Reduction of Duration of Antibiotic Therapy for Suspected Early-Onset Sepsis in Late-Preterm and Term Newborns After Implementation of a Procalcitonin-Guided Algorithm: A Population-Based Study in Central Switzerland.

Background: Suspected early-onset sepsis (EOS) is the main reason for antibiotic therapy at the start of life. Prolonged antibiotic therapy for culture-negative sepsis is often reported. Antibiotic stewardship is mandatory due to the potential negative effects of unnecessary antibiotics. Procalcitonin (PCT)-guided therapy is one possible strategy with published evidence to shorten antibiotic therapy. The aim of this study is to analyze the feasibility and the performance of the published PCT-algorithm in the clinical setting without study support. Methods: This is a retrospective, population-based study regarding duration of antibiotic therapy for suspected EOS in Central Switzerland between 2014 and 2018. All neonates >34 0/7 weeks of gestational age started on antibiotic therapy for suspected EOS within the first 3 calendar days of life were included. The Procalcitonin-guided algorithm according to the NeoPInS study was used as strategy to determine duration of antibiotic therapy. Results: In a population-based cohort of 35,642 life born neonates, the duration of antibiotic therapy of 879 neonates (2.5% of the cohort) treated for suspected EOS was 4 calendar days (median, IQR 2-5). We observed a statistically significant reduction from 4 (median, IQR 3-6) to 3 calendar days (median, IQR 2-4) from 2014 to 2018. Duration of antibiotic therapy was independent of gestational age (late-preterm vs. term neonates), of the presence of risk factors or clinical signs, but dependent on the presence of abnormal laboratory measurements (C-reactive protein > 10 mg/l or leukocytopenia <5 Giga/l) before start of antibiotic therapy (p < 0.01). Conclusions: PCT-guided therapy using the NeoPInS algorithm is feasible and may lead to reduced duration of antibiotic therapy for suspected EOS as reported in the original study. We observed a learning curve to the new algorithm which may be explained as change process. The use of biomarker to guide duration of antibiotic therapy for suspected EOS may have unintended consequences with prolongation of antibiotic therapy in some cases.