Evaluation of the accuracy of a portable glucose meter compared with laboratory-based spectrophotometric methods

Abstract

Background: The accuracy of portable blood glucose monitoring systems is critical for the safe management and follow-up of Diabetes Mellitus. This study evaluated the accuracy of a capillary glucose monitoring device compared with laboratory-based spectrophotometric analysis of venous blood according to ADA recommendations and selected ISO 15197:2015 accuracy limits. Methods: A cross-sectional study was conducted using 133 paired venous and capillary blood samples. Glucose concentrations were measured using the Humalyzer 2000 and Mindray BA-88A spectrophotometers and the Prodigy AutoCode® glucometer. Data were analysed using Kruskal-Wallis and Spearman correlation tests, with p<0.05 considered statistically significant. Results: The glucometer reported slightly lower glucose values than the laboratory-based spectrophotometric methods. Correlation showed a weak association with the Humalyzer 2000 (r=0.17) and a moderate association with the Mindray BA-88A (r=0.415). Seventy-four per cent of measurements were within the ISO 15197:2015 acceptance limits compared with the Humalyzer 2000, whereas 92% were within the acceptance limits compared with the Mindray BA-88A for glucose concentrations <100 mg/dL. Bland–Altman analysis showed low systematic bias versus both analysers (-0.44 to -3.89 mg/ dL), although agreement was more consistent with the Mindray BA-88A because of narrower limits of agree ment. Greater variability was observed at higher glucose concentrations, particularly with the Humalyzer 2000. Conclusion: The glucometer demonstrated moderate performance and partial agreement with selected ISO 15197:2015 accuracy limits compared with laboratory-based spectrophotometric methods. Increased variability at higher glucose concentrations limits its applicability in settings requiring high analytical precision; therefore, results should be interpreted with caution.