Predicting Neonatal Early Onset Sepsis: A 14-Year Cohort Study.

Abstract

In many infants, treatment is started for suspicion of early onset sepsis (EOS), of whom the majority do not have an infection. Early prediction of the absence of a culture-proven sepsis (CPS) would significantly reduce the time of antibiotic treatment and hospitalization. Our objective was to analyze 3 criteria in infants with CPS: positive blood culture (BC) at 24 hours after the onset of suspicion of EOS (OSEOS), C-reactive protein (CRP) ≥10 mg/L and clinical signs of infection, so we can consequently consider to stop antibiotic treatment in infants without these criteria. We included all infants with suspicion of EOS from 2007 until 2020. The proportion was calculated of (1) infants with CPS with, at 24 hours, a positive BC and/or CRP ≥10 mg/L and/or clinical signs of infection and (2) infants without CPS with CRP <10 mg/L between 12 and 24 hours after OSEOS. The BC showed growth of a pathogenic microorganism in 50 of 4120 included infants (1.2%). Time to positivity was ≥24 hours in 8 (16%) infants, of whom 7 infants had a raised CRP and/or clinical symptoms of infection within 24 hours. In 1095 (74%) of infants without CPS in whom CRP was measured between 12 and 24 hours after OSEOS, CRP was <10 mg/L. A combination of BC, CRP, and clinical signs of infection can diagnose 98% (49/50) of infants with CPS 24 hours after OSEOS. Based on normal CRP and the absence of a positive BC, the decision to stop antibiotics could have been brought forward to 24 hours in 74% of infants.