Abstract

        Background and Objective: According to recent changes in diagnostic criteria for diabetes, the harmonization of results obtained from various methods and systems by considering their accuracy and precision is essential. This study aimed to evaluate the accuracy, precision and consensus of some routine laboratory glucose kits in comparison with Hexokinase reference method.

           Material and Methods: The participants were 38 diabetic patients with fasting blood sugar (FBS) ≥126 mg/dl, nine prediabetic patients with FBS of 100-125 mg/dl, 15 non-diabetic people with FBS of 60-100 mg/dl and 9 hypoglycemic patients with FBS of ≤60 mg/dl. Their FBS were measured by four routine laboratory glucose kits:    Glucose oxidase on BT3000 analyzer with an open system and Hexokinase reference method on a close system (COBAS INTEGRA^®400plus analyzer, Roche kit). Accuracy and precision were determined and compared with reference method.

         Results: Glucose oxidase methods showed a good agreement with the reference method, in Correlation Coefficient>0.99. based on  regression analysis, the  slope of 1.114 for Pars Azmoon, 1.105 for Bionik, 1.121 for Elitech and 1.087 for Human were reported (P<0.05). Error of the mean for ParsAzmoon was 12.79, for Bionik 10.86, for Elitech 12.58 and for Human were 8.46. Coefficient of Variation   showed more imprecision for Bionik and Human kits.

       Conclusion: given the same almost standard errors, standard devisions and regression analysis, the precision in four methods is the same but in comparison with Hexokinase, reference method has not the accuracy.

          Keywords: Blood Glucose, Glucose Oxidase, Hexokinase, Methods, Consensus

ABSTRACT

        Background and Objective: The current challenge of diabetes mellitus is to prevent its complications. These complications are directly associated with hyperglycemia in diabetics. The HbA1c measurement is essential for long-term glycemic control. Synchronization of HbA1c measurement is important in order to avoid discrepancies between results reported by laboratories. This study aimed to evaluate the accuracy, precision and agreement of five HbA1c measurement methods with HPLC reference method.

       Methods: HbA1c levels of 55 samples were measured using six methods of microcapillary electrophoresis (Sepia), enzymatic method (Pishtaz Teb), immunoturbidometry (Pars Azmoon), boronate affinity (Nycocard), immunofluorescence (ichroma) and Tosoh G8 HPLC.

       Results: The five tested methods showed a good agreement with the HPLC method with correlation coefficient of less than 95%. Regression testing of HPLC method and other methods showed slope of 0.99 (P<0.05) for Sebia, 1.02 (P<0.05) for Pishtaz Teb, 0.79 (P<0.05) for Pars Azmoon, 0.82 (P<0.05) for Nycocard and 0.89 (P<0.05) for ichroma. Average inaccuracy for the Sebia, Pishtaz Teb, Pars Azmoon, Nycocard and ichroma in comparison with the HPLC reference method were -0.09, -0.004, -0.75, -0.79 and -0.78, respectively.

         Conclusion: The Sebia microcapillary method and Pishtaz teb enzymatic method have appropriate accuracy and precision. Therefore, these methods can be used as alternatives to the HPLC method for HbA1c measurement. Other methods such as Pars Azmoon, Nycocard and ichroma have significant shortcomings in terms of accuracy.