Abstract
Background:
HbA1c result provide information on metabolic control in diabetes mellitus (DM) and could also be used for its diagnosis. For its determination, the laboratory must be certified by the National Glycohemoglobin Standardization Program (NGSP) or the International Federation of Clinical Chemistry (IFCC) and comply with a strict quality control program.
Aims:
To determine the correlation and agreement between HbA1c results measured by three analytical methods (enzymatic, turbidimetric, and capillary electrophoresis) versus HPLC.
Methods:
Method comparison—1245 samples from equal number of subjects at 45 Association of High Complexity Laboratories (Asociación de Laboratorios de Alta Complejidad—ALAC) centers, centralizing sample processing and operator. Statistical analysis—analysis of variance (ANOVA) and nonparametric Friedman ANOVA test for related samples, means, and medians. Correlation and concordance—Pearson’s correlation and linear regression, intraclass correlation coefficient (Passing and Bablock and Bland and Altman).
Results:
The comparison of mean values obtained by the four methods showed statistically significant, but clinically irrelevant, differences: HbA1c by HPLC versus Electrophoresis 0.06% (0.42 mmol/mol) P = .000 (± 1.96 DS -0.070 -0.047), Enzymatic 0.087% (1 mmol/mol) P = .000 (± 1.96 DS 0.077 0.098), Turbidimetric 0.056% (0.38 mmol/mol) P = 0.000 (± 1.96 DS -0.067 -0.044). Their concordance showed intraclass correlation of single measures of 0.982 P < .001 (95% CI 0.987 - 0.9838).
Conclusions:
The three methods present low variability and high correlation versus the HPLC.
Keywords: capillary electrophoresis methods, electrophoresis methods, enzymatic methods, glycated hemoglobin, HbA1c, method comparison, turbidimetric methods
Introduction
Diabetes mellitus (DM) is a very frequent metabolic disorder characterized by insulin resistance and/or very low hormone levels in relation to the increasing demand of peripheral tissues, all of which leads to permanent hyperglycemia. 1
The steady increase of type 2 diabetes (T2D) frequency depends on lifestyle changes associated with overweight and obesity worldwide. 2
The delay in the diagnosis of T2D, which could take several years, implies a greater risk of developing macro and microvascular events that decrease the quality of life of people and increase the cost of their treatment. 3 Therefore, access to validated tools for early diagnosis is necessary to be able to administer a treatment capable of preventing or delaying the appearance of these complications. The assessment of HbA1c plays an important role both for the diagnosis and for the subsequent metabolic control of people with DM. 4
The American Diabetes Association (ADA) confirmed the usefulness of HbA1c as a diagnostic method as of 2010 5 based on conclusions drawn by an international committee of experts whose members had been appointed by the ADA, the European Association for the Study of Diabetes (EASD), and the International Diabetes Federation (IDF). 6
Per ADA recommendations, the following cutoff values were established: normal ≤5.6% (38 mmol/mol), ruling out DM diagnosis; between 5.7% (39 mmol/mol) and 6.4% (46 mmol/mol), prediabetes; ≥ 6.5% (47 mmol/mol) compatible with DM diagnosis. In 2011, the World Health Organization (WHO) acknowledged the usefulness of HbA1c for DM diagnosis. 7
The measurement of HbA1c, carried out by reliable methods, would offer advantages over the determination of fasting plasma glucose or 75-g oral glucose tolerance test (OGTT), since the blood sample can be taken at any time of the day and without requiring the fasting condition. Complementarily, the molecule is more stable, shows lower intraindividual variation, and enables predictions regarding the development of DM-related illnesses.1,8
Other advantages of HbA1c are the absence of short-term variations (caused by food consumption or stress exercises, among others), which makes the study more comfortable for patients and enables evaluation of the results of patient exposure to hyperglycemia in recent weeks. This allows the health team to make a global assessment within a given period of time. HbA1c disadvantages center on variations between different ethnic, age, and gender groups, and on the impact of the quantification of certain pathologies such as anemia and hemoglobinopathy, among others. 9
Other diagnostic and follow-up criteria are recommended in cases of pregnancy, recent blood loss, transfusions, or hemolytic anemia. 4
Laboratories specializing in clinical analyses play a key role in the diagnosis and follow-up of patients suffering from DM; however, these laboratories must guarantee standardization of the values they determine through the application of a strict quality control system.
