Background: some consensus guidelines include C-reactive protein (CRP) in the diagnostic workup of early-onset neonatal sepsis (EOS), but its routine use remains debated due to variable diagnostic performance. The experiences and data from individual centers can help clarify its clinical utility and inform local practice. Methods: Retrospective analysis at a level III center assessing the impact of discontinuing routine C-reactive protein (CRP) testing for suspected early-onset sepsis (EOS). Laboratory use, antibiotic therapy, and outcomes in neonates of all gestational ages were compared before (2021–2022) and after (2024–2025) the policy change. Results: A total of 638 neonates were included (period 1, n = 348; period 2, n = 290). CRP testing decreased markedly (218/348 in period 1 vs. 40/290 in period 2; p < 0.001), alongside a significant reduction in the number of complete blood counts performed (285/348 vs. 214/290; p = 0.02). Concurrently, both the proportion of short antibiotic courses (≤48 h) initiated within the first 3 days of life (98/181 vs. 88/133) and the median duration of antibiotic therapy (48.0 h vs. 40.0 h; p < 0.001) decreased without worsening outcomes. The duration of antibiotic therapy was even shorter in infants born before 34 weeks’ gestation (48.0 h vs. 37.5 h; p < 0.001). Conclusions: Restricting the use of CRP in the evaluation of EOS was associated with a reduction in unnecessary antibiotic exposure. This strategy may be considered a core component of neonatal antibiotic stewardship programs.

Background: Antimicrobial therapies used for treating group B streptococcus (GBS) early-onset sepsis (EOS) provide insight into clinicians’ adherence to antimicrobial stewardship (AMS) guidelines. Methods: We reviewed antimicrobial therapies given to treat newborns with GBS-EOS. Data were obtained from an Italian surveillance network (including 35 birthing centers) and were prospectively collected from 1 January 2003 to 31 December 2024. Empiric and definitive therapies were classified as adequate and inadequate. Results: There were 967,054 live births and 200 cases of GBS-EOS, of which 43 (21.5%) were preterm and 157 (78.5%) were full-term; 35 (17.5%) out of 200 showed no signs of illness. Fourteen (7.0%) died (one full-term and thirteen preterm newborns under 34 weeks of gestation). Based on the available information, antibiotics were adequate in 106/137 (77.4%) empiric and 48/119 (40.3%) definitive therapies. The duration of antibiotic courses did not differ between severe (median 10 days, IQR 8.0–14.0) and non-severe cases (median: 10 days; IQR: 10.0–12.5; p = 0.68). Antibiotic treatments lasted ≥ 15 days in 34 (20.1%) out of 169 cases with available information. Conclusions: In this large Italian multicenter study, deviations from international recommendations in antimicrobial therapies for GBS-EOS were critical. Our findings underscore the importance of timely antimicrobial de-escalation and the need to avoid excessively prolonged courses of antimicrobials.