Evaluation of an Enzymatic Hemoglobin A1C Assay

1 2 3
Numan Majeed, Fatima Kanani, Adnan Mustafa Zubairi
1-3
Indus Hospital and Health Network, Karachi, Pakistan

Abstract
Background: Different modalities have been designed for estimation of HbA1c, that include ion-exchange
High Performance Liquid Chromatography (HPLC), immunoassay, enzymatic assays and capillary
electrophoresis. Although based on different principles, these analytical processes have to exhibit robust
performance so that an overall comparability of assay values can be ensured which are traceable to the
available reference method.
Objectives: Objective of the study was to evaluate HbA1c measurement on Abbott Alinity C system against
ion exchange high performance liquid chromatography.
Methods: We performed different experiments to evaluate measurement of HbA1c on Alinity C system by
Abbott. These studies included accuracy, precision, linearity, carry over and method comparison in patients
without any hemoglobin variants. Evaluation Protocol (EP) evaluator was used for data analysis.
Results: All the experiments passed the minimum passing criteria. All the claims by the vendor were successfully
verified with and addition that vendor claims 12-hour stability while our study showed a stability of 18 hours.
In precision the coefficient of variance in our experiment showed better results.
Conclusion: Alinity C system can be used reliably and efficiently for the assessment of HbA1c in patients
without haemoglobin variants as our study did not include samples with variant hemoglobin.
Received: 14-11-2023 | Revision: 20-04-2024 | Accepted: 05-06-2024
Corresponding Author | Dr. Numan Majeed, Resident, Indus Hospital and Health Network, Karachi, Pakistan
Email: numan.mjd@gmail.com
Keywords: Glycated hemoglobin, Alinity C, HPLC, method validation

Introduction are remote and difficult to access.3 Accurate and prompt

diagnostic testing, which is often done by hospital clini-
D iabetes mellitus is one of the fastest growing non
communicable epidemic, affecting millions of
people globally, which affects quality of life by its many
cal laboratories using a variety of techniques, is necessary
for the diagnosis and efficient treatment of diabetic
4, 5
1 patients.
complications such as nephropathy, neuropathy, retino-
pathy.2 According to an estimate about 50% cases of Glycated hemoglobin (HbA1c) is an important marker
diabetes are yet undiagnosed, particularly in areas that of mean glycemic index of diabetic patients that can be
used for monitoring purposes and is a strong predictive
Production and Hosting by KEMU marker for complications of diabetes.5,6 HbA1c is also
https://doi.org/10.21649/akemu.v30i2.5567
2079-7192/© 2024 The Author(s). Published by Annals of
included in the diagnosis of diabetes by the American
7
KEMU on behalf of King Edward Medical University Lahore,
Pakistan.
Diabetes Association. Several methods have been
This is an open access article under the CC BY4.0 license
http://creativecommons.org/licenses/by/4.0/
developed for quantification of HbA1c including ion-

April - June 2024 | Volume 30 | Issue 02 | Page 184

exchange High Performance Liquid Chromatography being performed by enzymatic method. The reagent
(HPLC), immunoassay, enzymatic assays and capillary is provided as ready for use cassettes with an on-board
electrophoresis, cation exchange high performance stability of 30 days.
8
liquid chromatography. Some kits/assays designed Various experiments were carried out for the purpose
to estimate HbA1c levels may report false results due of evaluation, which are summarized below
to presence of hemoglobin variants such as HbS, HbE,
HbF and HbD. These interferences may result in clini- 1. Clinical & Laboratory Standards Institute provided
(12)
cally significant variations resulting in either over treat- Evaluation Protocol no. 15-A3 (CLSI EP15-A3)
ment and hypoglycemia or under-treatment leading to was followed for precision. Intra and inter day
hyperglycemia, both of which are potentially detri- precision were assessed by analyzing three patient
mental to human health.4 pools (with values of 5.62%, 10.2% and 13.9%)
for a period of 5 days with series of five replicates
To identify the fraction of hemoglobin A that is glycated each day.
4

