Objective To evaluate the interference of hemoglobin variant E (HbE) on five kinds of glycosylated hemoglobin (HbA1c) system. Two of them were ion-exchange high performance liquid chromatography (IE-HPLC) methods, one was affinity chromatography high performance liquid chromatography (AC-HPLC) method, one was turbidimetric inhibition immunoassay method (TINIA), one was enzyme (EM) method. Methods All 60 blood samples from May 2012 to October 2013 were collected from Zhongshan Hospital of Yat-sen University, and then divided into normal control group (20 cases), diabetic group (20 cases) and HbE group (20 cases).Variants were used to detect the whole blood concentration of HbA1c by five detection systems. Based on the judgment standards of National Glycohemoglobin Standardization Program (NGSP), comparison analysis and bias evaluation results for different groups of 5 kinds of detection systems were estimated. The statistical difference of mean blood glucose (eAG) and fasting blood glucose (FPG) was calculated using the estimation of HbE variant group detection system HbA1c results. The test results were analyzed by Deming regression analysis, to determine whether HbE has significant clinical effect on the results of HbA1c, using 6% HbA1c and 9% ± 10% relative error as the evaluation scope. Results The 95% confidence interval (95%CI) of the 4 kinds of detection system in normal control group and diabetic group were within ±0.7% HbA1c deviation, which was less than 6%, comparing to NGSP I laboratory certification, indicating no significant difference of determination results(P>0.05). There were no significant difference between eAG and FPG levels in HbE group using AC-HPLC, Variant HbE variant Ⅱ IE-HPLC method and TINIA method(P>0.05).While the difference between eAG level and FPG level of Variant Ⅱ Turbo IE-HPLC and enzymatic method was statistically significant(P 0.05). Also when in the 6% and 9% HbA1c concentrations, detection method and average differences between control samples were less than the clinically acceptable range,indicating that the interference of these 2 methods was not affected by HbE. The difference of 95% CIof Bio-Rad Variant Ⅱ Turbo IE-HPLC method and enzyme method were located outside the ± 0.7% HbA1c, comparing to AC-HPLC method. The determinations of the deviation were 13.9%- 40.1 % and -34.1%--1.3% respectively, which were greater than 6%. The difference was of statistically significance(P<0.01). Comparing to the control system, Bio-rad Variant ⅡTurbor IE-HPLC held the positive deviation, while the enzyme method held the negative deviation. Also when in the 6% and 9% HbA1c concentrations, the mean differences of samples under detection method and control method were both greater than the clinically acceptable range, indicating that HbE has significant clinical interference in these 2 kinds of methods. Conclusion HbE had different interference effects in clinical laboratory for HbA1c test, so one should pay attention to this kind of Hb variance in daily practice. Appropriate method should be selected to prevent the occurrence of interference to the determination or to find a surrogate marker of HbA1c.(Chin J Lab Med,2014,37:921-927) Key words: Hemoglobin E; Hemoglobins,abnormal; Hemogbobin A,glycosylated