Content uploaded by Ludek Slavik
Author content
All content in this area was uploaded by Ludek Slavik on Aug 22, 2016
Content may be subject to copyright.
Clin. Lab. 7/2016 1
Clin. Lab. 2016;62:XXX-XXX
©Copyright
ORIGINAL ARTICLE
Validation of a New Panel of Automated Chemiluminescence Assays
for Anticardiolipin Antibodies in the Screening for
Antiphospholipid Syndrome
D. Janek 1, L. Slavik 2, J. Ulehlova 2, V. Krcova 2, A. Hlusi 2, J. Prochazkova 2
1 Department of Hemato-Oncology, Faculty of Medicine and Dentistry, Palacky University Olomouc and Department of Hematology and Transfusion
Medicine, Hospital Karvina, Czech Republic
2 Department of Hemato-Oncology, Faculty of Medicine and Dentistry, Palacky University Olomouc and University Hospital Olomouc,
Czech Republic
SUMMARY
Background: Antibodies Anticardiolipin (aCL) and anti-β2-glycoprotein I (aβ2GPI) are two of three laboratory
criteria of antiphospholipid syndrome (APS). All of assays of antiphospholipid antibodies (aPL), coagulation as-
says as well as ELISAs, show methodological shortcomings, that affect their sensitivity and specificity. Therefore,
we decided to validate these parameters for a new chemiluminescent examination (CLIA).
Methods: aCL and aβ2GPI antibodies were measured by ELISAs (AIDA, Bad Kruznach, Germany) and aβ2GPI
with CLIA kits (Werfen, Barcelona, Spain).
Results: When we evaluated both assays, the coefficient of variation for CLIA was slightly lower (9.04 - 12.74%)
than for ELISA (11.05 – 15.3%) and the LOD was 0.2 IU/L. The dilution series showed significant linearity for all
CLIA methods, aCL IgG, aCL IgM, aβ2GPI IgG, and aβ2GPI IgM (0 - 3000 IU/L), and method comparison stud-
ies revealed good agreement with the currently used ELISA (Kappa values ranging 0.534 - 0.936) without deter-
mination of aβ2GPI IgG. The determination aβ2GPI IgG by CLIA method shows higher positivity in 31 samples.
These new aCL IgG, aCL IgM, aβ2GPI IgG, and aβ2GPI IgM tests have excellent analytical characteristics and
allow fully automated and simultaneous analysis on an analyzer.
Conclusions: Chemiluminescent determination of an automated analyzer can improve the fundamental parame-
ters of tests such as reproducibility between laboratories.
(Clin. Lab. 2016;62:xx-xx. DOI: 10.7754/Clin.Lab.2015.151129)
Correspondence:
Mgr. Ludek Slavik, Ph.D.
Department of Hemato-Oncology
Faculty of Medicine and Dentistry
Palacky University Olomouc and
University Hospital Olomouc
I. P. Pavlova 6
77520 Olomouc
Czech Republic
Email: ludek.slavik@fnol.cz
_____________________________________________
Manuscript accepted November 27 2015
KEY WORDS
Antiphospholipid Syndrome, Anticardiolipin antibodies,
anti- β2-glycoprotein I antibodies, chemiluminescence
immunoassay
INTRODUCTION
Antiphospholipid antibodies (aPLs) represent a hetero-
geneous group of antibodies that recognize phospholip-
ids (PL), PL-binding proteins or PL-protein complexes.
There is strong evidence that aPLs are pathogenic in vi-
vo, leading to a large variety of clinical manifestations
among which vascular thrombosis and recurrent fetal
loss are the most prevalent. The presence of these clini-
D. Janek et al.
Clin. Lab. 7/2016
2
cal events associated with the detection of aPLs in the
blood characterizes the antiphospholipid syndrome
(APS). Criteria for classification of APS were first de-
fined in Sapporo in 1999 [1] and were later revised in
Sydney in 2006 [2]. According to this last consensus
statement, laboratory tests that are used to define the
laboratory criteria of APS are the lupus anticoagulant,
anticardiolipin antibodies (aCL), and anti-β2-glycopro-
tein I antibodies (aβ2GPI). All three types of tests are
considered to be independent risk factors; therefore, the
positivity in any of them is sufficient to consider it as a
laboratory criterion of APS.
aCL were first detected by using a radioimmunoassay
(RIA) in 1983 [3]. Beginning in 1985, the detection of
aCL was replaced by enzyme-linked immuno-sorbent
assays (ELISA). Until the early 1990s, aCL testing was
mainly performed using in-house ELISAs. Once com-
mercial kits became available, routine laboratories start-
ed to use them with a high inter-laboratory variability of
the results.
