A new study in the Journal of Perinatology is drawing attention to one of the most difficult decisions in neonatal medicine: when extremely preterm infants should receive antibiotics for possible early-onset sepsis. The article, led by A. Guiterrez, I. Mir, K. Stumpf and colleagues, examines how sepsis risk can be assessed in babies born at the limits of viability and what those assessments could mean for antibiotic stewardship. The issue is deceptively complex. In the first hours after birth, infection can progress rapidly in a newborn whose immune system is immature, yet the early signs of sepsis can look almost identical to the normal physiological instability associated with extreme prematurity. Clinicians must therefore act before laboratory confirmation is available, balancing the danger of delayed treatment against the consequences of exposing vulnerable infants to unnecessary antimicrobial drugs.
Early-onset sepsis generally refers to a bloodstream or systemic infection that becomes apparent shortly after birth, often as a result of microorganisms transmitted from the mother during labor or delivery. In newborns, the most frequently discussed pathogens include group B Streptococcus, Escherichia coli and other Gram-negative bacteria, although the organisms involved vary according to local epidemiology and clinical circumstances. Extremely preterm infants are particularly challenging to assess because their lungs, skin barrier, gastrointestinal tract and immune defenses are incompletely developed. Respiratory distress, temperature instability, low blood pressure, feeding intolerance, apnea and changes in blood chemistry may all indicate infection, but each can also occur without infection simply because the infant was born prematurely. This overlap makes clinical judgment difficult and has historically encouraged broad, precautionary antibiotic use.
The study’s focus on risk assessment reflects a major change in neonatal care: the movement from treating every possible infection toward identifying which infants are most likely to benefit from immediate antimicrobial therapy. Risk assessment can incorporate maternal, perinatal and neonatal information, including suspected intra-amniotic infection, prolonged rupture of membranes, maternal fever, the circumstances of delivery, gestational age, birth weight, the infant’s clinical condition and early laboratory findings. In principle, combining these variables can help clinicians distinguish a newborn with a high probability of invasive infection from one whose symptoms are more consistent with respiratory immaturity or other noninfectious complications. For extremely preterm infants, however, the challenge is to ensure that a risk model does not falsely reassure clinicians when the cost of missing infection could be catastrophic.
Antibiotic stewardship in the neonatal intensive care unit is not simply a matter of reducing prescriptions. It means selecting the right drug, dose and duration for the infants who need treatment while avoiding exposure in those who do not. Antibiotics can be lifesaving when a bacterial infection is present, but they also alter the developing microbial communities of the intestine and skin. The neonatal microbiome is involved in digestion, immune development and resistance to colonization by harmful organisms. Disruption of these communities has been associated in clinical research with complications such as invasive fungal disease, antimicrobial-resistant infections and intestinal disorders, including necrotizing enterocolitis, although the relationships are biologically complex and influenced by many factors. Every unnecessary dose therefore represents more than a short-term medication decision; it may affect an infant’s ecology and vulnerability during a critical period of development.
The central diagnostic problem is that the tests used to confirm sepsis are imperfect and slow. A blood culture remains the reference method for detecting bacteria in the bloodstream, but it can take time to produce a result and may be negative when the blood volume collected is very small or antibiotics were given before sampling. Biomarkers such as C-reactive protein and procalcitonin can provide additional information, but they do not independently prove or exclude infection. Their concentrations can change because of inflammation, tissue injury, delivery-related stress or normal postnatal physiology. Molecular tests and advanced prediction tools may eventually provide faster answers, yet their usefulness depends on analytical accuracy, appropriate validation and the clinical setting in which they are deployed. In extremely preterm infants, a reliable assessment must account for how rapidly physiology changes during the first hours of life.
A risk-based approach also raises an important question about thresholds. If the threshold for starting antibiotics is set too low, nearly every unstable extremely preterm infant may be treated, creating substantial exposure without guaranteeing better outcomes. If it is set too high, a small but significant number of infants with genuine infection could experience delays in therapy. The safest strategy may therefore involve repeated assessment rather than a single decision made immediately after birth. Initial treatment can be guided by the infant’s condition and perinatal risk factors, followed by review of vital signs, blood culture results, laboratory trends and the infant’s clinical trajectory. Such reassessment is particularly important because early-onset sepsis is a dynamic process, while the decision to continue antibiotics is often made after more information has become available.
The implications extend beyond individual prescriptions to the organization of neonatal care. Successful stewardship requires clear protocols, rapid communication between obstetric and neonatal teams, reliable blood-culture practices and systems that prompt clinicians to reconsider therapy when evidence of infection does not emerge. It also requires careful documentation of why antibiotics were started and why they were continued or stopped. In extremely preterm infants, decisions cannot be reduced to an algorithm alone. A prediction model developed in one hospital may perform differently in another because rates of maternal infection, antibiotic resistance, delivery practices and laboratory procedures vary. Any tool must therefore be externally validated and monitored for missed infections, unnecessary treatment and unequal performance across different patient groups.
The publication arrives at a moment when neonatal specialists are increasingly examining the long-term effects of routine antimicrobial exposure. The goal is not to withhold treatment from infants at genuine risk, but to make treatment more precise. That distinction is crucial in extremely preterm care, where both infection and medication-related harm can be serious. A carefully calibrated risk assessment could help clinicians identify infants who need immediate antibiotics, support earlier discontinuation when cultures remain negative and clinical findings improve, and encourage closer observation for those whose risk is uncertain. The study’s emphasis on stewardship highlights a broader principle of modern medicine: better care does not always mean more intervention. It means matching intervention to biological risk with enough speed, evidence and flexibility to protect the smallest patients.
For families and clinicians, the practical significance of this work lies in its attempt to clarify a decision made under intense uncertainty. Early-onset sepsis cannot be managed safely through fear of infection alone, just as antibiotic stewardship cannot be pursued by ignoring the distinctive fragility of extremely preterm newborns. The emerging model is one of continual risk estimation, in which clinical examination, maternal history, microbiology and laboratory data are combined and updated over time. By bringing these competing priorities into the same framework, the research contributes to a larger effort to make neonatal intensive care both safer and more scientifically disciplined. The challenge now is translating risk assessment into protocols that work at the bedside without delaying lifesaving therapy or normalizing avoidable antibiotic exposure.
Subject of Research: Risk assessment of early-onset sepsis and antibiotic stewardship in extremely preterm infants
Article Title: Risk assessment of early-onset sepsis in extremely preterm infants: implications for antibiotic stewardship
Article References: Guiterrez, A., Mir, I., Stumpf, K. et al. Risk assessment of early-onset sepsis in extremely preterm infants: implications for antibiotic stewardship. J Perinatol (2026). https://doi.org/10.1038/s41372-026-02879-4
Image Credits: AI Generated
DOI: 10.1038/s41372-026-02879-4
Keywords: early-onset sepsis, extremely preterm infants, neonatal intensive care, antibiotic stewardship, risk assessment, antimicrobial therapy, neonatal infection, prematurity
Tags: balancing antibiotic use in newbornsclinical signs of neonatal sepsisdecision-making in neonatal infection managementEarly-onset sepsis risk assessment in preterm infantsimmature immune system in preemiesimpact of early antibiotic exposure on preemieslaboratory confirmation delays in neonatal infectionsneonatal antibiotic stewardshipneonatal sepsis diagnosis challengespathogens in neonatal early-onset sepsisrapid infection progression in preterm infantssepsis risk factors in extremely preterm infants
