Intensity of care is the major determinant of coagulase-negative staphylococcal bacteraemia in NICU: implications for risk adjusted surveillance of hospital-acquired infections

Abstract number: 1733_1439

Phillips P.H., Cortina Borja M., Millar M., Thalib L., Kempley S., Gilbert R.

Coagulase-negative staphylococcus (CNS) is the commonest cause of bacteraemia in UK Neonatal Intensive Care Units (NICU), accounting for two-thirds of positive blood cultures[1]. CNS bacteraemia is related to the use of invasive devices and can be reduced by up to 50% by improved hygienic practices [2]. We propose that the rate of CNS bacteraemia could be used as a marker for the quality of care related to hospital acquired infection in NICUs. To be interpretable across different units, surveillance must be risk adjusted to take into account the vulnerability of babies and the invasive procedures used. We examined how risk factors including intensity of care predicted CNS bacteraemia in a London NICU.

Methods: The study population included all NICU inpatients in the hospital over 9 years. Admission data were linked with the laboratory database for all blood cultures. Levels of care, used for resource allocation throughout the UK, were defined as Special Care, High Dependency or Intensive Care, and were used as surrogate markers for invasive procedures. Dates of admission to each level were used to determine the total days at risk at each level of care. Analyses determined the rates of CNS infection stratified by various risk factors and intensity of care. Multivariate Poisson regression was then used to estimate adjusted incidence rate ratios for CNS bacteraemia.

Results: 2492 infants were admitted to NICU. 393 infections were recorded, of which 164 (42%) were recurrent. The overall rate of CNS bacteraemia was 5.79 per 1000 baby days, with crude rates of 2.95, 9.63 and 8.92 in Special Care, High Dependency and Intensive Care respectively. Intensity of care was strongly associated with CNS infection, with those in High Dependency being most at risk: the incidence rate ratios adjusted for gestation at birth were 3.48 (95% CI 2.43–4.96) and 2.24 (95% CI 1.54–3.29) for High Dependency and Intensive Care compared with Special Care. Babies born between 32 and 37 weeks were significantly less likely to develop bacteraemia than term infants, but the adjusted rate ratios for lower gestations were not significant.

Conclusion: Level of care was the main determinant of the incidence of CNS bacteraemia. These predictors warrant evaluation in other datasets with a view to developing a marker for risk adjusted monitoring of hospital acquired bacteraemia.