This standardization of HbA1c determination is so important that the Standardization Committee of the American Association of Clinical Chemistry created a Subcommittee for the Normalization of HbA1c in 1996. 10 Its objective was to devise a plan by which clinical laboratories would be able to compare their HbA1c results with those obtained by the Diabetes Control and Complications Trial (DCCT), and thereby demonstrate the correlation between HbA1c values and the development and increase in chronic illnesses caused by DM. 11
Therefore, the methodology for assaying HbA1c must be certified by the National Glycohemoglobin Standardization Program (NGSP) and normalized or extrapolated to DCCT. As a complement, in 2004, the International Federation of Clinical Chemistry and the Laboratory Medicine Work Team (IFCC), which subscribed to the methodological traceability concept, set up a reference measurement procedure (RMP) for HbA1c. 12
We also recommend that laboratories maintain a quality control system comprising internal controls, and also participate in an external quality control program. In the event that a laboratory uses a reference value to trace changes, a ≤2.0% analytic coefficient of variation indicates that there is a 95% probability of an HbA1c ≥0.5% (3 mmol/mol) difference between successive blood samples from the same patient resulting from a significant change in glycemic control. 13
Considering all the above, our study evaluated the correlation and agreement of HbA1c results as measured by four analytic methods applied to the same samples (HPLC, Enzymatic, Turbidimetric, and Capillary Electrophoresis). Samples were processed in the same laboratory in a centralized fashion.
Material and Method
Study Population
For this study, blood samples were taken from 1245 males and females aged ≥18 who, after being properly informed, consented to participate in this study. Each of the 45 laboratories specialized in clinical analyses in the Argentine Republic taking part in this study recruited 30 patients who went to these laboratories as volunteers. These patients were randomly selected.
Criteria for inclusion of volunteers were people aged ≥18, and each laboratory informed them about this multicentered study, its characteristics, and implications. Once all information related to this study was properly read and questions were duly answered, the participants signed a written consent according to Good Clinical Practice.
Exclusion criteria were people <18 years old; pregnant women; patients with anemia or hemoglobinopathies (the volunteers were questioned about this and all those who were unaware of the response were studied by HPLC Varian II Turbo—BIORAD); patients with recent blood loss or recent blood transfusions; patients on insulin treatment or exposed to corticosteroids, high doses of aspirin (more than 1000 mg/day), vitamin C, iron intake, or immunosuppressive drugs; and those with chronic kidney disease, addictions, or active cancer diseases.
All participating centers have an ISO 9001:2015 certification, operate with certified internal quality controls, and take part in the “Buenos Aires” International Program to Guarantee External Quality Control of Clinical Analyses (Programa Internacional “Buenos Aires” de Aseguramiento Externo de Calidad en Análisis Clínicos—ProgBA–CEMIC). The Hidalgo laboratory method for analyzing HbA1c is Bio-Rad Variant II Turbo 2.0, with Level I Laboratory certification.
Procedures
This comparison method study followed an open cross design including centralized analyses of blood samples sent in by the laboratories that functioned as peripheral blood collection centers.
For the centralized processing of HbA1c, the BAIRESLAB laboratory, located in the Autonomous City of Buenos Aires, was home to the most representative equipment from Argentine laboratories specialized in clinical analyses. The other centers that participated in the recruitment of patients are members of the Association of High Complexity Laboratories (Asociación de Laboratorios de Alta Complejidad—ALAC). ALAC is a network that brings together highly complex clinical analysis laboratories installed in provincial capitals and main cities of Argentina. It is the only network of laboratories nationwide certified under ISO 9001: 2015 standards.
All these blood samples were then sent to BAIRESLAB. Blood samples were kept refrigerated at temperatures between 2ºC and 8ºC. Refrigeration was constant from extraction date to processing, a time lapse never longer than five days. Each blood sample was processed on the same day by the same operator applying all four methods. Results were automatically uploaded to a data bank.
