Technical and clinical accuracy of three blood glucose meters: clinical impact assessment using error grid analysis and insulin sliding scales

Aims: Manufacturers of point-of-care testing (POCT) glucose systems must adhere to International Organisation for Standardisation (ISO) 15197 performance standards, a tightened version of which becomes mandatory in May 2016. Using methods deviating from the evaluation requirements of ISO 15197, we investigated the accuracy and clinical utility of three glucose testing systems.

Methods: Whole blood glucose was measured on 62 EDTA-preserved samples with each system followed by plasma glucose determination with a laboratory hexokinase method. Technical accuracy was assessed using linear regression, Bland-Altman plots and bias. Clinical utility was evaluated using consensus error grid analysis and insulin dosing algorithms.

Results: The correlation coefficient of the glucose measuring devices with the comparator method was between 0.979 and 0.993 (n=61). Bias against the reference method was +0.7-+4.4%. The FreeStyle Lite (0.7±4.4%) and ACCU-CHEK Aviva (1.3±7.3%) systems had the least bias to the reference. Error grid analysis depicts all meters to be clinically accurate, with all results falling within Zones A and B.

Conclusions: Conformité Européene-marked devices are presumed to have adequate accuracy, but postmarket evaluations are imperative to assess total system accuracy and whether regulatory targets are achieved.