Diagnostic of early neonatal infection remains a serious challenge. Since clinical symptoms and biological markers are neither sensitive nor specific, a lot of newborns suspected of infection, undergo biological analysis and empirical antibiotic treatment awaiting bacteriological results. However the prevalence of the disease has dramatically decreased since the last twenty years, subsequently to the generalization of the per partum antibioprophylaxis strategy. Because the possible deleterious effects of neonatal antibiotic treatment are well described now, it appears very urgent to restrict their use to the infected newborns only. Recent studies underline the benefit of using procalcitonin (PCT) to diferentiate virai infections of bacterial infections. PCT in blood cord could become a new and efficient marker to help neonatologists taking care of infection-suspected newborns. An evidence based approach is necessary, combining anamnestic, clinical and biological data as PCT to identify the very low risk newborns population and to limit the neonatal antibiotic prescriptions.