Background: The accuracy of glucose meters (GMs) for self-monitoring of blood glucose (BG) is essential for pre-venting insulin dosage errors in insulin injecting diabetic patients. We evaluated the accuracy of 19 commercially available GMs in a clinical setting applying the new planned ISO criteria.
Methods: Capillary blood was taken at 10-minute intervals in insulin dependent patients during induction of a BG increase and decrease by administration of a standardized carbohydrate rich meal and/or injection of an appro-priate dosage of insulin under medical supervision. BG was measured simultaneously by GMs and a laboratory method (Hitado Super GL). The GM results were compared with a laboratory reference method according to the current ISO criteria (ISO 15197:2003) and the proposed (more stringent) criteria.
Results: Evaluating GM results according to the current ISO criteria, all GMs reached the minimum acceptable system accuracy criteria. Analysing GM results according to the more stringent planned ISO standard criteria, only 8 of 19 GMs reached the minimum acceptable accuracy criteria, i.e. ≥ 95% of the individual glucose results fell within ± 15 mg/dL of the results of the reference value at glucose concentrations < 100 mg/dL and within ± 15% at glucose concentrations ≥ 100 mg/dL. When results obtained in BG ranges < and ≥ 100 mg/dL were sepa-rately analysed, readings of 10 and 8 GMs, respectively, fell in the area of acceptable accuracy.
Conclusions: The study shows that more than half of the tested GMs would fail the proposed tighter ISO criteria. Most GMs reveal similar accuracy in BG measurement when compared at lower and higher BG ranges.