Despite considerable efforts, according to external qual-
ity survey results, standardization of aCL testing is far
from being achieved [4,5]. Nowadays, incipient tech-
nologies attempt to develop fully automated methods
for the detection of anticardiolipin antibodies.
The aim of this study is to compare a fully automated
chemiluminescence immunoassay (CLIA) to an ELISA
assay for the detection of both isotypes (IgG, IgM) of
aCL and aβ2GPI antibodies.
MATERIALS AND METHODS
The group of patients
This study was conducted on a set of blood samples
from 122 patients with suspected APS positivity, sent to
our laboratory between January and October 2014. Cit-
rate plasma samples were immediately analyzed or
stored at -80°C for CLIA and ELISA.
Blood collection
Blood sampling was carried out in a single vacuum tube
using a Vacuette® needle (Greiner Bio-One, Vienna,
Austria), with a buffered solution containing sodium
citrate at a concentration of 0.109 mol/L (3.2%). The
system ensured blood and anticoagulant mixture at a de-
sired 1:10 ratio. Then the blood was carefully mixed in
a test tube, with the tube being gently turned upside
down several times and transported to the laboratory.
Then the sample was centrifuged two times for 10 min-
utes at 3000 x g, 0.5 mL of the upper layer of platelet-
poor plasma (PPP) was aspirated, then frozen and stored
at -80°C until CLIA and ELISA was performed. For the
actual analysis, the sample was thawed in a water bath
at 37°C for 20 minutes.
Autoantibody assays
aCL and aβ2GPI antibodies were measured by ELISAs
(AIDA, Bad Kruznach, Germany) and aβ2GPI with
CLIA kits (Werfen, Barcelona, Spain) - the assay is cur-
rently in use in our lab. The results are expressed in
U/mL [9]. The cutoff value of the AIDA ELISAs was
locally determined as the 99th percentile of 50 healthy
volunteers. The ELISA method was performed with a
reader Multiscan FC (Thermo-LabSystems, Helsinki,
Finland).
The CLIA method was performed with Acustar (Wer-
fen, Barcelona, Spain), a random-access immunoana-
lyzer, using a two-step immunoassay method based on
the principle of chemiluminescence. β2GPI or cardiolip-
in/β2GPI complex is used to coat magnetic particles and
a human anti-IgG or anti-IgM is labeled with conjugate.
During the first incubation the specific antibodies pres-
ented in the sample, in the calibrators, or in the controls,
bind with the solid phase. During the second incubation,
the conjugate reacts with the antibodies captured on the
solid phase. After each incubation, the material that has
not bound with the solid phase is removed by suction
and repeated washing [6,7].
The quantity of marked conjugate bound to the solid
phase is evaluated by chemiluminescent reaction and
measured by the light signal. The generated signal, mea-
sured in RLU (Relative Light Units), is indicative of the
concentration of the specific antibodies present in the
sample. For aCL IgG or IgM, the concentrations of the
calibrators are expressed in U/mL (U = units) and cali-
brated against the “Harris” reference sera. For aβ2GPI
IgG or IgM, the concentrations of the calibrators are ex-
pressed in U/mL and calibrated against an internal refer-
ence standard, not further specified by the manufactur-
er.
Each sample was analyzed in duplicate (calibrators,
controls, reference population, and patient samples). We
determined in-house cutoff values using 50 healthy
volunteers with the method of percentiles (99th). Quali-
ty control material, provided by the manufacturer, was
analyzed in every run [8].
The APS IgM or IgG control set provides a ready-to-use
positive control, where we know quantity of aCL or
aβ2GPI antibodies, and a negative control containing
normal human serum. As a result of the positive control
imprecision characteristics were evaluated.
Statistics
Agreement between assays was quantified using Co-
hen's κ statistic. κ coefficients > 0.75 signify substantial
agreement [10].
RESULTS
Calculation of cutoff values
The cutoff values were calculated with the 99th percen-
tile. Table 1 presents the calculated cutoff values of all
Acustar assays in comparison with the reference values
given by the manufacturer.
Validation of a New Panel of Automated Chemiluminescence Assays for Anticardiolipin Antibodies in the Screening for
Antiphospholipid Syndrome
Clin. Lab. 7/2016 3
Table 1. The comparison of cutoff values for CLIA methods.
CLIA aCL IgM aCL IgG aβ2GPI IgG aβ2GPI IgM
Manufacturer 20 20 20 20
99th percentile 15.3 13.3 14.2 6.3
Table 2. The inter-assay imprecision for CLIA methods.
aCL IgM aCL IgG aβ2GPI IgG aβ2GPI IgM
ELISA 15.3 12.35 13.45 11.05
CLIA 12.74 10.56 9.04 9.11
Table 3. The comparison agreement of CLIA and ELISA detection by Kappa values.
aCL IgG aCL IgM aβ2GPI IgG aβ2GPI IgM
Kappa value * 0.472 0.742 0.150 0.538
Kappa value ** 0.764 0.936 0.211 0.534
* - (in-house cutoff CLIA, commercial ELISA), ** - (clinically relevant cutoff).
