Objective.To analyze trends in the incidence and pathogen distribution of healthcare-associated bloodstream infections (HABSIs) over a 20-year period (1992–2011).Design.Historical cohort study.Setting.Thirty-two-bed neonatal intensive care unit (NICU) in a tertiary referral hospital.Patients.Neonates with HABSIs defined according to the criteria of the National Institute of Child Health and Development (NICHD).Methods.A hospital-based ongoing surveillance program was used to identify HABSI cases in neonates. A distinction between definite or possible HABSI was made according to the NICHD criteria. Incidence, incidence densities (HABSIs per 1,000 hospital-days and HABSIs per 1,000 total parenteral nutrition–days), and case fatality rate were calculated. Logistic regression analysis was used to find time trends. Four periods of 5 years were considered when executing variance analysis.Results.In total, 682 episodes of HABSIs occurred on 9,934 admissions (6.9%). The median total incidence density rate was 3.1 (interquartile range, 2.2–3.9). A significant increasing time trend in incidence density was observed for the period 1995–2011 (P < .003). A significant decrease in the case fatality rate was found in the last 5-year period (P < .001). No neonate died following possible HABSIs, whereas the case fatality rate among neonates with definite HABSIs was 9.7%. Most HABSIs were caused by coagulase-negative staphylococci (n = 414 [60.7%]). A significant increase in Staphylococcus aureus HABSI was observed in the last 10-year period (P < .001).Conclusions.An increase in incidence density rate occurred, while the case fatality rate dropped. Better perinatal care could be responsible for the latter. A decrease in days before infection and a high incidence of coagulase-negative Staphylococcus HABSIs indicate the need for vigorous application of evidence-based prevention initiatives, in particular for catheter care.