The conundrum of early-onset sepsis.

In an ideal world, the practice of medicine would be based on scientific studies guiding the decisions involved in the care of an individual patient. However, clinicians must frequently rely on observational studies and the experiences of other practitioners (the “art of medicine”) because high-quality randomized clinical trials are not available. Within the field of pediatrics, nowhere is that more evident than in the evaluation and treatment of infants with possible sepsis. Scientific studies have identified the risk factors for sepsis (eg, chorioamnionitis), the most frequent pathogens responsible for sepsis (group B Streptococcus and Escherichia coli), the clinical signs associated with infection, the sensitivity and specificity of diagnostic tests, and the toxicities associated with treatment. To date, however, studies cannot accurately tell us whether an individual woman has chorioamnionitis nor whether an individual infant is infected or is instead showing clinical signs compatible with the normal transition to postnatal life or a noninfectious condition. If the treatment of sepsis were completely benign, it would make no difference if every infant with the slightest chance of infection was treated. However, treating an uninfected infant for 5 to 7 days means disrupting maternal bonding for an extended period of time, exposing the infant to drugs with potential toxicities, fostering the development of antibioticresistant flora, and increasing the probability that the infant will experience a more serious morbidity later in the course of hospitalization. In this issue of Pediatrics, Kiser et al describe their experience using published guidelines from the Committee on Fetus and Newborn (COFN) for the evaluation and management of late preterm and term infants born to women with suspected chorioamnionitis. Ninety-six percent of the infants in this study were clinically well at birth, but 20.2% of their population received antibiotic therapy for $7 days solely on the basis of abnormal laboratory data. The COFN recommended continuation of broad-spectrum antibiotics in the neonate with a negative blood culture when the mother had received broadspectrum antibiotics and laboratory data were abnormal. The duration of treatment was not specified. However, the COFN concluded, “Antibiotic therapy should be discontinued at 48 hours in clinical situations in which the probability of sepsis is low.” This combination of statements was confusing and was open to a variety of interpretations. In a subsequent commentary, the COFN reiterated, “Healthyappearing infants without evidence of bacterial infection should receive broad spectrum antimicrobial agents for no more than 48 to 72 hours. Although that recommendation applied to all infants who remained well by 72 hours of life, it was not explicitly stated in the commentary that it also applied to infants born to women with chorioamnionitis. Therefore, after considerable discussion, the COFN modified its recommendations: to not treat a well-appearing term infant with a negative blood culture (whose mother was treated for chorioamnionitis) longer than 48 to 72 AUTHORS: Richard A. Polin, MD,a Kristi Watterberg, MD,b William Benitz, MD,c and Eric Eichenwald, MDd College of Physicians and Surgeons, Columbia University, New York, New York; Department of Pediatrics/Neonatology, University of New Mexico, Albuquerque, New Mexico; Division of Neonatal and Developmental Medicine, Stanford University, Stanford, California; and University of Texas Health Science Center, Houston, Texas