Successful Performance of Laboratory Investigations with Blood Glucose Meters Employing a Dynamic...

Devices employing electrochemistry-based correction algorithms (EBCAs) are optimized for patient use and require special handling procedures when tested in the laboratory. This study investigated the impact of sample handling on the results of an accuracy and hematocrit interference test performed with BG*Star, iBG*Star; OneTouch Verio Pro and Accu-Chek Aviva versus YSI Stat 2300.

Venous heparinized whole blood was manipulated to contain three different blood glucose concentrations (64-74, 147-163, and 313-335 mg/dL) and three different hematocrit levels (30%, 45%, and 60%). Sample preparation was done by either a very EBCA-experienced laboratory testing team (A), a group experienced with other meters but not EBCAs (B), or a team inexperienced with meter testing (C). Team A ensured physiological pO₂ and specific sample handling requirements, whereas teams B and C did not consider pO₂. Each sample was tested four times with each device. In a separate experiment, a different group similar to group B performed the experiment before (D1) and after (D2) appropriate sample handling training.

Mean absolute deviation from YSI was calculated as a metrix for all groups and devices. Mean absolute relative difference was 4.3% with team A (B: 9.2%, C: 5.2%). Team B had much higher readings and team C produced 100% of "sample composition" errors with high hematocrit levels. In a separate experiment, group D showed a result similar to group B before the training and improved significantly when considering the sample handling requirements (D1: 9.4%, D2: 4.5%, P < 0.05).

Laboratory performance testing of EBCA devices should only be performed by trained staff considering specific sample handling requirements. The results suggest that healthcare centers should evaluate EBCA-based devices with capillary blood from patients in accordance with the instructions for use to achieve reliable results.