Diagnostic Tool Can Use Umbilical Cord Blood To Diagnose Early-Onset Sepsis In Preterm Infants, Study Finds

Healio (7/11, Kellner) reported a study “identified five biomarkers in umbilical cord blood that could indicate early-onset sepsis in preterm infants and then developed a machine learning model that could aid in diagnosing the condition.” The researchers developed a “diagnostic tool using machine learning to predict the odds of early-onset sepsis based on clinical factors and biomarker levels.” The tool found that “elevated levels of C-reactive protein, lipopolysaccharide-binding protein (LBP), serum amyloid A1, leucine-rich alpha-2-glycoprotein 1 and serine proteinase inhibitor A3 showed a ‘clear’ early-onset sepsis signal.” Using these findings, “the researchers evaluated a machine-learning model for identifying babies with early-onset sepsis.” The tool “grouped 41 infants with the healthy control infants and 12 with the early-onset sepsis group. All of the babies predicted to have early-onset sepsis had C-reactive protein, SAA1 and LBP levels close to or above those of the babies with confirmed sepsis.” The study was published in JCI Insight.