Bacterial sepsis is a frequently occurring disease in the first weeks of life, posing a significant threat particularly to those born prematurely. The incidence of sepsis is determined by laboratory surveillance, only taking into account culture positive episodes of sepsis. However, in up to 80% of neonates treated for sepsis, blood cultures fail to grow an organism. Therefore a variety of molecular techniques have been trialled to overcome these diagnostic difficulties. To describe the current incidence and causal pathogens of neonatal sepsis in Europe in this thesis, systematic literature reviews on neonatal bacterial and fungal infections were conducted. The review highlighted in particular the discrepancy between incidences of culture positive and clinical sepsis. A further literature review assessed molecular diagnostic techniques that have been employed to determine pathogens of neonatal sepsis. Based on the results obtained from the systematic reviews, a series of molecular tests, including quantitative multiplex PCRs, a 16S rDNA broad range PCR and a Candida multiplex PCR were developed. These tests were applied to two sets of samples obtained from neonates with suspected and confirmed early and late onset sepsis in Europe, collected in two separate clinical trials. Results identified a large amount of bacteria (74% in EOS and 50% in LOS), however failed to detect all cultured pathogens. A large number of samples were positive for CoNS and Enterobacteriacae in both sample sets. In particular, in EOS, S. pneumoniae was shown to be more predominant than anticipated from the literature, and in LOS Enterococci were more prevalent. Of concern is a high number of polymicrobial infections detected by PCR. Universal definitions for clinical sepsis need to be established to enable surveillance and comparison across countries. Molecular diagnostics have the potential to become an important additional tool to describe the epidemiology of neonatal sepsis.