Objectives: To describe the epidemiology of neonatal GBS-disease over five years (1997-2001) in the Netherlands, stratified for proven and probable sepsis and for very early (<12 hr), late early (12 hr-<7d) and late (7-90 d) onset sepsis. To evaluate the introduction in January 1999 of guidelines for prevention of early- onset GBS disease based on risk factors. Methods: Cases were collected in co-operation with the Dutch Paediatric Surveillance Unit and corrected for under-reporting by the capture-recapture technique. Results: The total incidence of proven very early- onset, late early-onset and late-onset GBS sepsis was 0.32; 0.11 and 0.14 per 1000 live births respectively. That of probable very early-onset, late early-onset and late-onset GBS sepsis was 1.10; 0.18 and 0.02 per 1000 live births respectively. In 46% of the proven early- onset cases maternal risk factors were absent. Significantly more infants with proven GBS sepsis were boys. In proven very early-onset GBS sepsis 64% of the patients were first born versus 47% in the general population. After the introduction of guidelines the incidence of proven early-onset (<7 days) sepsis decreased significantly from 0.54 per 1000 live births in 1997-1998 to 0.36 per 1000 live births in 1999-2001. However, the incidence of meningitis and the case fatality rate in the first week of life did not decrease. The incidence of late-onset sepsis remained unchanged. Conclusion: After introduction of risk factor based prevention guidelines, the decrease of the incidence of proven early-onset (<7 days) GBS sepsis was limited. We therefore recommend adaptation of the Dutch GBS prevention guidelines.