) have been enrolled. Patients who develop BPD were enrolled as cases, patients without BPD as controls. BPD was defined as need for oxygen supplementation for at least 28 days. Urine sample was collected within 48 hours of life for each patient and untargeted metabolic analysis was performed with ultra performance liquid chromatography system coupled to a Quadrupole Time-of-Flight mass spectrometer. Results: 161 neonates have been enrolled, of that 69 with BPD (42.9%) and 92 without BPD (57.1%). The median GA at birth was 29 (27-31) weeks and the median birth weight 1115 (800-1400) g. Urine untargeted metabolomic revealed a metabolites clustering in BPD neonates in comparison to neonates without BPD. The metabolomic derangements concerned metabolites belonging to the acylcarnitine class, uridine and pseudouridine. To control the effect GA on metabolome, 11 infants with BPD were matched with 11 without BPD of comparable GA. After adjusting the analysis for GA, the main discriminating metabolites were: L-cystathionine, isovalerylcarnitine, creatine, p-hydroxyphenylacetic acid. Compared to clinical data alone, the application of urinary metabolome analysis at birth increases the sensitivity and specificity in the prediction of BPD, respectively from 86.7% to 90.0% and from 84.0% to 86.0%. Conclusions: The metabolomic approach on urine collected within 48 hours of life revealed an interesting clustering of metabolites that discriminates between preterm infants at risk of developing BPD versus infants who will not develop it. The analysis of the identified metabolites at birth increases the prediction of BPD development in preterm infants. In the next future, early detection of infants at risk of developing BPD may be useful for the administration of targeted therapies.