There is growing interest for studying how early-life influences the development of respiratory diseases. Our aim was to apply metabolomic analysis to urine collected at birth, to evaluate whether there is any early metabolic signatures capable to distinguish children who will develop acute bronchiolitis and/or recurrent wheezing. Urine was collected at birth in healthy term newborns. Children were followed up to the age of 3 years and evaluated for the development of acute bronchiolitis and recurrent wheezing (≥3 episodes). Urine were analyzed through a liquid-chromatography mass-spectrometry based untargeted approach. Metabolomic data were investigated applying univariate and multivariate techniques. 205 children were included: 35 had bronchiolitis, 11 of whom had recurrent wheezing. Moreover, 13 children had recurrent wheezing not preceded by bronchiolitis. Multivariate data analysis didn't lead to reliable classification models capable to distinguish children with and without bronchiolitis or with recurrent wheezing preceded by bronchiolitis neither by PLS for classification (PLS2C) nor by Random Forest (RF). However, a reliable signature was discovered to distinguish children who later develop recurrent wheezing not preceded by bronchiolitis, from those who do not (MCCoob = 0.45 for PLS2C and MCCoob = 0.48 for RF). In this unselected birth cohort, a well-established untargeted metabolomic approach found no biochemical-metabolic dysregulation at birth associated with the subsequent development of acute bronchiolitis or recurrent wheezing post-bronchiolitis, not supporting the hypothesis of an underlying predisposing background. On the other hand, a metabolic signature was discovered that characterizes children who develop wheezing not preceded by bronchiolitis.
Keywords: bronchiolitis; metabolomics; neonate; recurrent wheezing; urine.