at one or both of the N-terminal valines of the beta-chain,

Abbott Diagnostics has developed an automated whole 2. Accuracy was analyzed using proficiency Testing
blood Hemoglobin A1c (HbA1c) test on the Alinity c Surveys (GH5I-A 2022) provided by College of
system. The N-terminal fructosyl dipeptides of the beta- American Pathologists at various concentrations.
chain of HbA1c are specifically measured using an Whole blood received (in EDTA containing,
enzyme approach that automatically lyses red blood vacuum tubes, by Becton, Dickinson Vacutainer)
cells. Glycated hemoglobin (HbA1c) and total hemo- were analyzed in the laboratory for HbA1c on the
globin (THb)are the two distinct measures, that are current method (HPLC) on Bio-Rad Turbo Variant
utilized to calculate the percent HbA1c and (National II and Alinity c for HbA1c samples without haemo-
Glycohemoglobin Standardization Program [NGSP] globin variants across the analytical measuring
units(mmol/mol).9 range. The samples were taken from hematology
Although based on different principles, these analytical department that came for Hb variant studies and
processes have to exhibit robust performance so that an were found not having any variant.
overall comparability of assay values can be ensured 3. Two experiments were conducted for method
10
which are traceable to the available reference method. comparison:
The International Federation of Clinical Chemistry and a. 80 samples with HbA1c levels within analytical
Laboratory Medicine (IFCC) Task Force on Implemen- measurement range and Hemoglobin levels
tation of HbA1c standardization (TF-HbA1c) has advo- between 12 and 16 mg/dL were analyzed by
cated the use of sigma metrics for the purpose of evalua- both methods. Specimens were categorized
tion and setting the quality targets of glycated HbA1c.11 In on HbA1c levels as follows:
13-15

the laboratory this metrics is a quality indicator that • 4.0 -5.7%

provides a benchmark for process performances. It • >5.7, <7.1%
makes sure that the analytical characteristics such as • 7.1-<10%
precision and bias are within the total allowable error.
4 • 10.0-14%
The assay is certified by NGSP, standardized to Inter- b. 20 specimens with hemoglobin levels below
national Federation of Clinical Chemistry (IFCC), and 7g/dL and not demonstrating any variant hemo-
traceable to the Diabetes Control and Complications globin were used to compare the methods at
Trial (DCCT). very low levels of Hemoglobin.
16
4. Linearity was done following CLSI EP06Ed2,
Methods by measuring five HbA1c levels in duplicate cove-
Alinity C system is a standalone as well as an integrated ring the entire AMR and using CAP linearity mate-
platform in total lab automation systems. It is based rial (Hemoglobin A1c Calibration Verification/
on photometric and turbidimetric methods. HbA1c is Linearity LN15-B 2020).

