658223
TAG0010.1177/1756283X16658223Therapeutic Advances in GastroenterologyJ Afonso, S Lopes
research-article2016
Therapeutic Advances in Gastroenterology
Original Research
Detection of anti-infliximab antibodies
is impacted by antibody titer, infliximab
level and IgG4 antibodies: a systematic
comparison of three different assays
Joana Afonso, Susana Lopes, Raquel Gonçalves, Paulo Caldeira, Paula Lago,
Helena Tavares de Sousa, Jaime Ramos, Ana Rita Gonçalves, Paula Ministro,
Isadora Rosa, Ana Isabel Vieira, Rosa Coelho, Patrícia Tavares, João Soares,
Ana Lúcia Sousa, Diana Carvalho, Paula Sousa, João Pereira da Silva, Tânia Meira,
Filipa Silva Ferreira, Cláudia Camila Dias, Yehuda Chowers, Shomron Ben-Horin
and Fernando Magro on behalf Portuguese IBD Study Group (GEDII)
Abstract:
Background: There is scant information on the accuracy of different assays used to measure
anti-infliximab antibodies (ADAs), especially in the presence of detectable infliximab (IFX). We
thus aimed to evaluate and compare three different assays for the detection of IFX and ADAs
and to clarify the impact of the presence of circulating IFX on the accuracy of the ADA assays.
Methods: Blood samples from 79 ulcerative colitis (UC) patients treated with infliximab were
assessed for IFX levels and ADAs using three different assays: an in-house assay and two
commercial kits, Immundiagnostik and Theradiag. Sera samples with ADAs and undetectable
levels of IFX were spiked with exogenous IFX and analyzed for ADAs.
Results: The three assays showed 81–96% agreement for the measured IFX level. However,
the in-house assay and Immundiagnostik assays detected ADAs in 34 out of 79 samples,
whereas Theradiag only detected ADAs in 24 samples. Samples negative for ADAs with
Theradiag, but ADA-positive in both the in-house and Immundiagnostik assays, were positive
for IFX or IgG4 ADAs. In spiking experiments, a low concentration of exogenous IFX (5 µg/
ml) hampered ADA detection with Theradiag in sera samples with ADA levels of between 3
and 10 µg/ml. In the Immundiagnostik assay detection interference was only observed at
concentrations of exogenous IFX higher than 30 µg/ml. However, in samples with high levels
of ADAs (>25 µg/ml) interference was only observed at IFX concentrations higher than 100 µg/
ml in all three assays. Binary (IFX/ADA) stratification of the results showed that IFX+/ADAand IFX-/ADAs+ were less influenced by the assay results than the double-positive (IFX+/
ADAs+) and double-negative (IFX-/ADAs-) combination.
Conclusions: All three methodologies are equally suitable for measuring IFX levels. However,
erroneous therapeutic decisions may occur when patients show double-negative (IFX-/ADAs) or double-positive (IFX+/ADAs+) status, since agreement between assays is significantly
lower in these circumstances.
Keywords: anti-infliximab antibodies, anti-Infliximab antibody methodologies, infliximab
trough levels, therapeutic drug monitoring
Introduction
Infliximab (IFX) is a chimeric (human–murine)
monoclonal IgG1 anti-tumor necrosis factor alpha
(TNF-α) antibody used in the treatment of inflammatory bowel disease (IBD) [Bendtzen, 2013].
Although IFX has profoundly improved the treatment of inflammatory diseases, not all patients
respond to induction therapy, and up to 50% of
patients experience the loss of clinical response
over time (secondary loss of response) [Yanai and
Ther Adv Gastroenterol
2016, Vol. 9(6) 781–794
DOI: 10.1177/
1756283X16658223
© The Author(s), 2016.
Reprints and permissions:
http://www.sagepub.co.uk/
journalsPermissions.nav
Correspondence to:
Fernando Magro, MD, PhD
Gastroenterology
Department, Medical
School, Centro Hospitalar
São João, Alameda Prof.
