[Effects of hemoglobin J-Bangkok traits on measurements of glycated hemoglobin by five methods].

OBJECTIVE: To evaluate the interference of hemoglobin variants J-Bangkok on glycated hemoglobin (HbA(1)c) detected by five measurement systems. METHODS: Seventy cases of blood samples were collected at Zhongshan Hospital of Sun Yat-sen University from July 2012 to January 2014, the blood samples were divided into the normal control group (40 cases) and Hb J-Bangkok variant group (30 cases), and the normal control group was divided into healthy control group (20 cases) and diabetic group (20 cases). HbA(1)c measurement systems were Primus Ultra2, Variant Ⅱ, Variant Ⅱ Turbo, Modular P and Leadman. Based on the standard of the American National Glycosylated Hemoglobin Standardization Program (NGSP), Primus Ultra2 was used as comparative system, and the other 4 systems were test systems. Comparative analysis and bias evaluation were conducted on the results from five detection systems in different groups, statistical analysis were used for evaluating the differences. The estimated average glucose (eAG) was calculated by HbA(1)c values and fasting plasma glucose (FPG) of Hb J-Bangkok variant group with the different detection systems. Deming regression analysis was used to determinate whether Hb J-Bangkok produced significant clinical effect on HbA(1)c results. HbA(1)c ± 10% and relative bias at 6% and 9% HbA1c were evaluation limits. RESULTS: The differences of the 95% confidence interval (95%CI) between the test systems and the comparative system in control group were within ±0.7% HbA(1)c, bias were less than 6%, there were no statistically significant difference (P>0.05). In Hb J-Bangkok group, the eAG calculated from HbA(1)c measured by using Primus Ultra2, Modular P and Leadman were (8.14±2.99), (8.10±3.06) and (8.23±3.00)mmol/L, which had no statistically significant difference compared with FPG ((8.21±3.12)mmol/L, t=0.996, 1.091, 1.479, all P>0.05), and the differences of 95%CI between the results measured by Modular P and the comparative system were all within ±0.7% HbA(1)c, bias were -4.3%-0.4% and -5.2%-4.9%, there were no statistically significant difference (P>0.05). At 6% and 9% HbA(1)c concentrations, the mean differences of the results from the three detection systems were less than the clinically acceptable range. These results showed that the systems of Primus Ultra2, Modular P and Leadman were not affected by Hb J-Bangkok. However, the eAG values calculated from HbA(1)c of Variant Ⅱ and Variant Ⅱ Turbo were (5.58±2.12) and(5.00±2.13)mmol/L, which showed statistically significant lower results compared with FPG level (t=12.29, 13.23 , all P<0.001). Compared with Primus Ultra2, the differences of 95%CI were outside of ± 0.7% HbA1c, bias were -31.9%--12.0% and -42.0%- -17.6% , greater than 6%, showed a negative bias.At 6% and 9% HbA(1)c concentrations, the mean differences of the results were all greater than the clinical acceptable range. These results indicated that Hb J-Bangkok had significantly clinical interference on Variant Ⅱ and Variant Ⅱ Turbo systems. CONCLUSION: Hb J-Bangkok has different interference on different HbA(1)c measurement systems, when performs the HbA(1)c test, clinical laboratory should pay attention to identify Hb variants, and select the appropriate methods to measure the HbA(1)c values in order to prevent the occurrence of interference by Hb variants.