April - June 2024 | Volume 30 | Issue 02 | Page 185

5. Carryover was assessed according to CLSI EP10- • At 13.90%
A3 (17) using samples with low, mid, and high o Between run CV was 0.1, while that claimed
concentrations of HbA1c (4.9, 7.0, and 9.6%, by Abbott is 0.3.
respectively) run in the order M-H-L-M-M-L-L- o Between day CV was 0.0, no claim made by
H-H-M, where L is low concentration, M is mid- Abbott
level and H is high level
We compared Alinity C system with Bio-Rad Variant
6. CLSI EP35 was used to perform stability experi- II for normal (non-anemic) and anemic samples. All
ment. Samples of 5 patients were taken. Each the samples were tested for variants and only those were
sample was run in duplicate at six-time intervals selected which had no haemoglobin variants. All the
without mixing according to following grid: experiments intercept and slope both are in allowable
- (time from draw) 0 hours, 1 hour, 4-hour, 8-hour, limits hence experiment was considered pass.
12 hours, 24 hours). The samples were stored at Table 2 shows comparison of proficiency testing samples
room temperature (24-26OC). run on the instrument with the assigned target values.
Total allowable error (TEa) was kept at 6%, which has All the samples were within the allowable limits hence
been set by College of American Pathology (CAP) the experiment was declared pass.
EP Evaluator was used for data analysis. Statistical The prescribed linearity of the instrument is 12.8. All
analysis/ experiment passing criteria as mentioned in
respective CLSI guidelines is shown in Table 1. Table 2: Results of accuracy experiment
Targe t Measured Values
Results Sample No
Run 1 Run 2 Range Midpoint Accuracy
The precision results passed at all three levels. The 1 (GH51-01) 7.1 7.1 6.4 to 7.7 7.1 Pass
coefficient of variance (CV) at all the three levels was 2 (GH51-02) 5.6 5.6 5.2 to 6.1 5.6 Pass
found to be <2%. 3 (GH51-03) 6.1 6 5.5 to 6.6 6 Pass
4 (GH51-04) 12.1 12.1 11 to 12.6 11.8 Pass
• At 5.62%
5 (GH51-05) 6.4 6.5 6 to 7 6.5 Pass
o Between run CV was 0.7 which was exactly
the same as that claimed by Abbott. the samples (CAP linearity material) were within the
o Between day CV was 0.0, no claim made by specified range. Slope was found to be 1.012 (0.996 to
Abbott 1.027) while intercept to be 0.10 (-0.23 to 0.02), which
• At 10.20% fall within the allowable limits declaring experiment
o Between run CV was 0.1, while that claimed as pass. Table 4 shows results of the linearity study.
by Abbott is 0.4.
o Between day CV was 0.5, no claim made by
Abbott
Table 1: Passing criteria for each experiment
No. Experiment Criteria Range of acceptability
1. Linearity Non-linearity: <2% TEa: 6%,
% of nonlinearity: 25%
2. Accuracy Midpoints of the target ranges for the lowest and highest Reportable range Proximity limits
specimens respectively are within proximity limits of the Low 4 3.6 to 4.4 %
Reportable Range Limits, and 2) these two specimens
also pass accuracy High 14 12.6 to 15.4 %
3. Precision SD Goal is within the 95% confidence limit Manufacturer claim, given in package insert.
4. Carry over Carryover is less than the Error Limit (0.1%) Error limit 6%
5. Method The slope is 1.00 (within 95% confidence) TEa:6%
comparison The intercept is 0.00 (within 95% confidence)

April - June 2024 | Volume 30 | Issue 02 | Page 186

 The linearity experiment showed liner results (pass)
in table 4

Figure 1: Scatter plot and recovery chart of the accuracy

study
Method comparison was done at two levels, i.e. hemo-
globin levels with in normal reference range and in Figure 2: Scatter plot and residual plot for the Linearity
anemic range study
Table 3 shows regression (both Deming and regular) The carry over experiment was declared pass with a
for the both anemic and nonanemic samples total carryover effect of -0.2%.
Table 3: Regression analysis for anemic and non-anemic samples
Non-Anemic Anemic
Deming Regular Deming Regular
Slope 1.030 (1.017 to 1.043) 1.026 (1.013 to 1.040) 0.979 (0.954 to 1.004) 0.976 (0.952 to 1.001)
Intercept -0.35 (-0.46 to -0.25) -0.32 (-0.43 to -0.22) 0.10 (-0.03 to 0.23) 0.12 (-0.01 to 0.25)
Std Err Est 0.20 0.20 0.07 0.07

Table 4: Linearity study results

Sample Assigned value Mean Poly Fit Line Fit Deviation from Linearity* Deviation**
1 (LN15-07) 5.4 5.30 5.36 5.36 0.00 0.00
2 (LN15-08) 6.69 6.90 6.74 6.74 0.00 0.00
3 (LN15-09) 7.97 8.00 8.10 8.10 0.00 0.00
4 (LN15-10) 9.26 9.45 9.47 9.47 0.00 0.00
5 (LN15-11) 10.55 10.85 10.85 10.85 0.00 0.00
6 (LN15-12) 11.84 12.25 12.22 12.22 0.00 0.00
*: Deviation from linearity: Deviation from linearity curve **: Deviation: Deviation from assigned value

