Abstract: Objective：To investigate the test methods and the comparability of the results of glycosylated hemoglobin A1c (HbA1c) from the clinical laboratories in Nanjing, analyze the inter-laboratory variation and the determination deviation, and explore the possible sources of error and improvement measures in order to provide the evidence for the mutual accredit of results from different laboratories. Methods：Fresh whole blood samples were mixed and target value was established. The prepared blood samples were distributed to 35 laboratories in Nanjing, and the determinations for HbA1c were completed within 24 hours. The samples were divided into different groups according to methodology, and statistical analysis for the results was performed. The acceptability of the results was decided according to the criteria recommended by the College of American Pathologists (CAP), i.e., the deviation within the range of 8% of target value and the "acceptable total error" within the range of 3.9% of target value. As the example, 6.5% of target value was designated as "level 2" to observe the distribution of the test results. Results： The coefficient of variation (CV) in the group of high performance liquid chromatography (HPLC), immunoturbidimetry, low pressure liquid chromatography (LPLC) and the microparticle color spectrum were 2.91% to 4.48%, 9.66% to 11.10%, 11.6% to 16.1% and 10.4% to 20.3%, respectively. The qualified rate of the test results from the laboratories of the topgrade hospitals and the secondary hospitals were 78.6% and 31.5%, respectively, judged by the criteria recommended by CAP, i.e., deviation within the range of 8% of target value. However, the qualified rates were 42.8% and 10.5% respectively judged by the "acceptable total error" standard (the range of 3.9% of target value). For the values of "level 2", the qualified rates were 71.4% (25/35) and 42.9% (15/35) respectively. Conclusion：The HbA1c results from the laboratories of the top-grade hospitals were more of a consensus, but there was large difference in the secondary hospitals. The disparity of methodology for determining the HbA1c level was really in presence, but the strict quality control for intralaboratory and keeping the external quality assessment of the test results were still the most important for maintaining the accuracy and conformity. The consciousness of quality control should be strengthened in all the laboratories, especially for the laboratories of the secondary or primary hospitals, and the contrast between laboratories should be extended to primary medical institutions.