Early‐onset neonatal sepsis (EONS) can have devastating consequences. Historically, clinicians treated at‐risk patients with antibiotics while awaiting blood culture results. This strategy, however, resulted in substantial antibiotic overtreatment over the last three decades.1 Consequences of early antibiotic overexposure include microbiome alterations2, 3 which are linked to development of asthma, food allergies and childhood obesity.4-7 Intrapartum antibiotic strategies have reduced the incidence of EONS since the 1990s, and clinicians now must weigh the risk of EONS with the risk of negative impacts from antibiotic overtreatment. Methods to prevent overtreatment include algorithm‐based risk factor assessments to distinguish infants truly at risk for EONS. A newer tool, the SRC, is a multivariate risk assessment method that synthesises risk factors and the newborn clinical condition to provide a probability of EONS for each infant.8, 9

The authors of this study estimated that, had the SRC been applied to their cohort of 3593 infants managed with NICE guidelines over a three‐month period, antibiotic therapy could have been avoided in nearly three‐quarters of treated infants.

However, applying a hypothetical management approach retrospectively to a cohort managed with another approach prospectively obviously does not provide the same level of evidence as head to head randomised trials. Studies such as this should be interpreted with caution. In this study, assessors of the SRC‐based approach were aware of the NICE guideline‐based assessment of risk, the actual use of antibiotics in study infants and the eventual clinical outcomes. Non‐blinding may have led to bias in estimating the potential effects of the SRC recommendations. Additionally, the true incidence of EONS in this cohort of patients was at least 1.2/1000 live births (6 infants out of 4992), but the authors used an estimated incidence of 0.5/1000 for the SRC. Using the known EONS incidence of this cohort of patients in the SRC may have resulted in a significant change in antibiotic usage.

Readers should note that several infants with EONS were deemed to be at low/no risk using either approach, but subsequently developed clinical signs of sepsis. These infants were identified by clinical vigilance, which may be the ultimate safety net for EONS detection and treatment. The authors do recommend close observation of newborns identified as high risk by NICE but not recommended antibiotics by SRC, and discharge after 12‐24 hours of monitoring. One reasonable conclusion from this study is that a combination of using the SRC along with clinical vigilance offers the best approach, where antibiotic overuse is avoided, and infants who clinically deteriorate are detected and treated early.

Even with promising results such as in this study, clinicians should cautiously develop local practice guidelines based on a comprehensive review of the evidence, local contextual and practice conditions (eg the feasibility of frequent clinical monitoring of infants at risk), the baseline rate of EONS and parental preferences and values. The guidelines should be accompanied by rigorous monitoring for unexpected consequences to ultimately ensure that the goal of achieving a balance between treating EONS early while avoiding needless antibiotics is met.

早发性新生儿败血症（EONS）可能具有毁灭性后果。从历史上看，临床医生在等待血培养结果的同时，用抗生素治疗高危患者。然而，在过去的三十年中，这种策略导致大量的抗生素过度治疗。1早期抗生素过度暴露的后果包括微生物组变化2，3，这些变化与哮喘的发展，食物过敏和儿童肥胖有关。4-7自1990年代以来，产前抗生素策略已降低了EONS的发生率，临床医生现在必须权衡EONS的风险和抗生素过度治疗带来的负面影响的风险。预防过度治疗的方法包括基于算法的危险因素评估，以区分真正有EONS危险的婴儿。一种更新的工具SRC是一种多变量风险评估方法，可以综合风险因素和新生儿的临床状况，为每个婴儿提供EONS的可能性。8、9

该研究的作者估计，如果将SRC应用于在3个月内接受NICE指南管理的3593例婴儿，则近四分之三的治疗婴儿都可以避免抗生素治疗。

但是，将假设管理方法追溯应用到前瞻性采用另一种方法管理的队列中，显然不能提供与随机对照试验相同水平的证据。此类研究应谨慎解释。在本研究中，基于SRC的方法的评估者了解基于NICE指南的风险评估，研究婴儿中抗生素的实际使用情况以及最终的临床结果。不盲目性可能导致在估计SRC建议的潜在影响时产生偏见。此外，该组患者中EONS的真实发生率至少为1.2 / 1000活产（4992名婴儿中有6例），但作者对SRC的估计发生率为0.5 / 1000。

读者应注意，使用这两种方法，几名患有EONS的婴儿被认为处于低风险/无风险，但随后出现了败血症的临床体征。这些婴儿是通过临床警惕来识别的，这可能是EONS检测和治疗的最终安全网。作者确实建议密切观察被NICE鉴定为高风险的新生儿，但不建议SRC推荐抗生素，并在监测12-24小时后出院。这项研究的一个合理结论是，将SRC与临床警惕相结合可提供最佳方法，避免抗生素过度使用，并尽早发现并治疗临床恶化的婴儿。

即使在本研究中取得了可喜的结果，临床医生也应根据对证据，当地背景和实践条件（例如，对有风险的婴儿进行频繁临床监测的可行性），EONS的基线发生率的全面审查，谨慎制定当地的实践指南以及父母的偏好和价值观。该指南应伴随有对意外后果的严格监控，以最终确保达到在早期治疗EONS与避免不必要的抗生素之间取得平衡的目标。