Comparison of two different device results measuring HbA1c by highperformance liquid chromatography (HPLC) method

Abstract

Aim HbA1c results, which are routinely measured in HA-8180T and HA-8180V model devices in our laboratory for HbA1c measurement, were divided into three groups based on diabetes mellitus diagnostic criteria with the aim of investigating the compatibility between the two devices at different HbA1c concentrations. Methods For HbA1c analysis, the HbA1c values of 260 patients were measured in two devices (Arkray Adams HA-8180T and Arkray Adams HA-8180V) using the ion exchange chromatography method. According to the measured % HbA1c values, 3 groups (1st Group; n=87<5.7%, 2nd Group; n=96 5.7%- 6.4% and 3rd Group; n=77> 6.5%). Results Correlation (r = 0.994, 95% Confidence Interval (CI) = 0.993-0.996, P < .001) of the measurement results obtained between the two devices and Passing-Bablok regression analysis [HA-8180T = 1.0xHA-8180V-0.20] (Slope 95% CI= 1.0-1.0, intercept 95% CI: -0.20-0.20) equation were obtained. According to the regression equation, the linearity between the devices was found to be (cusum test; P = .90). In the Bland-Altman plot to evaluate the compatibility of the two devices, it was observed that the percentage change between the %HbA1c results obtained with HA-8180-T and the %HbA1c results obtained with HA-8180-V was 3% (95% CI: 2.83 – 3.12) higher on average. Conclusion Due to the compatibility of the results measured between the two devices in this study, we think that the use of the HA-8180V device, which has a shorter result time, in laboratories with a higher number of tests may be appropriate in terms of reducing the workload.

Keywords

Introduction

Hemoglobin is a metalloprotein found in erythrocytes, containing iron and having oxygen-carrying capacity. In the normal adult human, the hemoglobin molecule (HbA) accounts for about 97% of hemoglobin from two alpha and two beta chains (α2β2). The terms glycated hemoglobin, HbA1c (A1C test, A1C) are common expressions used to describe the glycation product formed by non-enzymatic binding of hemoglobin (Hb) A0 to the N-terminal (1-deoxyfructosyl) valine amino acid glucose.¹
HbA1c indicates an average blood glucose level of 2-3 months and is a marker used not only to guide the diagnosis and treatment of diabetes but also to assess the quality of patient care and predict the risk of developing diabetes complications.^2-3 The fact that both intra-individual (CVI) and inter-individual (CVG) biological variation of HbA1c is lower than fasting plasma glucose and/or 2-hour plasma glucose (CVI and CVG values for HbA1c and plasma glucose are 1.2% and 5.0% and 4.8% and 8.1%, respectively), does not require pre-test preparation, is not affected by acute stress and has high preanalytic stability makes the HbA1c test advantageous.^1-2-3-4
Device changes and changes in measurement methods are quite common in laboratories. Studies show that there may be significant differences between HbA1c levels determined by different methods.^4-5 Due to the importance of comparable results all over the world, the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) in 1995 and the National Glycohemoglobin Standardization Program (NGSP) in the United States the following year initiated standardization programs for HbA1c measurement. Many different measurement methods have been developed for HbA1c. These methods are based on charge difference (ion exchange chromatography, electrophoresis, capillary electrophoresis, isoelectric focusing) and structural difference (affinity chromatography, immunochemical analyses, enzymatic methods).^6-7-8 HPLC (High-Performance Liquid Chromatography) is one of the most common analytical techniques used to measure HbA1C. In HPLC, ion exchange or affinity columns are used to distinguish HbA1C from other hemoglobin molecules.⁹
HbA1C ≥ 6.5% or ≥48 mmol/mol measured using a method standardized according to the Diabetes Control and Complications Study (DCCT) and certified by the National Glychemoglobin Standardization Program (NGSSP) is one of the diagnostic criteria for diabetes mellitus according to the American Diabetes Association (ADA) criteria. The HbA1C analysis method used by all laboratories in the United States is required to be calibrated according to the high-performance liquid chromatography (HPLC) method, which is the gold standard for HbA1C.^10-11
Depending on the technical features of different models of HPLC devices used in routine HbA1c measurements, such as the yield time (column elution time) and the ability to measure in variant mode, the number of tests run in the laboratory and the yield times may vary. In previous studies, method performance comparison studies of HA-8180T and HA-8180V (ADAMS Arkray) model devices belonging to the same manufacturer have been conducted and it has been reported that the compatibility between both devices is quite good.^12-13 The HA-8180T detects HbA2, HbF, HbA1c (Stable HbA1c, S-A1c), HbS, HbC, HbE, and HbD variants with an analysis time of 210 seconds. The HA-8180V device is a reverse phase cation exchange chromatography device that provides HbA1c and HbF in 48 seconds in fast mode and HbA1c and HbF in 90 seconds in variant mode, as well as HbS, HbC, HbE and HbD detection.
In this study, HbA1c results, which are routinely measured in HA-8180T and HA-8180V model devices in our laboratory and perform HbA1c measurement, were divided into three groups based on diabetes mellitus diagnostic criteria, and the aim was to investigate the compatibility between the two devices at different HbA1c concentrations.

