Prematurity is an important cause of perinatal death, and no reliable biochemical/biophysical markers exist to identify newborns with an increased mortality risk. We aimed to use S100B concentrations in urine as an early indicator of risk of neonatal death. We did a cross- sectional study using urine obtained from 165 preterm newborns, of whom 11 suffered neonatal death within the first week, 121 displayed no overt neurologic syndrome, and 33 suffered neonatal hypoxia and intraventricular hemorrhage (IVH) but not ominous outcome. Urine S100B concentrations were determined at four time- points and corrected for gestational age by conversion to multiples of median (MoM) of healthy controls of the same gestational age. Ultrasound imaging was assessed within the first 72 h from birth. In infants that died within the first week, S100B levels in urine were already higher than controls at first urination and increased progressively between the 24 and 96- h time- points. Multiple logistic regression analysis showed a significant correlation between urine S100B protein concentrations and the occurrence of neonatal death. An S100B concentration cut- off of 12.93MoM at first urination had a sensitivity of 100% and a specificity of 97.8% for predicting an ominous outcome. The positive predictive value was 78.6% , the negative predictive value was 100% . Measurement of urine S100B protein levels in preterm newborns could be useful to identify newborns at higher risk of neonatal death. We aimed to use S100B concentrations in urine as an early indicator of risk of neonatal death. We did a cross sectional sectional view to identify newborns with an increased mortality risk. of which 11 had neonatal hypoxia and intraventricular hemorrhage (IVH) but not ominous outcome. time-points and corrected for gestational age by conversion to multiples of median (MoM) of healthy controls of the same gestational age. Ultrasound imaging was assessed within the first 72 h from birth. In infants that died within the first week, S100B levels in urine were already higher than controls at first urination and increased progressively between the 24 and 96- h time- points. Multiple logistic regr ession analysis showed a significant correlation between urine S100B protein concentrations and the occurrence of neonatal death. An S100B concentration cut-off of 12.93MoM at first urination had a sensitivity of 100% and a specificity of 97.8% for predicting an ominous outcome. The positive The predictive value was 78.6%, the negative predictive value was 100%. Measurement of urine S100B protein levels in preterm newborns could be useful to identify newborns at higher risk of neonatal death.