Reference: J Pediatr. 2021 Jul;234:77-84
Early-onset neonatal sepsis can result in considerable morbidity and mortality in as many as 18% of affected newborns. A concerted effort by public health officials and clinicians caring for pregnant patients and their newborns resulted in a significant decline in early-onset neonatal sepsis (EOS) in the 1990s. Rates among term newborns declined to near 0.1%. To reduce potential overtesting and overtreating for an increasingly rare condition, many clinicians advocate for the use of the Early-Onset Sepsis Calculator, which demonstrated efficacy and safety in an initial before and after study. A previously published systematic review found this calculator reduced the need for empiric antibiotics by a relative risk reduction of 55%, seemingly without missing cases of EOS.
These same investigators conducted a follow-up systematic review and patient-level data meta-analysis of cohort studies published after the original derivation study. Studies of high-risk infants and studies without a single case of EOS were excluded. The investigators included a total of 18 cohort studies, mostly from the US, resulting in a population of over 400,000 live births and a total of 234 cases of EOS. EOS, defined as a newborn with a positive blood or cerebrospinal fluid culture within 72 hours of life, was most commonly caused by Group B streptococcus and Eschericia coli (51% and 16%, respectively). Overall risk of bias was assessed as high in 11 of 18 studies and the quality of evidence for the primary outcome of EOS was low due to the retrospective nature. Using patient-level data, the EOS Calculator recommendations were determined at the population rate of 0.6 cases per 1000 live births at initial assessment and 12 hours later. The EOS Calculator recommended routine care for 103 of 234 EOS cases (44%, 95% CI 37.6%-50.6%) at initial assessment and 65/234 (27.8%, 95% CI 22.1%-34.0%) at 12 hours. Among the infants with EOS who were clinically well-appearing or had a low-risk estimate on initial assessment, clinical illness occurred by 48 hours in 76.9%. A total of 31 infants remained clinically well and were diagnosed with EOS by cultures obtained for other reasons.
Unnecessary antibiotic exposure early in life can increase bacterial resistance, disrupt the microbiome, and has been associated with other long lasting effects. Excess testing and monitoring increases newborn length of stay and may increase parental anxiety and vulnerable child syndrome. The Kaiser EOS Calculator provides hard ‘numbers’ and recommendations for further steps in management—-a coveted tool. This meta-analysis questions whether this clinical decision support tool accurately stratifies infants, however. Limitations of this review include high risk of bias in most included studies and inclusion of retrospectively identified known EOS cases. However, even among the prospective cohort data, a considerable number of babies with EOS were recommended to have routine care at birth and 12 hours (40.0% and 23.3%, respectively). Also, data on true negatives were not reported, so the negative predictive value (NPV) of the EOS calculator remains unknown. We’re guessing the NPV would have been very high. Although imperfect, these data appear to represent implementation of the EOS Calculator in the real world—-where clinical judgment and patient factors vary considerably from the original design. The conservative clinician would take the EOS Calculator recommendations into account but carefully monitor for clinical change during the first two days of life.
For more information, see the topic Early-onset Neonatal Sepsis in DynaMed.
DynaMed EBM Focus Editorial Team
This EBM Focus was written by Carina Brown, MD, Assistant Professor at Cone Health Family Medicine Residency. Edited by, Alan Ehrlich, MD, Executive Editor at DynaMed and Associate Professor in Family Medicine at the University of Massachusetts Medical School; Dan Randall, MD, Deputy Editor at DynaMed; Katharine DeGeorge, MD, MS, Associate Professor of Family Medicine at the University of Virginia and Clinical Editor at DynaMed; Nicole Jensen, MD, family physician at WholeHealth Medical; Tanya Tupper, MM, RT(N), CNMT, PET, Senior Medical Writer at DynaMed; Vincent Lemaitre, PhD, Medical Writer at DynaMed, and Christine Fessenden, Editorial Operations Assistant at DynaMed.