Abstract / Summary
Preterm birth, which is the delivery of an infant prior to 37 weeks of gestation, is associated with neurodevelopmental impairment in children that reach school age, due to unmet needs of metabolites and oxygen in the first period of life. A personalized nutrition approach aimed at maintaining euglycemia may support the child growth and be associated with improved clinical outcomes. However, optimizing glucose control remains a challenging task, and the availability of a nutritional clinical advisor (NCA), capable of recommending optimal feeding strategies to support the infants’ development would be highly beneficial. The first step toward the design of such a tool is to build a model able to simulate reliable key metabolite concentrations in neonates, to safely and effectively test such advisory systems. Here, we developed a neonate glucose simulator capable of generating realistic glycemic time courses potentially usable for in silico optimizing a NCA. The model captures the dynamic interplay between glucose and insulin, assuming that glucose kinetics are similar to those of older individuals but are appropriately adjusted to reflect known neonatal metabolic characteristics. Glucose profiles were validated against Continuous Glucose Monitoring (CGM) traces collected in 19 infants born preterm receiving variable glucose infusion rates to maintain euglycemia. A sensitivity analysis was also performed to assess the impact of key model parameter uncertainty on simulation outcomes. This Neonate Glucose Simulator is currently being employed in the PROMETEUS European project to build a NCA able to maintain euglycemia in virtual neonates.