Human cytomegalovirus (HCMV) represents one of the most significant viral risks of birth defects and long-term sequelae. The severity of the infection depends on the form of the disease, which can be symptomatic or asymptomatic with or without sequelae. The aim of this study was to investigate in a population of newborns the impact of HCMV infection on the urine metabolome by using (1)H-nuclear magnetic resonance (NMR) spectroscopy combined with multivariate statistical analysis. Twenty-three children born from women with a primary HCMV infection during pregnancy were recruited. Twelve were HCMV infected at birth whereas eleven were not infected (control). The (1)H-NMR spectra were analyzed using a PLS-DA mathematical model in order to identify the discriminant metabolites between the asymptomatic and the control group. The most important metabolites characterizing the clustering of the samples were: myoinositol, glycine, 3-hydroxybutyrate, 3-aminoisobutyrate, creatine, taurine and betaine. These findings suggest the use of metabolomics as a useful new tool in the investigation of HCMV congenital infection.