Glucometer versus analyzer: comparable results with negligible clinical risk

Abstract

This study assessed the reliability of Roche Accu-Chek Inform II glucometers in a real-world setting. A retrospective analysis was conducted on 6,695 paired results. Capillary samples were tested using Roche Accu-Chek Inform II glucometers, while venous samples were analyzed using Roche Cobas c503/702 analyzers. Compliance was assessed using modified criteria based on the ISO 15197 guideline and the CLSI EP09-A3 guideline using Passing-Bablok regression analysis, Bland-Altman plots, and Surveillance Error Grid (SEG) analysis. The overall compliance of glucometer results within ±15% or 0.83 mmol/L (15 mg/dL) of the reference method was 81.5%, below the acceptance criterion of 94.6%. SEG analysis showed that 90.3% of the paired results fell within the No-risk zone, with less than 0.001% in the moderate/lower-risk zone. The Emergency Department results indicated 87.8% overall compliance and 92.2% of pairs falling in the No-risk zone. Based on the regression analysis, the glucometer results showed a positive constant bias of nearly 0.33 mmol/L (6 mg/dL). The Bland-Altman plots showed a positive mean difference of 0.43 mmol/L for results ≤5.55 mmol/L (≤100 mg/dL) and a positive mean percentage difference of 3.77% for results >5.55 mmol/L (>100 mg/dL), within the permissible deviation. The compliance values ranged from 76.0% to 90.3% for clinical concentration groups, with the highest compliance found between >16.65-22.20 mmol/L (>300-400 mg/dL). The Accu-Chek Inform II glucometers demonstrated in real-world reliability, with most results falling within acceptable risk categories. However, compliance still needs improvement, so manufacturers should assess opportunities for advancement.

Keywords: Blood glucose; blood glucose analysis; hospitals; post-marketing product surveillance; surveillance error grid.