D. Janek et al.
Clin. Lab. 7/2016
4
Validation of a New Panel of Automated Chemiluminescence Assays for Anticardiolipin Antibodies in the Screening for
Antiphospholipid Syndrome
Clin. Lab. 7/2016 5
Figure 1. The comparison of ELISA and CLIA results for aCL and aβ2GPI IgG and IgM.
Figure 2. The comparison of ELISA and CLIA results for β2GPI IgG.
D. Janek et al.
Clin. Lab. 7/2016
6
Performance characteristics
Inter-assay imprecision characteristics were calculated
from the results of the commercial positive control ma-
terial (Table 2). Coefficients of variation (CV) for the
positive control material ranged from 9.8% to 12.9%.
The negative control yielded negative results in every
run the manufacturer's cutoff is applied. As for in-house
cutoff for IgG, the normal control sample was above
this cutoff.
Comparison of results for aCL and aβ2GPI IgG and
IgM measured by CLIA and ELISA
According to Figure 1, the overall agreement between
Acustar and ELISA was calculated, as well as the sensi-
tivity and specificity of CLIA compared to ELISA for
the cohort of all patient samples.
Samples were categorized as positive or negative based
on the in-house cutoff value and the positive cutoff of
CLIA; for ELISA the manufacturer cutoff is used.
Overall agreement, sensitivity, and specificity between
CLIA and ELISA was also calculated for the APS pa-
tient population (Table 3).
In the total cohort, Kappa values range from poor to
moderate to good agreement (from 0.150 - 0.742), for
the four parameters with a 99th percentile cutoff for
CLIA and commercial cutoff for ELISA. When we ap-
ply the clinically relevant cutoff for CLIA of 50 U/mL
and for ELISA 40 IU/mL, the Kappa values significant-
ly increase to very good (0.926 - 0.936) for aCL. The
Kappa value for aβ2GPI IgM remains stable (0.534) re-
gardless of the type of cutoff. Only in the aβ2GPI IgG
does the Kappa value remain unchanged with a poor
value. This is due to the significantly higher number of
positive samples (31 patients) only in the CLIA method.
DISCUSSION
There is no doubt that the determination of aCL and
aβ2GPI antibodies is limited by standardization and
clinical relevance of tests [3]. Although, according to
recommendation where the ELISA method is preferred,
there is a new immunofluorescence method [11,12].
The aim of this study was to evaluate a new system of
chemiluminescense technology for aPL detection and to
compare the qualification of aCL and aβ2GPI antibod-
ies’ IgG and IgM with ELISA method. We also evaluat-
ed the degree of concordance between the results of
these different methodologies. As a result we deter-
mined the fully automated and computerized immuno-
analytical method, which reduced significantly the
hands-on time in comparison with the labor-intensive
ELISA assays.
The imprecision characteristics performed with the pos-
itive control material for all CLIA assays, expressed as
CV, were less than 15% (Table 1) and therefore fulfilled
the criteria of the Australian aCL Working Party [13].
Since there is no golden standard for aPL antibody as-
says, we also evaluated the diagnostic relevance of the
new assays comparing them with the presence of clini-
cally relevant cutoff values.
Laboratories are advised (according to the guidelines) to
calculate in-house cutoff values [11] using the 99th per-
centile of a normal population of at least 50 healthy vol-
unteers [1]. Our calculated cutoff values for CLIA
method with the 99th percentile are slightly different
from those which are given by the manufacturer (Table
1). Calculated cutoffs of the 99th percentile are lower
and provide a significant number of slightly positive
values. This is essential for a uniform assessment of
CLIA and ELISA methods.
There was moderate agreement (Kappa-values ranging
0.472 - 0.742) between CLIA and ELISA method re-
sults with 99th percentiles for CLIA method and manu-
facturer cutoff for ELISA without determination of
aβ2GPI IgG. The application of clinically relevant cut-
offs increased agreement for all parameters (Kappa val-
ues ranging 0.534 - 0.936) without aβ2GPI IgG. The de-
termination of aβ2GPI IgG by CLIA method shows
higher positivity in 31 samples (see Figure 1).
In the detailed evaluation of these 31 samples, there
were 26 positive findings only by CLIA method in a
different class of antibodies, and only 5 patients´ other
antibodies were negative. This indicates lower sensitivi-
ty or specificity of ELISA assays against a class of
aβ2GPI IgG antibodies. On the other hand, CLIA meth-
ods indicate higher susceptibility of the group aβ2GPI
IgG (see Figure 2).
