Abstract
BACKGROUND: Invasive Group B streptococcus (GBS) is a major cause of serious neonatal infection. Current strategies to reduce early-onset GBS disease have no impact on late-onset disease (LOD). Although GBS LOD is viewed as a sporadic event in the community, LOD arising within the neonatal intensive care unit (ICU) raises questions about mode of acquisition. METHODS: Following a cluster of 4 GBS LOD cases, enhanced surveillance for all GBS LOD was undertaken over 2 years in the neonatal ICU supported by neonatal rectal screening. GBS isolates were serotyped and genome-sequenced. RESULTS: Twelve late -onset invasive GBS episodes were identified (incidence 0.6/1000 live births). Genomic analysis revealed that 11/12 GBS isolates (92%) were linked to at least one other LOD isolate. Isolates from the first cluster were serotype V, resistant to macrolides and lincosamides, and sequencing confirmed isolates were indistinguishable, or distinguishable by only one SNP difference, from each other. Rectal carriage was rare. Prospective surveillance identified three further clusters of LOD due to serotypes Ia (3 cases), Ib (2 cases), and III (2 cases), that would not have been identified without surveillance and genome sequencing, leading to a re-evaluation of interventions required to prevent GBS LOD. CONCLUSION: Acquisition routes for LOD GBS in the neonatal ICU are poorly understood; cases may not necessarily be sporadic. Within this neonatal ICU, our data suggest that a single case of LOD GBS sepsis should be considered a potential nosocomial transmission event warranting prompt investigation, heightened infection prevention vigilance and action where required.