April - June 2024 | Volume 30 | Issue 02 | Page 187

All the samples were run at specified intervals without HbA1c in order to ensure the quality and clinical relia-
mixing/shaking and the stability was found out to be bility of results produced. This program needs to be
21
13.2 hours, which is more than the stability claimed accredited by the NGSP or the IFCC.
by the manufacturer. (figure 3) Our study demonstrated that maximum imprecision
(CV=0.7%), the imprecision was higher at low concen-
tration but improved as the value of HbA1c got higher
till its linearity (13%).
The TOP-HOLE study is the one of the very few studies
so far that performed validation of Alinity C system.
But unlike our study in which we compared HPLC using
Bir-Rad Variant II Turbo, the TOP-HOLE study used
enzymatic Architect C-4000 with Alinity c stand-alone
(S-A) system. Both of the instruments use same method
and reagents, yet demonstrated a bias of -3.1at low level
(31.9 mmol/mol equaling to 5.1%) and 4.59 higher level
Figure 3: Stability study cut-point (87.2 mmol/mol equaling to 10.1% using control mate-
rial.19 But as both the platforms had same reagent and
Discussion: method, we cannot suggest which one is more reliable.
Another study reported a total imprecision at different
Estimation of HbA1c is an essential investigation for the
concentration levels ranging from 0.5-1.3%.9
diagnosis as well as monitoring of diabetes mellitus as
per international guidelines including American Diabetes Most of the studies conducted on enzymatic assay for
Association (ADA) and National Institute for Health estimation of HbA1c have used used other instruments
18
and Care Excellence (NICE ). For the estimation of as Alinity C system is comparatively newer and very few
HbA1c levels, different methods and instruments have studies so far has been done. A study done by Monaco
been developed. IFCC and NGSP has set specific analy- et al on the Abbott Architect system, a system that utilizes
tical pre-requisites for and instrument to be deemed fit same reagent has shown very similar results as ours,
for clinically acceptable results of HbA1c.18 Moreover, our results showed a better agreement as compared to
in the diagnosis of DM, HbA1c testing offers significant that shown in this study.22 Similarly, another study by
advantages over fasting plasma glucose. HbA1c is stable Sriwimol et al done on same principal, using Mindray,
23
in blood samples, does not require fasting, has less vari- similar results were obtained. Chalia et al also showed
ability within individuals, and is not affected by acute similar findings, they compared three systems, enzy-
situations. Due to its critical clinical role, there is a global matic on Abbott Architect, Roche turbidimetric and
21
consensus that HbA1c results should be standardized. To HPLC showing strong correlation with all the methods.
achieve this, the IFCC has defined an internationally On Siemens Dimension (also an enzymatic assay) simi-
recognized HbA1c reference measurement system lar results were reported but the study also suggested
24
(RMS), ensuring the global interchangeability of result that Dimension is not as cost effective as Biorad.
interpretation, independent of the assay used. Manufac- Similarly, our study showed that the samples are stable
turers of in vitro diagnostic systems are required to for a period of 13.2 hour at room temperature and can
design commercial systems that meet the traceability be analyzed even without mixing and stirring. It is
19
requirements to this RMS. Additionally, laboratory worthwhile to mention that Abbott in their kit literature
professionals must verify the alignment process to mentions 8 hours of stability and recommends mixing
higher-order references and ensure that the marketed of all samples before analysis. Although we did not use
measuring systems' performance is appropriate for clini- the total lab automation TLA) track system but it can
cal use.20 decrease the turnaround time (TAT), as reported in the
Labs must adhere to a program of analytic performance TOP-HOLE study to 1 hour only.
control (CV <2%) applied to the method used to measure