Hernâni Monteiro,
4200-319 Porto, Portugal
fm@med.up.pt
Joana Afonso, MSc
Filipa Silva Ferreira, MD
Department of
Pharmacology and
Therapeutics, Faculty of
Medicine, University of
Porto, Porto, Portugal
MedInUP, Centre for Drug
Discovery and Innovative
Medicines, University of
Porto, 4200 Porto, Portugal
Susana Lopes, MD
Rosa Coelho, MD
Patrícia Tavares, MD
Gastroenterology
Department, Centro
Hospitalar São João,
Porto, Portugal
Raquel Gonçalves, MD
João Soares, MD
Gastroenterology
Department, Hospital de
Braga, Braga, Portugal
Paulo Caldeira, MD
Ana Lúcia Sousa, MD
Gastroenterology
Department, Centro
Hospitalar do Algarve,
Faro, Portugal
Paula Lago, MD
Gastroenterology
Department, Centro
Hospitalar do Porto, Porto,
Portugal
Helena Tavares de Sousa,
MD
Gastroenterology
Department, Centro
Hospitalar do Algarve,
Portimão, Portugal
Departament of Medicine
e Medical Biosciences,
University of Algarve, Faro,
Portugal
Jaime Ramos, MD
Diana Carvalho, MD
Gastroenterology
Department, Centro
Hospitalar de Lisboa,
Lisboa, Portugal
http://tag.sagepub.com
Creative Commons Non Commercial CC-BY-NC: This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 3.0 License (http://www.creativecommons.org/
licenses/by-nc/3.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the
SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).
781
Therapeutic Advances in Gastroenterology 9(6)
Ana Rita Gonçalves, MD
Gastroenterology
Department, Centro
Hospitalar Lisboa Norte,
Lisboa, Portugal
Paula Ministro, MD
Paula Sousa, MD
Gastroenterology
Department, Hospital de S.
Teotónio, Viseu, Portugal
Isadora Rosa, MD, PhD
João Pereira da Silva, MD
Gastroenterology
Department, Instituto
Português de Oncologia de
Lisboa, Lisboa, Portugal
Ana Isabel Vieira, MD
Tânia Meira, MD
Gastroenterology
Department, Hospital
Garcia de Orta, Almada,
Portugal
Cláudia Camila Dias, MSc
Department of
Pharmacology and
Therapeutics, Faculty of
Medicine, University of
Porto, Porto, Portugal
Yehuda Chowers, MD, PhD
Gastroenterology
Department, Rambam
Health Care Campus and
Bruce Rappaport School of
Medicine, Technion Israel
Institute of Technology,
Israel
Shomron Ben-Horin, MD,
PhD
IBD Service, Department
of Gastroenterology,
Sheba Medical Center
and Sackler School
of Medicine, Tel-Aviv
University, Israel
Figure 1. (A) Infliximab assays. (B) Anti-infliximab antibodies assays.
Hanauer, 2011; Steenholdt et al. 2013]. When
managing loss of response, clinicians may empirically intensify treatment with the existing drug
(increase dosage and/or increase frequency),
switch to another TNF-α antagonist or switch to a
totally different class of drug. This empirical
approach has disadvantages: risk of irreversible tissue damage while the physician searches for an
effective new drug, and significant economic consequences of unsuccessful trial and errors
[Bendtzen and Svenson, 2011; Steenholdt et al.
2014a]. A more astute strategy is probably to use
therapeutic drug monitoring (TDM), which enables clinicians to identify patients in whom a medication or change in medication is likely to be
effective [Roblin et al. 2014; Steenholdt et al. 2014b;
Yanai et al. 2015]. Indeed, a rational evidencebased and tailored therapy according to individual
needs may reduce delays in effective treatment
[Bendtzen, 2013; Steenholdt et al. 2014b].
Awareness of the value of TDM has led to the
development of different techniques for assessing
levels of infliximab and anti-infliximab antibodies
(ADA) in patients, but these different methodologies have distinctive limitations and may yield
different results. This potential bias may have a
significant impact on TDM results and interpretation. There is still little information allowing us
to compare different assays, in particular in relation to ADAs detection, which is likely to be subject to interference by detectable levels of IFX
[Casteele et al. 2012; Kopylov et al. 2012;
Steenholdt et al. 2013].