Materials and Methods

Whole blood samples of 260 patients who applied to the laboratory of Ordu Training and Research Hospital for HbA1c analysis were taken into K3-EDTA tubes (VacutainerTM Becton-Dickinson, Rutherford, NJ, USA). This study was approved by the Clinical Researchers Ethics Committee of Ordu University.[Date: 20.01.2023, No: 2023/27] For HbA1c analysis, the HbA1c values of 260 patients were measured in two devices using the ion exchange chromatography method (Arkray Adams HA-8180T and Arkray Adams HA-8180V, Japan). According to the HbA1c values of the patients, 3 groups (1st group <5.7%, 2nd group 5.7%–6.4%, and 3rd group >6.5%) were formed. No variant Hb was observed in any of the patients. Quality control procedures were applied for both devices throughout the study. Two levels (normal and pathological level) of the ICC sample (extendSURE ®) were used daily for internal quality control (ICC). The lot numbers of the controls were the same during the study period for HA-8180T (Lot no: 7125) and HA-8180V (Lot no: 7119). Statistical Analysis All statistical analyses were performed with the MedCalc (version 20.009; Ostend, Belgium) statistical package program. The conformity of the variables to the normal distribution was assessed using the Kolmogorov-Smirnov test. Values were given as mean, standard deviation (SD), median, and Q1-Q3. Regression analysis and Bland-Altman compatibility Graph were used to evaluate the compatibility of the results obtained in both devices with each other. The significant difference between the groups was evaluated at the level of P < .05.

Results

HbA1c values of 260 patients measured in both devices are shown in Table 1. The %HbA1c values of the patient samples measured in both devices according to the groups are shown in Table 2. The correlation between the two devices (r = 0.994, 95% Confidence Interval (CI)=0.993–0.996, P < .001) and the equation [HA-8180T = 1.0 × HA-8180V – 0.20] (Slope 95% CI=1.0–1.0, intercept 95% CI: –0.20–0.20) were obtained in the Passing-Bablok regression analysis of 260 patients (Figure 1). According to the regression equation, the cusum test between the devices was found to be (P = .90). In the Bland-Altman plot to evaluate the compatibility of the two devices, it was observed that the percentage change between the %HbA1c results obtained with HA-8180T and the %HbA1c results obtained with HA-8180V was higher on average by 3% (95% CI: 2.83–3.12) (Figure 2).

