ABSTRACT

Changes in a device and the modification of measurement methods are frequent issues in medical laboratories. The effects of such modifications on assay results should be investigated. HbA1c is a widely used analyte in treatment and in the follow-up of diabetic patients. We aimed to evaluate the effects of two different assay methods on the detection of the percentage of HbA1c.

We used blood samples with K3-EDTA of 57 diabetic patients who were admitted to our laboratories for the HbA1c assay. HbA1c assays were performed using immunoturbidimetric (Architect C 8000; Abbot Laboratories Inc., Middletown, USA) and ion exchange chromatography (MQ- 2000PT; Shanghai Hui Zhong Medical Technology Co. Ltd., Shangai, China) methods. HbA1c assays were repeated two times in both devices. Results were analyzed using MedCalc software.

Mean HbA1c level in immunoturbidimetric and ion exchange chromatography assays were 6.6 (min:4.1 and max:11.4) and 6.9 (min:4.9 and max:11.8), respectively. In the linear regression analysis, we detected an r value of 0.9533 (r<0.975). In Passing–Bablock analysis, we found the following equation, y=0.4+1.0 x (intercept CI:−0.22−0.68; slope CI:0.97−1.09). We did not observe any constant or proportional systematic errors between the assay methods. We found a 0.37 difference between the two methods in the Bland–Altman graphs of mean HbA1c measurements (Bias 5.7%).

Researches on the harmonization of HbA1c are still increasing worldwide. However, at present, there are variations in methods and devices. NGSP suggests that the difference between methods should not exceed HbA1c±0.70. We found that mean HbA1c results were higher by 0.37 times in ion exchange chromatography assay compared with those in immunoturbidimetric assay. This difference is within the range suggested by NGSP. (JAREM 2015; 5: 52-5)

Keywords: HbA1c, standardization, method comparison