There was no significance of correlation was revealed
between aPL and the clinical symptoms because of het-
erogeneity of aPL. Certain laboratory tests present us
with a variety of sensitivity and specificity issues which
can cause problems [3]. This is the reason why it is
highly important to develop new, more sensitive, and
specific tests with lower inaccuracies.
Tests for the detection of aPL antibodies must be suffi-
ciently sensitive to correctly classify patients with sus-
picion of APS to start the right patients on anticoagulant
therapy for prevention of recurrent thrombotic events
[14].
We can say though while there are now relatively accu-
rate recommendations for the detection of antiphospho-
lipid antibodies [11], laboratory diagnosis of APS re-
mains a problem. Chemiluminescent determination of
an automated analyzer can improve the fundamental
parameters of tests such as reproducibility between lab-
oratories [3].
Acknowledgement:
Supported by grants LF-2016-001 and RVO-2016.
Declaration of Interest:
The authors declare there is no conflict of interest.
Validation of a New Panel of Automated Chemiluminescence Assays for Anticardiolipin Antibodies in the Screening for
Antiphospholipid Syndrome
Clin. Lab. 7/2016 7
References:
1. Miyakis S, Lockshin MD, Atsumi T, et al. International consen-
sus statement on an update of the classification criteria for Defi-
nite antiphospholipid syndrome (APS). J Thromb Haemost 2006;
4(2):295-306.
2. Wilson WA, Gharavi AE, Koike T, et al. International consensus
statement on preliminary classification criteria for definite anti-
phospholipid syndrome: report of an international workshop. Ar-
thritis Rheum 1999 Jul;42(7):1309-11 Consensus Development
Conference Research Support, Non-U.S. Gov't Review.
3. Devreese K, Hoylaerts MF. Laboratory diagnosis of the antiphos-
pholipid Syndrome: a plethora of obstacles to overcome. Eur J
Haematol 2009;83(1):1-16.
4. Buller HR, Agnelli G, Hull RD, Hyers TM, Prins MH, Raskob
GE. Antithrombotic Therapy for venous thromboembolic disease:
the Seventh ACCP Conference on Antithrombotic and Thrombo-
lytic Therapy. Chest 2004;126(3 Suppl):401S-28S.
5. Tincani A, Filippini M, Scarsi M, Galli M, Meroni PL. European
attempts for the Standardization of the antiphospholipid antibod-
ies. Lupus 2009; 18(10):913-9.
6. Kuwana M, Matsuura E, Kobayashi K, Okazaki Y, Kaburaki J,
Ikeda Y, Kawakami Y. Binding of ß2 glycoprotein I to anionic
phospholipids facilitates processing and presentation of a cryptic
epitope that activates pathogenic autoreactive T-cells. Blood.
2005;105:1552-7.
7. Ichicawa K, Khamashta MA, Koike T, Matsuura E, Hughes GR.
ß2 Glycoprotein-I reactivity of monoclonal anticardiolipin anti-
bodies from patients with the antiphospholipid syndrome. Arthri-
tis Rheum. 1994;37 (10):1453-61.
8. Zucker S, Cathey MH, West B. Preparation of Quality Control
Specimens for Coagulation. Am J Clin Pathol. 1970;53:924-7.
9. Decavele AS, Schouwers S, Devreese KM. Evaluation of three
commercial ELISA kits for anticardiolipin and anti-beta2-glyco-
protein I antibodies in the laboratory Diagnosis of the antiphos-
pholipid syndrome. Int J Lab Hematol 2011;33(1):97-108.
10. J.L. Fleiss Statistical methods for rates and proportions (Second
Ed.) John Wiley, New York (1981).
11. Tincani A, Filippini M, Scarsi M, Galli M, Meroni PL. European
attempts for the standardisation of the antiphospholipid antibod-
ies. Lupus 2009;18(10):913-9.
12. Brandt JT, Triplett DA, Alving B, Scharrer I. Criteria for the di-
agnosis of lupus anticoagulants: an update. On behalf of the Sub-
committee on Lupus Anticoagulant/Antiphospholipid Antibody
of the Scientific and Standardisation Committee of the ISTH.
Thromb Haemost 1995;74(4):1185-90.
13. Wong RC; Australasian aCL Working Party. Consensus guide-
lines for anticardiolipin antibody testing. Thromb Res 2004;114
(5-6):559-71.
14. Pengo V. A contribution to the debate on the laboratory criteria
that define the antiphospholipid syndrome. J Thromb Haemost
2008;6(6):1048-9.