Discussion

Hemoglobin A1c is the main parameter for monitoring glycemic control in diabetic patients. In 2010, HbA1c was included in the ADA Diabetes Care Standards as a diagnostic criterion. The World Health Organization concluded in 2011 that HbA1c could be used as a diagnostic test.¹⁴ Due to increased standardization, its use as a diagnostic criterion in diabetes is also increasing. Therefore, it is even more important that HbA1c measurement methods have sufficient diagnostic precision and accuracy and are comparable with other methods.¹⁵
For the diagnosis of diabetes and effective treatment follow-up, HbA1c measurement must be reliable, reproducible, and highly accurate. However, results that do not reflect the correct value in HbA1c measurement may be obtained as a result of various factors that may interfere with the measurement, such as hemoglobinopathies, iron deficiency anemia, or vitamin B12 deficiency.¹⁶ Due to the presence of Hb variants, high or low HbA1c results lead to errors in the diagnosis and treatment of diabetes. Identification of Hb variants during or before the HbA1c measurement process requires the selection of an HbA1c measurement method that is not affected by the variant or derivative in question. Thus, it is possible to measure the glycated Hb accurately.¹⁷
Many methods have been developed for HbA1c measurement. These methods assess the charge (ion-exchange chromatography and electrophoresis) and structural difference (boronate affinity chromatography and immunological tests) between glycolyzed and non-glycolyzed hemoglobin species.¹⁸ While there are studies indicating that there is a very good agreement between HPLC and the turbidimetric inhibition immunoassay (TINIA) method,^19,20 there are also studies indicating that HbA1c values measured by the HPLC method are higher than TINIA.^4,21,22 The reason for this height in the HPLC measurement method is the interaction of the HbA1c peak with other contents and abnormal Hb variants. The HPLC (Ion exchange) method is based on the charge of the globin component of Hb. They stated that since abnormal Hb fragments are less positively charged than HbA, similar to glycosylated hemoglobin, their elution together from the column may affect the measurement results.²¹ In another study conducted by Kın Tekçe et al., they stated that the HbA1c results obtained with the MQ-2000 PT device using the ion exchange chromatography method for HbA1c measurement were measured on average 0.37 higher than Architect C 8000 using the TINIA method, but this difference was within the limits predicted by NGSP.²² In another study by Cihan et al., two different NGSP-approved HbA1c measurement methods (HPLC and TINIA) were compared. The HbA1c levels of the patients were measured in three different hospitals as two devices using the HPLC method and a device using the TINIA method. At the end of the study, the mean HbA1c values measured by HPLC methods were found to be relatively higher than TINIA.⁴
Similar to our study, in a study conducted by Urrechaga, the correlation between the results of the HA-8180T device and the HA-8180V device in the variant mode was found to be compatible, and the regression equation y=1.000x+0.0 (Slope 95% CI=1.000–1.000; intercept 95% CI=−0.00–0.00) was obtained.¹² In this study, in which we evaluated the effect of two different models of ARKRAY HPLC devices using the Ion Exchange Chromatography method on HbA1c values, we determined that the devices can be used interchangeably according to the Passing-Bablok regression equation [HA-8180T=1.0 × HA-8180V – 0.20] (Slope 95% CI=1.0–1.0, intercept 95% CI: –0.20–0.20). In addition, we did not observe any deviation from the linearity according to the equation (Cusum test P = .90). In another study by Urrechaga, HA-8180T and HA-8190V devices were compared and the regression equation obtained was y=1.022x–2.34 (Slope 95% CI=1.010–1.029; intercept 95% CI=−2.67–1.77). In the Bland-Altman compatibility graph, the difference in the averages of the devices was stated as 1.2 mmol/mol (0.11%).¹³ In our study, in the Bland-Altman graph to evaluate the compatibility of both devices, we found that the HbA1c results obtained with HA-8180T were 3% higher on average (95% CI: 2.83–3.12) as a percentage change compared to the HbA1c results obtained with HA-8180V. However, especially in patients with high HbA1c concentration (for Group 3; >6.5% HbA1c), we observed that the percentage change difference between the measurement results of the devices decreased and the consistency between the results increased.
The difference between the measurement methods should not exceed the level of clinical significance determined by the NGSP. The maximum permissible bias for HbA1c is 1.9% and the desired bias is 1.2%.²³ In this study, in which we compared both devices working with the HPLC method, the HA-8180V device was operated in fast mode with shorter results, and it was observed that there was an average of 3% bias between the devices. Although the HbA1c measurement method by HPLC method is the same for both devices, we think that the average percentage bias between the devices may occur due to the different elution times of the HbA1c fractions and the use of different calibrators of the devices. Since the aim of this study is not to compare the performance of the two devices by performing all the verification steps, the fact that all the verification studies of the devices could not be carried out constitutes the main limitation of our study.

Conclusion

In conclusion, in this study, we found that the HbA1c measurement results performed in the fast mode of the HA-8080V device in patients without hemoglobinopathy fit quite well compared to the HA-8080T device, which gives results in a longer time. We think that it may be appropriate to use the HA-8180V device in laboratories with a high number of tests, as it can reduce the workload of laboratories with a short time to give results.

