Background Infections caused by coagulase negative staphylococci (CoNS) form a large proportion of neonatal late-onset sepsis (LOS) episodes, contributing to mortality, healthcare costs and morbidity. CoNS is a commensal organism and true bacteraemia must be distinguished from contamination. The UK National Neonatal Audit Programme (NNAP) and the Vermont Oxford Network (VON) data collecting systems both record neonatal LOS rates. However, each uses a different definition of true sepsis when CoNS is grown: VON require 5 days minimum antibiotic treatment; NNAP requires only 3 or more clinical signs to be present. Neither incorporates laboratory markers of sepsis, which are used clinically in our unit.
Aims To compare the estimated CoNS neonatal LOS rates using the VON, NNAP and “Clinical” definitions.
Methods Data from clinical and laboratory records were collected retrospectively for all episodes of suspected sepsis where CoNS grew from blood culture between January 2013 and March 2014. Cohen’s kappa test was used to analyse agreement between definitions.
Results CoNS was grown during 35 episodes of suspected sepsis. Rates of agreement between the 3 definitions are shown in table.
Conclusions The 3 different definitions for LOS caused by CoNS gave differing estimates of infection rates. Strongest agreement occurred between the VON and “Clinical” assessments. This information is important when comparing LOS rates between different neonatal units.
Acknowledgements Thanks to Dave Browning, (Microbiology Statistician); the staff on the neonatal unit in Southampton.