For determination of HbA1c values, four methods were applied to the same sample: HPLC (BIORAD), Enzymatic (Abbott), Turbidimetric (Roche), and capillary electrophoresis (SEBIA). For each analyzer, the same reagent lot was used in all determinations.
To guarantee quality, the equipment used complied with the manufacturer’s internal quality control and with the external quality control as stipulated by ProgBA–CEMIC.
Once the samples arrived at the main laboratory center, all HbA1c determinations were processed on the same day by applying the four methods in the same conditions, in order to eliminate any possible inter-laboratory or inter-operator variations.
The methods selected for the measurement of HbA1c were high performance liquid chromatography (HPLC) in VARIANT II TURBO (BIO-RAD), Turbidimetric immunoassay performed with COBAS C501 (ROCHE), Enzymatic method process using ARCHITECT C8000 (ABBOTT), and Electrophoresis capillary made in CAPILLARYS (SEBIA). These are the most representative methods in the local market and are the most used in the laboratories belonging to ProgBA, with the exception of Capillary Electrophoresis, which was included since it was the last method introduced in our market.
Prior to the beginning of this work, we made in collaboration with a NGSP-certified laboratory (Laboratorio Hidaldgo, Buenos Aires) a comparison between our HPLC Variant Turbo II (BIORAD) and their equipment. We have run samples and calibrators twice in both chromatographs; this comparative analysis showed a lineal determination coefficient of 99.7% (data not shown).
Statistical Analysis
The programs IBM SPSS Statistics 26, MedCalc Version 12.5, Stata 14.2, and Epidat version 4.2 were used for statistical analysis.
Descriptive statistical methods were used for quantitative data expressed as mean ± standard deviation and their corresponding medians as well as for qualitative data, expressed in percentages. For the sake of inference, 95% confidence intervals were also calculated, not only regarding proportions but also mean data.
Differences among mean data obtained through various techniques were evaluated by Analysis of variance (ANOVA) for repeated measures; the Bonferroni procedure was the “post hoc” test. Whenever applicable, the Friedman nonparametric technique was used to detect significant differences.
To evaluate the different methods, the following was applied: (a) Pearson linear correlation and lineal regression coefficients; (b) agreement analysis through the intraclass correlation coefficient; (c) the Passing and Bablok regression method—based on the rank and order principle that takes into account observational errors in the contrast between two analytic methods, the slope of the regression line being calculated as the median of all possible slopes; and (d) complementation of analyses by Bland and Altman graphics.
Values of P < .05 (two tails) were considered statistically significant.
Ethical Aspects
Written consent after proper informative sessions was signed by all patients who met the requirements of the inclusion criteria, unlike those who fell into the exclusion criteria group, none of whom signed this written consent. All procedures followed national and international regulations and complied with the recommendations of the Helsinki Declaration. The protocol was approved by the Ethics Committee of the Foundation for Pharmacological and Medicinal Studies (Comité de Ética de la Fundación de Estudios Farmacológicos y Medicamentos—FEFyM)
Results
One thousand two hundred and forty-five people participated in this study, 63.4% of whom were female (IC 95%: 60.7-66.1). Average age was 49 and 45 for men and women, respectively, with a significant statistical difference (P < .01).
According to body mass index (BMI), 0.9% were underweight, 29.6% had normal weight, 33.1% were overweight, and 36.4% were obese; the number of people with normal weight was significantly higher among women than in men (37.9 vs 15.4%, P < .001).
Only one patient in this sample registered high fetal hemoglobin values, and three of the 1245 included subjects were compatible with high hemoglobin A2.
The means for HbA1c values ranged from 5.46% (36 mmol/mol) to 5.60% (38 mmol/mol) (Table 1); medians were between 5.27% (34 mmol/mol) and 5.42% (36 mmol/mol) for the different methods (Table 1 and Figure 1).
Table 1.
