Comparative accuracy of two veterinary-calibrated point-of-care glucometres for measurement of blood glucose concentration in dogs.

OBJECTIVES

To investigate and compare the accuracy of two veterinary portable blood glucose metres (AccuTell; AccubioTech; and AlphaTrak2; Abbott Laboratories).

MATERIALS AND METHODS

One hundred twenty venous blood samples were obtained for immediate whole blood glucose concentration measurement using two portable blood glucose metres designed for use in dogs. Plasma glucose concentration was measured by a laboratory analyser using a hexokinase reaction method. Packed cell volume was measured for each sample. Linear regression analysis was performed and Bland-Altman plots constructed with International Standards Organisation 15197:2013 standards applied to assess relationship and agreement between results of portable blood glucose metres and laboratory analyser methods, respectively. Clarke's error grid analysis was used to assess clinical accuracy and usefulness.

RESULTS

The AccuTell and the AlphaTrak2 had mean differences of -0.69 ± 0.70 mmol/L (bias ±sd) and -1.09 ± 1.22 mmol/L (bias ± sd) with the reference analyser, respectively. Eighty-eight of 120 (73.3%) samples for the AccuTell, and 73 of 120 (60.8%) samples for the AlphaTrak2 were plotted in the zone of the error grid analysis that indicated less clinical risk of misdiagnosis of glucose status. Neither device fulfilled the International Standards Organisation 15197:2013 standards for system accuracy. A statistically significant effect of packed cell volume was identified on the difference between blood glucose concentration obtained by the AlphaTrak2 and the hexokinase reaction method, but not for the AccuTell.

CLINICAL SIGNIFICANCE

Both devices do not meet International Standards Organisation 15197:2013 standards. Users should expect 95% of the samples measured with the AccuTell to be between -0.7 and +2.1 mmol/L higher than actual values, and 95% of the samples measured with the AlphaTrak2 to be between -1.3 and +3.5 mmol/L higher than actual values.