Blood Glucose Monitoring in the Neonatal Period: A Study on Evaluation Using the RAPIDPoint 500/500e.
BACKGROUND Rapid glucose assessment is essential in neonatal intensive care units (NICUs) for early detection of hypoglycemia. We evaluated the agreement between neonatal blood glucose levels measured using bedside blood gas analyzers (GasBS) and those measured in the central laboratory (LabBS), and identified factors associated with clinically significant discrepancies. METHODS We retrospectively analyzed 5436 paired specimens from 2105 infants admitted to the NICU. For the primary agreement analyses, 5375 specimens with GasBS > 20 mg/dL were included. Agreement between GasBS and LabBS was assessed using Bland-Altman analysis, correlation analysis, and linear mixed models. Factors associated with clinically significant discrepancies (absolute difference ≥ 10 mg/dL) were evaluated using generalized linear mixed models. Exploratory receiver operating characteristic analysis was performed to assess the ability of GasBS to detect LabBS < 40 mg/dL. RESULTS GasBS and LabBS showed good agreement, with a mean difference of 1.88 mg/dL (95% limits of agreement, -12.56 to 16.32 mg/dL) and a strong correlation (Spearman's ρ = 0.939, p < 0.001). Lower base excess, lower oxygenated hemoglobin, and higher postnatal age were associated with greater discrepancies, whereas higher hematocrit was associated with an increased risk of clinically significant discrepancies. GasBS also showed excellent performance in detecting LabBS < 40 mg/dL, with an area under the receiver operating characteristic curve of 0.985 and an optimal cutoff of 43.5 mg/dL. CONCLUSIONS Blood gas analyzers provide clinically useful glucose measurements in newborns. However, discrepancies from central laboratory measurements may occur under certain physiological conditions, and confirmatory testing remains essential when severe hypoglycemia is suspected.