In order to incorporate therapeutic drug monitoring into clinical practice it is pertinent to
782
recognize the potential for assay heterogeneity
and accuracy. Therefore, the objective of this
study was to evaluate and compare three different
methodologies used to detect IFX and ADA and
to clarify the importance of detectable IFX levels
when measuring ADA levels namely on the accuracy of ADA assays.
Methods
Patients and sera
Trough blood samples were collected from 79
IFX treated ulcerative colitis (UC) patients.
Blood samples were centrifuged, and the serum
collected and stored at –80°C. This was a multicenter, open-label, single-arm trial. Study participants were recruited from ten IBD centers in
Portugal. The trial was conducted in accordance
with the Declaration of Helsinki and Ethical
Principles of Good Clinical Practice and was
approved by the local Ethics Committees. All
participants gave their written informed consent.
Evaluation of IFX levels
IFX levels were evaluated using a sandwich
enzyme-linked immunosorbent assay (ELISA)
from three different sources (Figure 1A): one inhouse ELISA and two commercial ELISA kits.
The upper limit of the measurement for the three
assays was calculated as the highest concentration of the standard curve × sample dilution factor used.
IFX levels were evaluated using the in-house
ELISA as previously described by Baert and
http://tag.sagepub.com
J Afonso, S Lopes et al.
colleagues with some modifications [Baert et al.
2003; Ben-Horin et al. 2011]. Briefly, 100 ml of
1:100 diluted serum was added to a plate precoated with 750 ng/ml TNFα (Peprotech, Rocky
Hill, NJ, USA) and incubated for 60 min.
Following washing, horseradish peroxidase
(HRP) labelled goat anti-human Fc fragment
antibody (MP Biomedicals, Solon, OH, USA) at
a concentration of 0.62 µg/ml was added for 60
min. After washing, a 3-min reaction was performed with tetramethylbenzidine (TMB) substrate and stopped with H2SO4, 2 M. The results
were read at 450/540 nm using a Power Wave
340 (Biotec Instruments). The infliximab concentration was quantified using a standard curve
constructed using exogenous infliximab (Schering
Plough, NJ, USA). The lower limit of quantification was 0.1 µg/ml.
IFX levels were also evaluated in parallel using
the
commercial
TNFα-Blocker
ELISA
(Immundiagnostik AG, Germany) and LisaTracker Premium Infliximab ELISA (Theradiag,
France) kits according to the instructions provided by the manufacturer. The lower limit of
quantification was 0.5 and 0.1 µg/ml, for
Immundiagnostik and Theradiag, respectively.
Evaluation of ADA levels
Antibodies to infliximab were evaluated using
three distinct methodologies (Figure 1B), one inhouse assay [anti-human lambda chain assay
(AHLC)], and two commercially available kits [a
semi-quantitative assay, TNFα- Blocker ADA
(Immundiagnostik AG, Germany); and a quantitative assay, Lisa-Tracker Premium Infliximab
ELISA (Theradiag, France)]. The upper limit of
the measurement for the three assays was calculated as the highest concentration of the standard
curve × sample dilution factor used.
In-house AHLC. ADA levels were determined in
all samples using an in-house ELISA as previously described by Ben-Horin and colleagues
[Ben-Horin et al. 2011]. The AHLC is a sandwich
ELISA that uses anti-human lambda chain conjugate antibody in the detection step, taking advantage of the exclusively kappa chain composition
of IFX [Kopylov et al. 2012]. Briefly, IFX was
added to a plate precoated with TNFα (Peprotech, Rocky Hill, NJ, USA). After Diluted serum
was added and incubated for 60 min at room
temperature. Goat anti-human lambda chain
HRP-labeled antibody (Serotec, Oxford, UK)
http://tag.sagepub.com
was added and incubated for 60 min, at room
temperature. After a 6-min reaction with TMB
the reaction was stopped with H2SO4. Absorbances were read at 450/540 nm and the results
are expressed as µg/ml-equivalent (µg/ml-e) after
normalization against results obtained using a
standard curve of goat anti-human F(ab’)2 fragment antibody (MP Biomedicals). For the purpose of brevity, the results are thereafter expressed
as µg/ml, rather than µg/ml-e. The lower limit of
quantification was 1.2 µg/ml.
