Impact of hematocrit levels on the accuracy of specific blood glucose meters: A hospital‐based study

Abstract

ABSTRACTAims/IntroductionBlood glucose meters are commonly used at the bedside, but most of the meters used in Hung Vuong Hospital (Ho Chi Minh City, Vietnam) are built for self‐monitoring and might not be suitable for determining glucose levels in patients. In this study, we aimed to validate the performance of six frequently used meters in our hospital using the Clinical & Laboratory Standards Institute (CLSI) standard, and investigate the hematocrit impact on the accuracy of these meters.Materials and MethodsA total of 135 pregnant women who underwent a 75‐g oral glucose tolerance test consented to participate in the study at Hung Vuong Hospital. Whole blood glucose levels were measured in duplicate using meters, and hematocrit levels were measured using an Alinity h‐series analyzer. Within 5 min, plasma glucose levels were measured twice in a row using the Cobas c502 reference analyzer. For accuracy and precision, the hematocrit effect was assed using CLSI POCT12‐A3.ResultsOut of six evaluated meters, three meters qualified. For CLSI criterion at glucose concentration of 5.55 mmol/L, Accu‐Chek Inform II, Accu‐Chek Performa and OneTouch VerioVue achieved 97.31%, 98.08% and 99.62%, respectively. For CLSI criterion at 4.17 mmol/L, these three achieved 100%. Accu‐Chek Inform II and Accu‐Chek Performa showed an inverse correlation between glucose level and hematocrit with slopes of −0.500 (95% confidence interval −0.678 to −0.322) and −0.396 (95% confidence interval −0.569 to −0.224), whereas OneTouch VerioVue was not affected by hematocrit, with a slope of 0.207 (95% confidence interval −0.026 to 0.440).ConclusionsBlood glucose meters' measurements can be affected by hematocrit, and might provide readings not within an acceptable bias. Medical organizations need to verify or validate before using on patients.