April - June 2024 | Volume 30 | Issue 02 | Page 188
As Alinity system by Abbott is comparatively new, so References
as of now on Alinity systems very few studies have been 1. Teodoro-Morrison T, Janssen MJ, Mols J, Hendrickx
done for validation of HbA1c enzymatic assay. Most of BH, Velmans MH, Lotz J, et al. Evaluation of a next
the studies done were by comparing Alinity with other generation direct whole blood enzymatic assay for
enzymatic or immunoturbidimetric methods only one hemoglobin A1c on the ARCHITECT c8000 chemistry
study reported comparison with HPLC, but it did not system. Clin Chem Lab Med. 2015;53(1):125-32. doi:
10.1515/cclm-2014-0310. PubMed PMID: 25153395.
fulfil the CLSI recommended experiments criteria
2. Butt N, Bano F, Ghani M, Ahmed AM, Majeed N. Asso-
(Berman et al).
ciation of serum advanced glycation (AGEs) end pro-
The major limitation of the study was that the samples ducts, apolipoprotein-B and zinc in severity of T2DM
were not run on TLA track system and TAT was not retinopathy.PJPS. 2021;34(2):803-8.
calculated. Secondly effect of various haemoglobin 3. Majeed N, Ahmad TM, Hayat A, Israr S, Saddique A,
variants on HbA1c could not be verified because of lack Shabbir S. Role of Serum Lipid and Apolipoprotein
of availability of such samples in sufficient numbers Ratios in Prediction of Diabetic Retinopathy. Biosight.
at various percentages. 2023;4(1):3-10. doi: https://doi.org/10.46568/bios. v4i1.
85.
Our study further opens the doors for further studies 4. Lenters-Westra E, English. Evaluating new HbA1c
using more robust study on impact of variants as well methods for adoption by the IFCC and NGSP reference
as track-based time effectiveness and most of all cost networks using international quality targets. CCLM.
calculations. 2017;55(9):1426-34.
5. Gillery P. HbA1c and biomarkers of diabetes mellitus
Conclusion in Clinical Chemistry and Laboratory Medicine: ten
years after. CCLM. 2023;61(5):861-72.
Abbot Alinity C system which is based on enzymatic
6. Rohlfing C, Hanson S, Weykamp C, Siebelder C, Higgins
estimation of HbA1c has passed the five experiments
T, Molinaro R, et al. Effects of hemoglobin C, D, E and
recommended by Clinical & Laboratory Standards
S traits on measurements of hemoglobin A1c by twelve
Institute that include linearity, accuracy, precession, methods. 2016;455(1):80-3.
method comparison (both with normal and low level 7. Association AD. Addendum. 2 Classification and diag-
hemoglobin) and carry over experiments following their nosis of diabetes: standards of medical care in diabetes
recommended protocols, hence it can be used reliably — 2021. Diabetes care. 2021; 44(Supplement_1): S15-
for estimation of samples without any hemoglobin S33.
variant. 8. Jaisson S, Desmons A, Renard B, Chevelle B, Leroy
Ethical Approval: The Institutional Review Board of N, Gillery PJCCA. Analytical performances of a new
enzymatic assay for hemoglobin A1c. 2014; 434 (1):
Indus Hospital & Health Network, Karachi approved
48-52.
the study vide No. IHHN-IRB# IHHN_IRB_2022_
9. Berman M, Templin L, editors. Performance of the
09_027. Enzymatic Alinity c Hemoglobin A1c Assay. 71st AACC
Conflict of Interest: The authors declare no conflict Annual Scientific Meeting Abstracts; 2019: AACC.
of interest. 10. Chakraborty S, Chakraborty S, Ghosh A, Chakraborty
SS. A Comparative Evaluation of Estimation of Glycated
Funding Source: None Haemoglobin by Immunoturbidimetry and High Perfor-
Authors' Contribution: mance Liquid Chromatography in a Tertiary Care Hos-
pital of Kolkata. IJRR. 2019;6(3):155-158.
NM: Conception & design, analysis & interpretation
11. Weykamp C, John G, Gillery P, English E, Ji L, Lenters-
of data, drafting of article
Westra E, et al. Investigation of 2 models to set and eva-
FK: Conception & design, critical revision for impor- luate quality targets for HbA1c: biological variation
tant intellectual content, final approval and sigma-metrics. Clinical chemistry. 2015; 61(5): 752-
9.
AMZ: Conception & design, final approval
12. Stokes C, Frost S, Frost L, Mcinerney N, O’Dwyer S.
A-070 A fully automated HbA1c Assay on the DxC 500

