ABSTRACT: The term neonatal sepsis is used to describe a generalized bloodstream infection of bacterial, viral, or fungal origin which is associated with hemodynamic changes and other clinical symptoms and signs, however, there is no unified definition. There are no basic criteria regarding differentiation of early-onset sepsis (EOS) versus late-onset sepsis (LOS). Stratification used in studies on neonatal sepsis also rarely includes the general condition of the newborn according to unambiguous assessment at birth, which hampers the establishment of a clear, uniform epidemiological description of neonatal sepsis. We aim to review the published data about the epidemiology and microbiology of sepsis in Organization for Economic Cooperation and Development (OECD) countries. Data was also collected on sepsis prevention programs that can be implemented in neonatal units. The outcomes of interest were incidence or incidence density of EOS and LOS, microbiology of EOS and LOS, and data on the methodology of the research, in particular the criteria for inclusion and exclusion of newborns from the study. Pubmed, EMBASE, LILACS Embase, Scopus, and Google Scholar were used. For the preselection step, inclusion criteria included: "bloodstream infection" or "neonatal sepsis" (MesH), "very low birth weight", and "country" full-text studies, human, and English language. Exclusion criteria included: studies published in languages other than English and studies available only as an abstracts. For proper selection, inclusion criteria included: information about epidemiology or microbiology bloodstream infection (BSI), study population and case definitions, exclusion criteria, narrative reviews, commentaries, case studies, pilot studies, study protocols, pediatric studies, and only clinical data (without microbiology or epidemiology) or studies with only one etiological factor analysis. The data review indicated the lack of an unequivocal, unified definition and no unambiguous basic criteria with regard to differentiation of EOS versus LOS. Among infants <1500 g, studies reported an EOS rate from 7% to 2%. For studies using other definitions (mostly all inborn babies), the rate of EOS ranged from 1% to 3%. The LOS incidences were much more varied among countries; the highest rates were in the multicenter studies focused on very low birth weight (VLBW) infants. The main pathogens in EOS are GBS and Gram-negative bacteria in LOS. Our review data shows that LOS microbiology is very diverse and that Gram-positive cocci, especially staphylococci, predominate versus Gram-negative rods. Unfortunately, the lack of uniform, international prevention programs results in high newborn morbidity and insufficient postnatal prevention of late-onset infections.
