Agreement between point-of-care glucometry, blood gas and laboratory-based measurement of glucose in an equine neonatal intensive care unit

Objective: To test the agreement between 3 common methods of glucose measurement in a population of critically ill foals presenting to a neonatal intensive care unit. Design: Prospective clinical study. Setting: University large animal hospital neonatal intensive care unit. Animals: Sequentially admitted critically ill neonatal foals <30 days of age. Interventions: Venous blood obtained from a jugular vein was used for determination of blood glucose concentration using point-of-care (POC) glucometry (GLU), a laboratory chemistry technique (CHEM) and a multi-electrode blood gas analyzer (BG). Paired data were compared using Lin's concordance correlation, Pearson's correlation and robust regression. Bias and limits of agreement were investigated using the technique of Bland and Altman. Measurements and main results: Concordance was significant for all comparisons and was strongest for CHEM-BG while weakest for GLU-BG. Pearson's correlation was excellent for all comparisons: CHEM-BG, 0.98; GLU-BG, 0.94; GLU-CHEM, 0.96. All comparisons had significant robust regression coefficients: CHEM-BG, 0.99; GLU-BG, 0.80; GLU-CHEM, 0.79. For Bland-Altman analysis, mean differences (mean±SD, 95% limits of agreement) were: GLU-BG (-33±18 mg/dL, -68.6 to 1.5); GLU-CHEM (-20±16 mg/dL, -51.0 to 10.9); CHEM-BG (13±11 mg/dL, -8.7 to 34.7). Conclusions: This study demonstrates that glucometry has less than ideal agreement with a laboratory standard and another POC test, blood gas analysis. These differences may be clinically important and decisions regarding management of glucose concentrations in critically ill foals should be made with these differences in mind.