Immundiagnostik semi-fluid phase enzyme immunoassay. ADA levels were determined in a semifluid phase enzyme immunoassay (SFPE) using a
commercial kit (TNFα-Blocker ADA) purchased
from Immundiagnostik (Germany) according to
the instructions provided by the manufacturer.
The SFPE uses an initial acid buffer treatment to
dissociate the IFX–ADA immune complexes and
two IFX conjugates with different conjugates.
Acidified samples are incubated with IFX–biotin
conjugate which immobilizes ADAs on the plate,
and IFX–peroxidase conjugate is used for detection [Imaeda et al. 2012].
Theradiag bridging ELISA. ADA levels were determined in a bridging ELISA (BE) performed using
a commercial kit (Lisa-Tracker Premium Infliximab ELISA) purchased from Theradiag (France)
according to the instructions provided by the
manufacturer. The BE uses a double-antigen
bridge: ADAs create a bridge between IFX immobilized on the plate and IFX enzyme-linked conjugate. A total of 4 out of 79 samples determined
with Theradiag were higher than the upper limit
of the kit: in these cases the upper limit was considered as the result. The lower limit of quantification was 10 ng/ml.
Exogenous IFX in ADA-positive sera samples
Exogenous IFX (Schering Plough, NJ, USA) was
used to spike ADA-positive sera samples that
contained undetectable levels of IFX.
Evaluation of exogenous IFX incubation time. The
incubation time of exogenous IFX in sera was
evaluated. Sera samples with ADAs and undetectable levels of IFX were incubated with several concentrations of IFX (5, 10, 15, 30, 100
and 300 µg/ml) for different lengths of time (0,
0.5, 3, 6, 12 and 24 h). Levels of ADAs were then
determined using the AHLC assay as described
above.
783
Therapeutic Advances in Gastroenterology 9(6)
Evaluation of stratified ADA levels (four groups)
with exogenous IFX spiking using AHLC methodology. Four groups with undetectable IFX (<0.1
µg/ml) were stratified according to ADA-positive
concentration as previously measured by AHLC
assay: group A (ADAs within 1.7–3 µg/ml); group
B (ADAs within 3–10 µg/ml); group C (ADAs
within 10–25 µg/ml); group D (ADAs >25 µg/
ml). Sera were preincubated with several concentrations of exogenous IFX (5, 10, 30, 100 and 300
µg/ml) for 30 min, at room temperature. Levels of
ADAs were then determined using the in-house
AHLC ELISA as described above.
Evaluation of ADA levels with exogenous IFX
spiking in two groups of sera samples using three
different methodologies. Two groups of ADApositive sera with undetectable IFX (<0.1 µg/ml)
were selected from sera previously measured by
AHLC assay: high levels of ADAs (>25 µg/ml)
and intermediate levels of ADAs (>3 µg/ml and
<10 µg/ml). Sera were preincubated with several
concentrations of exogenous IFX (5, 10, 15, 30,
100 and 300 µg/ml) for 30 min, at room temperature. Levels of ADAs were then determined using
the three different methodologies: in-house
AHLC, Immundiagnostik SFPE and Theradiag
BE as described above.
Measurement of ADA IgG4 in the sera
Sera samples containing ADAs were evaluated for
ADA IgG4. ADA IgG4 levels were determined
using an ELISA developed and described by
Bendtzen and colleagues and adapted for IgG4
detection [Bendtzen and Svenson, 2011;
Steenholdt, 2013]. Briefly, diluted serum was
added and incubated for 60 min at room temperature to a plate precoated with anti-IgG4 antibody
(Serotec, Oxford, UK). Plates were then washed
and HRP-labeled infliximab was added for 60
min, at room temperature. A 5-min reaction with
TMB substrate was performed and stopped with 2
M H2SO4. Absorbances were read at 450/540 nm
and the results were expressed as µg/ml-e after
normalization against results obtained using a
standard curve of goat anti-human F(ab’)2 fragment antibody (MP Biomedicals).
Statistical analysis
The intraclass correlation coefficient (ICC) was
used to determine the quantitative agreement
between IFX levels evaluated by the three assays.