Parametric Statistics on the Four Methods for Measurement of HbA1c %.
| Measurement methods | Mean | Standard deviation | mean CI 95% | Percentile 25 | Median | Percentile 75 | Minimum | Maximum | P value (Bonferroni)* | |
|---|---|---|---|---|---|---|---|---|---|---|
| Lower lim. | Upper lim. | |||||||||
| HbA1c HPLC (a) | 5.54(37)** | 0.90(7) | 5.49(37) | 5.59(38) | 5.12(33) | 5.37(35) | 5.67(39) | 4.25(23) | 12.99(119) | <.001 vs b. c & d |
| HbA1c Electrophoresis (b) | 5.60(38) | 0.89(7) | 5.55(37) | 5.65(38) | 5.18(33) | 5.42(36) | 5.71(39) | 4.48(26) | 13.37(123) | <.001 vs a & c |
| HbA1c Enzymatic (c) | 5.46(36) | 0.89(7) | 5.41(36) | 5.50(37) | 5.05(32) | 5.27(34) | 5.55(37) | 4.36(24) | 13.25(121) | <.001 vs a. b & d |
| HbA1c Turbidimetric (d) | 5.60(36) | 0.90(7) | 5.55(37) | 5.65(38) | 5.16(33) | 5.4(36) | 5.7(39) | 4.35(24) | 13.45(124) | <.001 vs a & c |
Note. n = 1245. Abbreviation: HPLC, High Performance Liquid Chromatography.
ANOVA repeated measure (Bonferroni post hoc test).
The number between parentheses after the HbA1c value in % (NGSP) corresponds to the value in mmol/mol (IFCC), expressed without decimals.
Figure 1.
Box plots of values for each of the four methods to measure HbA1c.
The box-and-whisker plot shows the five statistics as they increase: lower limit for the evaluation of extreme cases, percentile 25, median, percentile 75, and upper limit for the evaluation of extreme cases.
*Outliers (1.5 interquartile range).
○Extreme outlier (3.0 interquartile range).
Comparison of the results obtained by applying the capillary electrophoresis and turbidimetric methods showed no statistically significant differences. The difference between the means obtained by HPLC and the turbidimetric and electrophoretic methods was 0.056% (0.38 mmol/mol) and 0.06% (0.42 mmol/mol) in each case, and 0.087% (1 mmol/mol) versus the enzymatic method (P = .001), statistically significant but clinically irrelevant (Table 1). The maximum difference between means registered was 0.14% (1 mmol/mol) (Table 1).
The parametric correlations between each of the methods showed R coefficients over 0.98; also, evaluation of agreement among the four described methods shows a coefficient for intraclass correlation of single measures of 0.982 (Table 2). These results are consistent with what the Passing and Bablock method reflects (Figure 2). This method makes no assumption about the distribution of samples or measurement errors.
Table 2.
Evaluation of Intraclass Agreement Among the Four Methods Studied.
| Agreement | Intraclass correlation (R) | IC 95% | P value | ||
|---|---|---|---|---|---|
| Lower limit | Upper limit | ||||
| 4 measurements | Single measures | 0.9823 | 0.9807 | 0.9838 | <.001 |
| Average measures | 0.9955 | 0.9951 | 0.9959 | <.001 | |
| HPLC vs Electrophoresis | Single measures | 0.9870 | 0.985 | 0.988 | <.001 |
| Average measures | 0.9930 | 0.993 | 0.994 | <.001 | |
| HPLC vs Enzymatic | Single measures | 0.9880 | 0.9861 | 0.9888 | <.001 |
| Average measures | 0.9940 | 0.9930 | 0.9944 | <.001 | |
| HPLC vs Turbidimetric | Single measures | 0.9860 | 0.9847 | 0.9878 | <.001 |
| Average measures | 0.9930 | 0.9923 | 0.9938 | <.001 | |
Abbreviation: HPLC, High Performance Liquid Chromatography.
Figure 2.
Passing and Bablock regression method for HbA1c assay by the three methods considering HPLC as reference: Capillary electrophoresis (a), Enzymatic (b), Turbidimetric (c).
Agreement between HPLC and each of the other methods, evaluated by applying the Bland and Altman technique was very good (Figure 3).
Figure 3.
Agreement calculated by Bland and Altman method for each method of determining HbA1c versus HPLC. Abbreviation: HPLC, High Performance Liquid Chromatography.
