BACKGROUND: Rapid and reliable bed-side determination of blood glucose concentration is very important in the management of acutely ill infants and especially in premature newborns. HemoCue is an easy-to-use glucose analyser. The aim of the present study was to examine the usefulness of the HemoCue glucose analyser compared to a reference plasma glucose method (SYS, BM/Hitachi 747/737) in a neonatal intensive care unit (NICU). METHODS: Seventy-eight consecutive neonates admitted to our NICU were enrolled in the study. At the time of the study all patients were grouped according to nutritional management (parenteral or enteral nutrition), haematocrit values and birth weight. The effects of feeding management, haematocrit values, and birth weight on accuracy and precision of the device were evaluated. RESULTS: Overall data linear regression analysis yielded an r-value of 0.905 and the Bland-Altman method demonstrated that HemoCue overestimates plasma glucose by 0.932 mmol/L. Evaluation of our data by receiver operating characteristic curve demonstrated 100% sensitivity cutoff at 4.1 mmol/L. CONCLUSIONS: HemoCue cannot be used satisfactorily in the management of glycaemia in the NICU. In the preterm population, birth weight had a dramatic influence on HemoCue accuracy. Low haematocrit and parenteral feeding further contributed to a decrease in the accuracy of this device.
BACKGROUND: Rapid and reliable bed-side determination of blood glucose concentration is very important in the management of acutely ill infants and especially in premature newborns. HemoCue is an easy-to-use glucose analyser. The aim of the present study was to examine the usefulness of the HemoCue glucose analyser compared to a reference plasma glucose method (SYS, BM/Hitachi 747/737) in a neonatal intensive care unit (NICU). METHODS: Seventy-eight consecutive neonates admitted to our NICU were enrolled in the study. At the time of the study all patients were grouped according to nutritional management (parenteral or enteral nutrition), haematocrit values and birth weight. The effects of feeding management, haematocrit values, and birth weight on accuracy and precision of the device were evaluated. RESULTS: Overall data linear regression analysis yielded an r-value of 0.905 and the Bland-Altman method demonstrated that HemoCue overestimates plasma glucose by 0.932 mmol/L. Evaluation of our data by receiver operating characteristic curve demonstrated 100% sensitivity cutoff at 4.1 mmol/L. CONCLUSIONS: HemoCue cannot be used satisfactorily in the management of glycaemia in the NICU. In the preterm population, birth weight had a dramatic influence on HemoCue accuracy. Low haematocrit and parenteral feeding further contributed to a decrease in the accuracy of this device.