ICC within 0.4 and 0.75 was considered
784
moderate, lower than 0.4 poor and higher than
0.75 excellent. Agreement between assays was
also assessed using Bland–Altman analysis, in
which the mean differences and limits of agreement are descriptive. The KAPPA coefficients
of 0–0.20 were considered to indicate slight,
0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80
substantial and 0.81–1 almost perfect agreement
[Landis and Koch, 1977]. The mean difference
(bias) and its 95% confidence interval (CI) were
computed. Limits of agreement (LOAs) were
defined as mean difference ± 2 standard deviations of difference (LOA = bias ± SDD). Ninetyfive per cent of differences are expected to lie
between LOAs.
The quantitative agreement between assays for
ADA could not be assessed because data were
reported using different and arbitrary units
(AU/ml).
Kappa coefficient and percentage agreement were
therefore used to determine the qualitative agreement between IFX levels, ADAs and status
(IFX+/ADAs-; IFX+/ADAs+; IFX-/ADAs-;
IFX-/ADAs-). Correlations between assays were
determined only for IFX levels using linear correlation analysis expressed as Pearson’s correlation coefficient (Pearson’s r).
Results
IFX levels
A total of 79 sera samples were evaluated for IFX
levels using three types of ELISA. Samples were
collected from multicentric cohort of UC patients,
composed of 44% men and with a median [interquartile range (IQR)] age of 35 (23–48) years. All
patients were in the maintenance phase (after 14
weeks). A total of 35% of the patients were in IFX
monotherapy and the remain cohort in concomitant therapy [38% azathioprine (AZA); 12%
5-aminosalicylic acid (5-ASA); 7% AZA + 5ASA;
1% methotrexate; 1% corticosteroids; 2% AZA +
5-ASA + corticosteroids; 1% AZA + corticosteroids + antibiotics].
Quantitative agreement was calculated using the
ICC and mean differences (Table 1). Comparisons showed that ICC varied from a moderate
agreement of 0.694 (0.618–0.805) between
Immundiagnostik and Theradiag up to an excellent agreement of 0.957 (0.933–0.972) between
the in-house ELISA and Immundiagnostik kit.
http://tag.sagepub.com
J Afonso, S Lopes et al.
Table 1. Intraclass correlation (ICC) and mean differences between different assays.
ICC
IFX
In-house ELISA/ Theradiag ELISA
In-house ELISA/ Immundiagnostik ELISA
Immundiagnostik ELISA / Theradiag ELISA
0.762
0.957
0.694
95% CI
0.627–0.848
0.933–0.972
0.618–0.805
Differences
Mean
95% CI
3.99
–0.43
4.43
2.63–5.36
–1.31 to 0.44
2.76–6.10
CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; IFX, infliximab.
Mean differences (bias) confirmed the ICC evaluation. Moreover, a Bland and Altman plot analysis
(graphics not shown) indicate that the significant
bias between in-house and Theradiag is a result of
dispersion of values along all of infliximab concentration, even for minor concentrations. Regarding
in-house and Immundiagnostik, bias is more significant for concentrations higher than 15 µg/ml.
The Pearson correlation between different assays
was also calculated: 0.84 (p<0.001) for in-house
versus Theradiag; 0.79 (p < 0.001) for Theradiag
versus Immundiagnostik and 0.92 (p<0.001) for
in-house versus Immundiagnostik.
In order to quantify the qualitative agreement of
the different methods, two different cutoffs were
applied to the measured concentrations: one used
the limit of detection determined by the manufacturers for each assay as the cutoff, conveying an
analytical perspective; and the other used clinically relevant cutoffs, as defined in the literature
[Kopylov et al. 2012; Ungar et al. 2015].
Regarding the analytical approach, IFX was considered positive when above 0.1 µg/ml using the
in-house and Theradiag methods. For the
Immundiagnostik kit, positivity was considered for
IFX concentrations above 0.5 µg/ml. The in-house
ELISA technique detected trough levels in 61 of
79 samples, a detection rate similar to that verified
using the Immundiagnostik method (78%). Using
Theradiag the IFX trough levels were detected in
fewer sera (71%). The accuracy and KAPPA coefficients for these methods using the analytical cutoff are depicted in Table 2(A). In-house and
Theradiag ELISA showed a moderate agreement
(0.509), whereas the in-house/Immundiagnostik
and Immundiagnostik/Theradiag had a substantial agreement (0.69 and 0.726, respectively).