April - June 2024 | Volume 30 | Issue 02 | Page 189

 AU Clinical Chemistry Analyser. Clinical Chemistry. as a first-line laboratory test: the TOP-HOLE (Towards
2023;69(Supplement_1):hvad097. 67. OPtimal glycoHemOgLobin tEsting) project. CCLM.
13. Clinical LSI. EP09‐A3: Measurement Procedure Com- 2022;60(3):441-50.
parison and Bias Estimation Using Patient Samples. 20. Braga F, Panteghini M. Verification of in vitro medical
2013. diagnostics (IVD) metrological traceability: responsi-
14. Petrides AK, Dahlin JL, Melanson SE. Method validation. bilities and strategies. Clinica Chimica Acta. 2014;
Contemporary Practice in Clinical Chemistry: Elsevier; 432(1):55-61.
2020. p. 57-75. 21. Chaila MZ, Viniegra M, Gagliardino JJ, Martínez A,
15. Cihan M, Abdullah Ü, KILINÇKAYA MF, Noyan T. Simesen de Bielke MG, Frusti M, et al. Glycated Hemo-
Comparison of high performance liquid chromatography globin Measurement: Comparison of Three Methods
and turbidimetric inhibition immunoassay methods Versus High Performance Liquid Chromatography.
for measurement of hemoglobin A1c. JHSM. 2022; JDST . 2022;16(3):724-31. doi: 10.1177/ 193229682
5(5): 1281-5. 1997179. PubMed PMID: 33686874.
16. Lee JH, Kim S, Jun S-H, Seo JD, Nam Y, Song SH, et 22. Luis Monaco B, Pablo Salgado B, Carlos Buso M,
al. Analytical performance evaluation of the Norudia Claudio Daniel Gonzalez M. Glycated Hemoglobin
HbA1c assay. J. Clin. Lab. Anal . 2020; 34(11): e23504. Measurement: Comparison of Three Methods Versus
doi: https://doi.org/10.1002/jcla.23504. High Performance Liquid Chromatography. JDST.
17. Núñez-Jurado D, Montenegro-Martínez J, Rubio- 2021;77(1):1-8.
Sánchez R, Conde-Sánchez M, Domínguez-Pascual 23. Sriwimol W, Choosongsang P, Choosongsang P, Tree-
I. Evaluation of the Premier Hb9210 instrument for rut P, Muenniam B, Makkong P, Pocathikorn A. Strong
HbA1c determination. Scandinavian. J. Clin. Lab. 2023; correlation and high comparability of capillary electro-
83(8):569-75. phoresis and three different methods for HbA1c measure-
18. Ohde S, Deshpande GA, Yokomichi H, Takahashi O, ment in a population without hemoglobinopathy. Scan-
Fukui T, Yamagata Z. HbA1c monitoring interval in dinavian. J. Clin. Lab. 2020;80(2):139-50. doi: 10. 1080/
patients on treatment for stable type 2 diabetes. A ten- 00365513.2019.1703213.
year retrospective, open cohort study. Diabetes research 24. Mukherjee B, Hooda C, Singh V, Sahu R, Singh MU.
clinical practice. 2018;135(1):166-71. Comparison of NGSP Certified Methods and their Cost
19. Pasqualetti S, Carnevale A, Dolci A, Panteghini M. A Analysis for Estimation of HbA1c-A Cross-sectional
step towards optimal efficiency of HbA1c measurement Study. JCDR . 2023;17(2):BC01-BC4.

April - June 2024 | Volume 30 | Issue 02 | Page 190