Abstract Objective To assess the accuracy of umbilical cord bilirubin values to predict jaundice in the first 48 h of life and neonatal infection. Method Newborn infants treated at a regional well-baby nursery born at ≥36 weeks of gestation were included in this retrospective cohort study. All infants born in a 3-year period from mothers with O blood type and/or Rh-negative were included and had the umbilical cord bilirubin levels measured. Hyperbilirubinemia in the first 48 h was defined as bilirubin levels above the phototherapy threshold. Neonatal infection was defined as any antibiotic treatment before discharge. Results A total of 1360 newborn infants were included. Two hundred and three (14.9%) newborn infants developed hyperbilirubinemia in the first 48 h of life. Hyperbilirubinemic infants had smaller birth weight, higher levels of umbilical cord bilirubin, a higher rate of infection and were more often direct antiglobulin test positive. Umbilical cord bilirubin had a sensitivity of 76.85% and a specificity of 69.58% in detecting hyperbilirubinemia in the first 48 h, with the cut-off value at 34 µmol/L. The area under the receiver operating characteristic curve was 0.80 (95% CI: 0.78-0.82). Umbilical cord bilirubin had a sensitivity of 27.03% and specificity of 91.31% in detecting perinatal infection. The area under the receiver operating characteristic (ROC) curve was 0.59 (95% CI: 0.57-0.63). Conclusions A positive correlation was found between umbilical cord bilirubin and hyperbilirubinemia in the first 48 h of life. Umbilical cord bilirubin is a poor marker for predicting neonatal infection.