Moreover, a good accuracy was found among the
three assays (81–88%).
http://tag.sagepub.com
Considering the clinical approach, IFX was
defined as positive for concentrations above
1 µg/ml. The cutoff was settled taking in consideration cutoffs reported previously in the literature [Kopylov et al. 2012; Ungar et al. 2015].
The in-house ELISA detected positive IFX
trough levels in 48 of the 79 samples (61%), while
the Immundiagnostik kit did so in 45 of the 79
samples (57%) and the Theradiag kit did so in 49
of the 79 samples (62%). The data regarding IFX
trough levels obtained using the different assays is
shown in Table 2(B).
The accuracy and KAPPA coefficient between
the three different assays were calculated (Table
2B). All assays showed excellent agreement
regarding IFX levels, with the KAPPA coefficient
ranging from 0.835 (0.709–0.960) between
Immundiagnostik and Theradiag results, to 0.922
(0.836–1.000) between the in-house assay and
Immundiagnostik results. A high level of accuracy
was found between the three assays, with the inhouse and Immundiagnostik assays showing the
highest accuracy (96%).
ADA levels
The ICC and mean difference for ADAs could
not be evaluated because the different assays
used different units of measurement (µg/ml for
in-house assay and Theradiag and AU/ml for
Immundiagnostik).
From an analytical point of view, ADAs positivity
was defined as levels above 1.2 µg/ml for in-house
method, 0.01 µg/ml for the Theradiag kit and 10
AU/ml for Immundiagnostik kit. All techniques
have identified the same number of ADAspositive samples (46%). Accuracy and KAPPA
coefficient were calculated and are depicted in
Table 3(A): a good accuracy was found for the
785
Therapeutic Advances in Gastroenterology 9(6)
Table 2. Agreement between different assays
regarding levels of IFX: (A) analytical approach and
(B) clinical approach.
Agreement
Agreement
Accuracy
KAPPA
[95% CI]
(A) IFX
(>0.1 µg/ml in-house and Theradiag; 0.5 µg/ml
Immundiagnostik)
In-house ELISA/
81%
0.509
Theradiag ELISA
[0.295–0.722]
In-house ELISA/
87%
0.697
Immundiagnostik
[0.526–0.868]
ELISA
Immundiagnostik
88%
0.726
ELISA / Theradiag
[0.559–0.893]
ELISA
(B) IFX
(>1 µg/ml)
In-house ELISA/
92%
0.840
Theradiag ELISA
[0.716–0.963]
In-house ELISA/
96%
0.922
Immundiagnostik
[0.836–1.000]
ELISA
Immundiagnostik
92%
0.835
ELISA / Theradiag
[0.709–0.960]
ELISA
CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; IFX, infliximab.
three assays (82–85%), whereas KAPPA coefficient shown a moderate agreement.
A clinical approach to the ADAs positivity is
shown in Table 3(B). Samples were considered
ADA-positive when levels exceeded 1.7 µg/ml for
the in-house (AHLC) as applied by Kopylov and
colleagues [Kopylov et al. 2012]. For Theradiag
the same cutoff was applied. For the
Immundiagnostik (SFPE) assay ADA-positive
were considered when >10 AU (cutoff defined by
the manufacturer for positivity). The in-house and
Immundiagnostik assays detected ADAs in 34 out
of 79 samples (43%), while Theradiag detected
ADAs in 24 out of 79 samples (30%). Although
the in-house and Immundiagnostik assays
detected the same number of samples with ADAs
not all samples matched. In fact, ADAs were confirmed by both the in-house and Immundiagnostik
assay in 28 out of the 34 samples (82%). Regarding
the comparative performance of Theradiag with
the in-house and Immundiagnostik assays, of the
24 samples with ADAs identified by Theradiag,
21 (87.5%) were also positive in both the in-house
786
Table 3. Agreement between different assays
regarding levels of ADAs: (A) analytical approach and
(B) clinical approach.
