Objective: To differentiate between neonates with high and low risk of infections caused by group B beta-haemolytic streptococci (GBS), by using the urinary group B streptococcal antigen test.
Design: Retrospective.
Setting: Medical Centre Leeuwarden and Public Health Laboratory, Friesland, the Netherlands.
Methods: In a period of two years clinical, haematological and microbiological (including urinary group B streptococcal antigen detection) data were collected in newborns and their mothers who met one or more of the following criteria: a previous affected child, prolonged (> or = 12 hrs) rupture of membranes, fever in labour, unexpected preterm delivery, unexplained perinatal asphyxia. On the basis of surveillance cultures a colonization rate was made. GBS infection was 'suspected' in an unwell infant with a 'high' colonization rate; infection with GBS was 'proved' by a positive blood culture with GBS.
Results: 6 of 342 neonates had an infection with GBS. Risk of invasive infection increased with higher colonization rates. Sensitivity of the antigen test to detect colonization was low, sensitivity to detect neonatal infection was high (51 versus 100%). The negative predictive value of urinary antigen testing was 100%. Prolonged rupture of membranes (1.5% risk of infection) and maternal fever (5%) were the most important risk factors.
Conclusion: In healthy neonates with risk factors but with a negative antigen detection test the risk of GBS infection is extremely low. In children with a risk factor a positive test result can indicate heavy colonization or infection. These children should be carefully observed and examined.