References

1. Little RR, La’ulu SL, Hanson SE, Rohlfing CL, Schmidt RL. Effects of 49 different rare Hb variants on HbA1c measurement in 8 methods. J Diabetes Sci Technol. 2015;9(4):849-856. doi:10.1177/1932296815572367
   Article · PubMed · Google Scholar
2. Sacks DB, Bruns DE, Goldstein DE, Maclaren NK, McDonald JM, Parrott M. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Clin Chem. 2002;48(3):436-472. doi:10.1093/clinchem/48.3.436
   Article · PubMed · Google Scholar
3. Goldstein DE, Lorenz RA, Malone JJ. Tests of glycemia in diabetes. Diabetes Care. 2000;23(1):80-82.
   PubMed · Google Scholar
4. Cihan M, Üner A, Kılınçkaya MF, Noyan T. Comparison of high-performance liquid chromatography and turbidimetric inhibition immunoassay methods for measurement of hemoglobin A1c. J Health Sci Med. 2022;5(5):1281-1285.
   PubMed · Google Scholar
5. Genc S, Omer B, Aycan-Ustyol E, Ince N, Bal F, Gurdol F. Evaluation of turbidimetric inhibition immunoassay and HPLC methods for glycated haemoglobin determination. J Clin Lab Anal. 2012;26(6):481-485. doi:10.1002/jcla.21550
   Article · PubMed · Google Scholar
6. Cohen MP, Witt J, Wu VY. Purified haemoglobin preparations in the evaluation of HbA1c determination by ion exchange chromatography. Ann Clin Biochem. 1993;30:265-271. doi:10.1177/000456329303000307
   Article · PubMed · Google Scholar
7. Marshall SM, Barth JH. Standardization of HbA1c measurements: a consensus statement. Ann Clin Biochem. 2000;37(1):45-46. doi:10.1258/0004563001901506
   Article · PubMed · Google Scholar
8. Kurt İ. Glikozile hemoglobin (HbA1c) ölçümü ve diabetes mellitusun uzun dönem glisemik kontrolünde kullanılması. Gulhane Med J. 2003;45(4):387-395.
   PubMed · Google Scholar
9. Gupta S, Jain U, Chauhan N. Laboratory diagnosis of HbA1c: a review. J Nanomed Res. 2017;5(4):00120.
   PubMed · Google Scholar
10. American Diabetes Association. 2. Classification and diagnosis of diabetes: standards of medical care in diabetes—2021. Diabetes Care. 2021;44(suppl 1). doi:10.2337/dc21-s002
    Article · PubMed · Google Scholar
11. Noyan T, Üner A, Cihan M. Interference in the measurement of hemoglobin A1c: presence of hemoglobin D. Turk Klinik Biyokimya Derg. 2022;20(3):132-135. doi:10.56615/tkbd.2022.15
    Article · PubMed · Google Scholar
12. Urrechaga E. Analytical evaluation of the ADAMS A1c HA-8180T analyzer for the measurement of HbA1c. J Clin Lab Anal. 2018;32(1). doi:10.1002/jcla.22155
    Article · PubMed · Google Scholar
13. Urrechaga E. Analytical evaluation and quality assessment of the ARKRAY ADAMS A1c HA-8190V for HbA1c. Scand J Clin Lab Invest. 2023;83(2):136-142. doi:10.1080/00365513.2023.2167231
    Article · PubMed · Google Scholar
14. van der Hagen EAE, Leppink S, Bokkers K, Siebelder C, Weykamp CW. Evaluation of the ARKRAY HA-8190V instrument for HbA1c. Clin Chem Lab Med. 2021;59(5):965-970. doi:10.1515/cclm-2020-1300
    Article · PubMed · Google Scholar
15. Miedema K. Laboratory tests in diagnosis and management of diabetes mellitus: practical considerations. Clin Chem Lab Med. 2003;41(9):1259-1265. doi:10.1515/cclm.2003.193
    Article · PubMed · Google Scholar
16. Seydel GŞ, Güzelgül F. The effect of hemoglobin variants on high-performance liquid chromatography measurements of glycated hemoglobin. Int J Med Biochem. 2021;4(1):25-28.
    PubMed · Google Scholar
17. Bry L, Chen PC, Sacks DB. Effects of hemoglobin variants and chemically modified derivatives on assays for glycohemoglobin. Clin Chem. 2001;47(2):153-163. doi:10.1093/clinchem/47.2.153
    Article · PubMed · Google Scholar
18. John WG, Mosca A, Weykamp C, Goodall I. HbA1c standardisation: history, science, and politics. Clin Biochem Rev. 2007;28(4):163-168.
    PubMed · Google Scholar
19. Sriwimol W, Choosongsang P, Choosongsang P, et al. Strong correlation and high comparability of capillary electrophoresis and 3 different methods for HbA1c measurement in a population without hemoglobinopathy. Scand J Clin Lab Invest. 2020;80(2):139-150. doi:10.1080/00365513.2019.1703213
    Article · PubMed · Google Scholar
20. Coskun A, Yavuz Ö, Memisogullari R, Kurt H. HbA1c ölçümünde HPLC ve türbidimetrik immüno-inhibisyon yöntemlerinin karşılaştırılması. Duzce Med J. 2007;9(1):15-19.
    PubMed · Google Scholar
21. Ozcelik F, Yiginer O, Serdar M, et al. Comparison of 3 methods for measurement of HbA1c. Turk J Biochem. 2010;35(4):344-349.
    PubMed · Google Scholar
22. Tekçe BK, Tekçe H, Aktaş G, Tosun M. HbA1c ölçümünde Architect C8000 ile MQ-2000PT sonuçlarının karşılaştırılması. JAREM. 2015;5(1):52-55.
    PubMed · Google Scholar
23. Ricós C, Fernández-Calle P, Gonzalez-Lao E, et al. Critical appraisal and meta-analysis of biological variation studies on glycosylated albumin, glucose, and HbA1c. Adv Lab Med. 2020;1(3):20200029. doi:10.1515/almed-2020-0029
    Article · PubMed · Google Scholar

Tables

Additional Information

Publisher’s Note
Bayrakol MP remains neutral with regard to jurisdictional and institutional claims.

Rights and Permissions