Definitions of SIRS and sepsis in correlation with early and late onset neonatal sepsis

Objective: To examine the applicability of the 2002 International Pediatric Sepsis Consensus Conference definitions of the systemic inflammatory response syndrome (SIRS) and sepsis to term and preterm newborns in the diagnosis of early (EOS) and late onset sepsis (LOS). Methods: Retrospective cohort study including 1) all newborns with hospitalization within the first 72 hours of life and 2) infants with episodes of suspected LOS at a tertiary care neonatal intensive care unit between 2004 and 2008, and correlating the definitions of SIRS and sepsis with culture proven and clinical EOS and LOS. Results: Association with EOS: Among term newborns SIRS and sepsis definitions applied to 62 and 39/245 newborns (25% and 16%, respectively) and to 5/13 cases of culture proven EOS (38%) and 34/66 cases of clinical EOS (52%), respectively. Among preterm newborns SIRS and sepsis definitions applied to 202 and 124/505 newborns (40% and 25%, respectively) and to 17/24 cases of culture proven EOS (71%) and 107/160 cases of clinical EOS (67%), respectively. Sensitivity of SIRS and sepsis definitions was higher in preterm compared to term newborns in case of culture proven and clinical EOS (p=.047 and p=0.03, respectively). Association with LOS: SIRS and sepsis definitions applied to 5/5 episodes of culture proven LOS (100%) and to 4/9 episodes of clinical LOS (44%) in newborns who were term at onset of sepsis (corrected gestational age) and to 14/19 episodes of culture proven LOS (74%) and 24/28 episodes of clinical LOS (86%) in preterm newborns. Conclusion: The definitions of SIRS and sepsis correlated well with LOS but poorly with EOS, where nearly two thirds of term and one quarter of preterm newborns would have been missed. Postnatal age rather than gestational age had a positive influence on the correlation.