*Carla Lucarelli, Bahía Blanca, Buenos Aires; Sebastián Reyes, Comodoro Rivadavia, Chubut; Santiago Fares Taie, Mar del Plata, Buenos Aires; Héctor Luis Milani, Junín, Buenos Aires; Claudia Heim, Tandil, Buenos Aires; Gisela Gramajo, Trelew, Chubut; Nazareno Seren, Tandil, Buenos Aires; Carlos Furnari. Pergamino, Buenos Aires; Rafael Pérez Elizalde, Mendoza, Mendoza; Néstor Lejtman, Catamarca, Catamarca; Fabián Schurmann, Paraná, Entre Ríos; Alejandra V. Kossman, Cipolletti, Río Negro; Sergio Riesco, Gral Pico, La Pampa; Milva Sánchez. Martínez, Buenos Aires; Andrea Anabel Bianciotti, Santa Rosa, La Pampa; Maximiliano Croci, Colón, Entre Ríos; Alejandra Lombardo and Gustavo Dip, Rosario, Santa Fe; Alfredo Martinez, CABA, Buenos Aires; Natalia Carolina Lugo, Resistencia, Chaco; Alejandro Rapela, CABA, Buenos Aires; María Gabriela Simesen de Bielke, San Miguel de Tucumán, Tucumán; Patricia Castagnino, Avellaneda, Buenos Aires; Marcela Miró, Río Gallegos, Santa Cruz; Camila Garceron, La Rioja, La Rioja; Natalia Piaggio, Gualeguaychú, Entre Ríos; Matías Viniegra. San Justo, Buenos Aires; Norma Bálsamo. Bariloche, Río Negro; Andrés Albrecht, Rafaela, Santa Fe; Adrián Aymard, CABA, Buenos Aires; Guillermo J. Coronel, Formosa, Formosa; Liliana Bearzi, Puerto Madryn, Chubut; Juan Carlos Nicolás, Córdoba, Córdoba; German Andrés Correa, San Luis, San Luis; Luis Monaco, Quilmes, Buenos Aires; Gladys Ibañez, Salta, Salta; Susana Carminati, Trelew, Chubut; Carlos Chichizola. Santa Fe, Santa Fe.
Taking an HbA1c value of 6.5% (48 mmol/mol) as the cutoff point, validity is generally very good, except for the enzymatic method whose values are rather different from the other techniques when compared to the HPLC (Table 3). This is consistent with lower HbA1c means observed when applying this method.
Table 3.
Frequency Distribution, Percentage and 95% CI for HbA1c With a Cutoff Value of 6.5% (48 mmol/mol).
| Grouped HBA1c (cutoff 6.5%) | Frequency | % | IC 95.0% LI | IC 95.0% LS | |
|---|---|---|---|---|---|
| HbA1c HPLC | up to 6.5(48)** | 1,167 | 93.7% | 92.3% | 95.0% |
| over 6.5(48) | 78 | 6.3% | 5.0% | 7.7% | |
| HbA1c Electrophoresis | up to 6.5(48) | 1,165 | 93.6% | 92.1% | 94.8% |
| over 6.5(48) | 80 | 6.4% | 5.2% | 7.9% | |
| HbA1c Enzymatic | up to 6.5(48) | 1,176 | 94.5% | 93.1% | 95.6% |
| over 6.5(48) | 69 | 5.5% | 4.4% | 6.9% | |
| HbA1c Turbidimetric | up to 6.5(48) | 1,165 | 93.6% | 92.1% | 94.8% |
| over 6.5(48) | 80 | 6.4% | 5.2% | 7.9% | |
Abbreviation: HPLC, High Performance Liquid Chromatography.
The number between parentheses after the HbA1c value in % (NGSP) corresponds to the value in mmol/mol (IFCC), expressed without decimals.
Discussion
To guarantee the quality and clinical reliability of issued results, laboratories must comply with a program of analytic performance control (CV <2%) applied to the method selected to measure HbA1c, a program that must be certified by the NGSP 14 or the IFCC. 13
Results showed the methods under study are scarcely prone to variability, and also correlate highly, these conclusions were obtained considering the HPLC method as the reference one for the reasons explained at the Material and Methods section. All methods tested had a reasonably good analytic performance.
