Identifying Neonates at Lowest Risk for Sepsis

In this issue of Pediatrics, Flannery et al¹  consider whether infants with a clear low risk of transmission of perinatal infection can be automatically excluded from consideration of sepsis screening and/or antibiotic therapy. They examine outcomes for infants who met low-risk criteria, including birth by cesarean delivery without labor, no rupture of membranes before delivery, and no clinical signs of sepsis in the mother or neonate, and find that even in premature infants, this constellation of clinical criteria was not associated with any cases of early-onset sepsis (EOS). Therefore, the authors conclude that infants who meet these criteria need not be screened for sepsis, even extremely and very premature infants.

Sepsis remains one of the most feared occurrences in any neonatal setting. Practitioners who have experienced a neonate becoming ill with sepsis have good reason to be cautious; neonatal sepsis can evolve rapidly and can be fatal even after initiation of appropriate antibiosis. EOS, or sepsis within 7 days of birth, carries a ∼20% mortality rate^2,3  and has led to the development of several important public health interventions aimed at reducing EOS, including screening pregnant women for carriage of group B Streptococcus and widespread prenatal counseling to avoid foods with a high risk of Listeria contamination during pregnancy.^4,5  Between these public health measures and medical improvements in monitoring and testing, the incidence of EOS has declined with time.⁶  However, despite that decline, preemptive use of antibiotics remains widespread in the newborn nursery and the NICU.

Contemporary knowledge about the negative impacts of antibiotic overuse, including its contributions to antibiotic resistance and alterations in the microbiome of the recipient, have helped the field of newborn care to start to reconsider our use of antibiotics.^7–9  The neonatal community has come to embrace antibiotic stewardship, and it is widely accepted that empirical antibiotics are not typically indicated for well-appearing term infants without high-risk infectious characteristics. The widely used neonatal EOS risk calculator¹⁰  only generates a sepsis risk score for infants ≥34 weeks’ gestation. For those who are well appearing without risk factors for infection, the general recommendation is observation and routine care with vital sign monitoring, which would be routine for any infant at this gestational age.

The extension of the Flannery et al¹  article down to extremely premature infants may be more of a paradigm shift to the field of neonatology. Extremely and very premature infants are relatively immunocompromised compared with late preterm- and term-born infants,¹¹  and the clinical tendency to empirically begin antibiotics on these infants is strong. However, the article presents evidence that the same high-risk characteristics can be used to stratify risk even for these higher-risk premature infants because no confirmed cases of sepsis were identified in those infants with low-risk clinical characteristics, regardless of gestational age.

Nonetheless, there are a few caveats to the conclusions. First, the sample size of the extremely and very premature infants was much smaller than the entire sample size, and the study was not powered to look specifically at that group. Secondly, it is not typical for an extremely premature infant to be borne by cesarean delivery without any laboring and/or evidence of fetal distress. The authors indicate most of these births were secondary to maternal indications, such as preeclampsia or maternal hypertension. Preeclampsia has been shown to be associated with neutropenia in very premature infants, which increases the risk of sepsis,^12,13  which raises concern as to whether this should be considered a low-risk scenario. A further point of caution is that 2 cases of EOS were identified that were initially stratified as low risk but, on further chart review, had high-risk characteristics. This highlights the continued need for scrutiny by clinical providers when applying these parameters as a clinical decision-making tool, which the authors themselves assert.

Finally, as clinicians, we all acknowledge that the risks and outcomes of sepsis are different for the population of extremely premature infants compared with the term infants. Application of the EOS risk calculator for the term infant results in decreased use of antibiotics because the risk for sepsis and decision for evaluation and use of antibiotics heavily rely on the clinical examination of the infant. Applying the same risk assessment tool in preterm infants would undoubtedly increase the use of antibiotics because their initial clinical signs and symptoms (temperature instability, need for respiratory support, and need for inotropic support) are likely to be present. Distinguishing the etiology of these symptoms between prematurity and sepsis is only possible if at least a baseline sepsis evaluation is done: by sending complete blood cell counts and a blood culture. It is almost impossible to imagine a clinical scenario in which extremely premature infant will not have any blood draws. Therefore, having a low threshold for sending baseline studies ensures no case of sepsis is missed, with minimal additional workup for the infant. This differs significantly from the approach to a term infant, who is presumably rooming in with the mother and who may warrant no further evaluation.

More reliable clinical signs or laboratory tests to detect sepsis would clearly be of value in discriminating clinical symptoms of sepsis from symptoms of prematurity. In the meantime, additional studies aimed at validating these screening criteria in extremely and very premature infants would be helpful for clinicians.

EOS

early-onset sepsis