Precision and accuracy of blood glucose measurements using three different instruments

Assessment of insulin sensitivity by dynamic metabolic tests such as the hyperinsulinemic euglycemic clamp critically relies on the reproducible and fast measurement of blood glucose concentrations. Although various instruments have been developed over the last decades, little is known as to the accuracy and comparability. We therefore compared the best new instrument with the former gold standard instruments to measure glucose concentrations in metabolic tests.

Fasting blood samples of 15 diabetic and 10 healthy subjects were collected into sodium-fluoride tubes, spiked with glucose (0, 2.8, 6.9 and 11.1 mmol/l) and measured either as whole blood (range 3.3-26.3 mmol/l) or following centrifugation as plasma (range 3.9-32.0 mmol/l). Plasma samples were analyzed in the YSI-2300 STAT plus (YSI), EKF Biosen C-Line (EKF) and the reference method, Beckman Glucose analyzer-II (BMG), whole blood samples in EKF instruments with YSI as reference method.

The average deviation of the EKF from the reference, BMG, was 3.0 ± 3.5% without any concentration-dependent variability. Glucose measurements by YSI were in good agreement with that by BMG (plasma) and EKF (plasma and whole blood) up to concentrations of 13.13 mmol/l (0.5 ± 3.7%), but deviation increased to -6.2 ± 3.8% at higher concentrations. Precision (n = 6) was ±2.2% (YSI), ±3.9% (EKF) and ±5.2% (BMG).

The EKF instrument is comparable regarding accuracy and precision to the reference method BMG and can be used in metabolic tests, while the YSI showed a systematic shift at higher glucose concentrations. Based on these results we decided to replace BMG with EKF instrument in metabolic tests.