Karen Fairchild, MD and Eric Gibson, MD
Early diagnosis of late onset neonatal sepsis remains challenging due to the complex and often nonspecific presentation of this illness. Predictive informatics for late onset sepsis may help reduce mortality by allowing for earlier and more rapid antibiotic administration.
Late onset sepsis is a leading cause of death in preterm infants. Despite a decrease in incidence over the last few years, late onset sepsis continues to be associated with significant mortality and neurologic morbidity. The key to sepsis prevention may lie in earlier detection of symptoms, which could be accomplished with early warning scores systems combined with algorithms and decision support tools for more accurate treatment. The HeRO and HRC heart rate monitoring systems are examples of currently available early warning systems for sepsis prediction, and a considerable decrease in mortality among patients with blood culture-positive sepsis has been demonstrated with their use. Currently, research efforts are directed at the development of combined cardiorespiratory monitoring which may perform better than heart rate analytics alone as an early warning system for sepsis. Future research efforts for early detection should also include measurements of serum biomarkers and cytokines as a surrogate of inflammation.