Although mean values obtained by capillary electrophoresis and turbidimetry deviated only minimally from the HPLC method (around 0.06%), these differences—albeit statistically significant—are irrelevant in clinical practice. Capillary electrophoresis mean value did not differ from data obtained by the turbidimetric method. The enzymatic method showed the highest deviation from the HPLC mean value, 0.09% (1 mmol/mol) and a 0.14% (1 mmol/mol) difference when compared with both capillary electrophoresis and turbidimetry.
The values near 1 of intraclass correlation clearly showed high agreement among the four methods; it is also quite likely that a patient with an HbA1c ≥6.5% (48 mmol/mol) has similar results when evaluated by any of the four methods.
Results show the methods are highly standardized and an excellent correlation and agreement among them (as seen in Figures 2 [a, b, and c] and 3 [a, b, and c] and Table 2).
When results are compared with the capillary electrophoresis and the turbidimetric methods in relation to the HPLC, we observed that 75 of 78 patients having ≥6.5% (48 mmol/mol) results by HPLC had values higher than ≥6.5% (48 mmol/mol). When compared with the enzymatic method, 69 of 78 results were ≥6.5% (48 mmol/mol).
Of the 1245 patients that took part in this study, only one had high fetal hemoglobin and three were compatible with high hemoglobin A2. Although it is possible that the existence of abnormal hemoglobin might have an influence on results of the HbA1c assay,15,16 these cases were not excluded from the population participating in this study, since their HbA1c results showed the same performance by the four methods.
Even though these results evidence a very high degree of consistency, they must be considered with caution given certain characteristics of this study such as there was a careful and standardized procedure to obtain the blood samples, which were then processed simultaneously in one single laboratory and with just one operator—conditions unlikely to be reproduced in a typical laboratory working day.
Anyhow, we believe our results are objective enough to guarantee their credibility and reproducibility by these different methods available on the market. In this sense, clinical laboratories are hugely responsible when selecting and using a method for quantification of HbA1c, considering the wide range of methods offered by manufacturers of diagnostic tests. What should be highlighted is that, besides the selection of a method certified by the NGSP or the IFCC, compliance with a strict internal quality control program and participation in an external evaluation program are necessary in order to guarantee a <2% Interlaboratory Coefficient of Variation.
Due to the significant differences in the results obtained by different methods present on the market in Argentina—-many not studied in this work—and their lack of standardization, the Argentine Diabetes Society, together with the Uruguayan Society of Diabetes and Nutrition, do not recommend today the use of HbA1c as a diagnostic method.
Even for the follow-up of patients already diagnosed and treated, the use of suboptimal methods for the determination of HbA1c is a problem in clinical practice in these countries.
In order to review the issue and within the framework of a cooperation agreement between the Argentine Diabetes Society and the Association of High Complexity Laboratories, different actions were initiated, this work being one of them; we hope that the results obtained will help sustain the necessary changes.
Acknowledgments
Estela Zanuso de Viniegra, Viniegra – Zanuso VZ Laboratorios, Buenos Aires; Laboratorios BAIRESLAB, Buenos Aires; Valeria Mohr, Tres Arroyos, Buenos Aires; Marcelo Pugliessi, Rosario, Santa Fe; Víctor Pessacq, La Plata, Buenos Aires; Reinaldo Marcomini, Corrientes, Corrientes; Fernando Elías, Rosario, Santa Fé; Carlos Insaurralde, Posadas, Misiones; Esteban Benelbaz, San Juan, San Juan; Carlos Zocchi, Dra. Jesica Barrionuevo. Neuquén, Neuquén; Laboratorio Hidalgo, Buenos Aires, Argentina; Roche Diagnostics Argentina; SEBIA - BG Analizadores; Abbott Laboratories Argentina; Biogiagnóstico S.A.; FEFyM (Fundación de Estudios Farmacológicos y Medicamentos). We thank all the participants for their contributions to this study.
Footnotes
Abbreviations: ADA, American Diabetes Association; DCCT, Diabetes Control and Complications Trial; DM, Diabetes Mellitus; EASD, European Association for the Study of Diabetes; HPLC, High Performance Liquid Chromatography; IDF, International Diabetes Federation; IFCC, International Federation of Clinical Chemistry; NGSP, National Glycohemoglobin Standardization Program; OGTT, Oral Glucose Tolerance Test; RMP, reference measurement procedure; WHO, World Health Organization.
Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by Association of High Complexity Laboratories—ALAC.
ORCID iDs: María Zulema Chaila 
https://orcid.org/0000-0001-8577-8417
Víctor Francisco Commendatore
https://orcid.org/0000-0002-1757-5690
References
- 1. Fox CS, Coady S, Sorlie PD, et al. Increasing cardiovascular disease burden due to diabetes mellitus: the Framingham Heart Study. Circulation. 2007;115:1544-1550. [DOI] [PubMed] [Google Scholar]
- 2. Nguyen NT, Nguyen XM, Lane J, Wang P. Relationship between obesity and diabetes in a US adult population: findings from the National Health and Nutrition Examination Survey, 1999-2006. Obes Surg. 2011;21:351-355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Herman WH, Ye W, Griffin SJ, et al. Early detection and treatment of type 2 diabetes reduce cardiovascular morbidity and mortality: a simulation of the results of the Anglo-Danish-Dutch study of intensive treatment in people with screen-detected diabetes in primary care (ADDITION-Europe). Diabetes Care. 2015;38:1449-1455. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. American Diabetes Association. 2. Classification and diagnosis of diabetes: standards of medical care in diabetes-2019. Diabetes Care. 2019;42(Suppl 1):S13-S28. [DOI] [PubMed] [Google Scholar]
- 5. Position Statement. Standards of medical care in diabetes—2010. Diabetes Care. 2010;33(Suppl 1):S11-S61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. International Expert Committee. International Expert Committee report on the role of the A1C assay in the diagnosis of diabetes. Diabetes Care. 2009;32:1327-1334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. World Health Organization. Use of glycated haemoglobin (HbA1c) in the diagnosis of diabetes mellitus. Abbreviated Report of a WHO Consultation. WHO/NMH/CHP/CPM [Internet] 2011;11:1-25. Accessed February 23, 2020. https://apps.who.int/iris/bitstream/handle/10665/70523/WHO_NMH_CHP_CPM_11.1_eng.pdf?sequence=1&isAllowed=y [PubMed] [Google Scholar]
- 8. Kramer CK, Araneta MR, Barrett-Connor E. A1C and diabetes diagnosis: the Rancho Bernardo Study. Diabetes Care. 2010;33:101-103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Marini MA, Succurro E, Arturi F, et al. Comparison of A1C, fasting and 2-h post-load plasma glucose criteria to diagnose diabetes in Italian Caucasians. Nutr Metab Cardiovasc Dis. 2012;22:561-566. [DOI] [PubMed] [Google Scholar]
- 10. Little RR, Rohlfing CL, Sacks DB; National Glycohemoglobin Standardization Program (NGSP) Steering Committee. Status of hemoglobin A1c measurement and goals for improvement: from chaos to order for improving diabetes care. Clin Chem. 2011;57:205-214. [DOI] [PubMed] [Google Scholar]
- 11. Diabetes Control Complications Trial Research Group; Nathan DM, Genuth S, Lachin J, et al. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329:977-986. [DOI] [PubMed] [Google Scholar]
- 12. Weykamp C, John G, Gillery P, et al. ; IFCC Task Force on Implementation of HbA1c Standardization. Investigation of 2 models to set and evaluate quality targets for HbA1c: biological variation and sigma-metrics. Clin Chem. 2015;61:752-759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Sacks DB, Arnold M, Bakris GL, et al. ; National Academy of Clinical Biochemistry; Evidence-Based Laboratory Medicine Committee of the American Association for Clinical Chemistry. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Diabetes Care. 2011;34:e61-99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Little RR, Rohlfing C, Sacks DB. The national glycohemoglobin standardization program: over 20 Years of Improving Hemoglobin A1c Measurement. Clin Chem. 2019;65:839-848. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Little RR, Roberts WL. A review of variant hemoglobins interfering with hemoglobin A1c measurement. J Diabetes Sci Technol. 2009;3:446-451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Nitta T, Yamashiro Y, Hattori Y, Ezumi T, Nishioka M, Nakamura J. The interference by HbF on HbA1c (BM Test HbA1c) measurement in enzymatic method. Ann Clin Biochem. 2015;52:569-575. [DOI] [PubMed] [Google Scholar]