Accuracy
KAPPA
[95% CI]
(A) ADAs
(>1.2 µg/ml for in-house and >0.01 µg/ml
Theradiag; >10 AU for Immundiagnostik)
In-house/ Theradiag
85%
0.695
[0.539–0.851]
In-house/
82%
0.643
Immundiagnostik
[0.472–0.814]
Theradiag /
82%
0.635
Immundiagnostik
[0.463–0.807]
(B) ADAs
(>1.7 µg/ml for in-house and Theradiag; >10 AU/
ml for Immundiagnostik)
In-house/ Theradiag
81%
0.602
[0.435–0.769]
In-house/
85%
0.692
Immundiagnostik
[0.531–0.852]
Theradiag /
83%
0.653
Immundiagnostik
[0.492–0.814]
AU, arbitrary units; CI, confidence interval; ELISA,
enzyme-linked immunosorbent assay; IFX, infliximab.
and Immundiagnostik assays. Nevertheless, the
in-house and Immundiagnostik assays considered
positive 13 and 12 samples, respectively, that were
negative with Theradiag.
Moderate agreement was found between the inhouse and Theradiag assays, with a KAPPA coefficient of 0.602 (0.435–0.769) while the in-house
and Immundiagnostik assays showed good agreement with a KAPPA coefficient of 0.692 (0.531–
0.852). Accuracy ranged from 81% to 85%
among the different pairs of assays (Table 3B).
IFX and ADA status
Figure 2 depicts the levels of IFX and ADAs evaluated by all assays. The vertical dotted line separates samples positive for IFX from samples
negative for IFX. The horizontal dotted line separates samples positive for ADAs from samples
negative for ADAs. Both analytical and clinical
perspectives are represented in Figure 2. From
the analytical point of view, 71–77% of total samples are included in IFX+ group, and 45% in
anti-IFX+. On the other hand, from a clinical
perspective all the three assays identified 25% of
samples positive for ADAs and IFX-negative.
http://tag.sagepub.com
J Afonso, S Lopes et al.
occurred for Theradiag and Immundiagnostik
(considering Theradiag as the standard assay).
Figure 2. Levels of infliximab (IFX) and anti-infliximab
antibodies (ADAs) determined by all methodologies.
Dotted lines represent cutoffs for analytical approach
(A, B, D, E and F) and clinical approach (C and D).
Cutoffs for IFX levels: (A) 0.1 µg/ml (in-house and
Theradiag); (B) 0.5 µg/ml (Immundiagnostik); (C) 1 µg/ml (all
assays).
Cutoffs for ADAs levels: (D) 1.7 µg/ml (in-house and
Theradiag) and 10 AU/ml (Immundiagnostik); (E) 1.2 µg/ml
(in-house); (F) 0.01 µg/ml (Theradiag).
However, the group of samples that had a doublepositive status for both IFX and ADAs were identified almost exclusively by the in-house and
Immundiagnostik assays. Theradiag detected
only three samples that were positive for ADAs in
the presence of the drug. However, these three
samples were considered IFX-negative by the
other two assays.
Thereafter, we focused on the double status of
IFX and ADAs. As the IFX/ADAs group definition depends directly on the cutoffs applied, the
two different approaches (analytical and clinical)
were taken into consideration. Tables 4 and 5
show the agreement between assays for different
combinations of IFX/ADA status from an analytical and clinical perspective, respectively.
IFX+/ADAs-. The analytical approach has shown
that 50% of samples had an IFX+/ADAs-status
(Table 4). Agreement between assays was considered good (85%).
The clinical perspective exhibited a strong agreement ranging between 75.6 and 100% (Table 5).
The poorest agreement between assays was
observed when using Theradiag as the reference
assay and Immundiagnostik as the confirmatory
assay (75.6%).
IFX+/ADAs+. The
double-positive
samples
obtained using the cutoffs given by the analytical
perspective is shown in Table 4. The agreement
was low (31.3–50%). The lowest agreement rate
http://tag.sagepub.com
In contrast, for the double-positive status (IFX+/
ADAs+) defined using the clinical approach,
there was a significant discrepancy between
Theradiag and the in-house and Immundiagnostik
assays. The in-house assay detected a total of 11
double-positive sera and Immundiagnostik a total
of 10 IFX+/ADAs+. Theradiag was able to
detect three double-positive samples but neither
were in agreement with the other assays. In fact,
in those three samples, both the in-house and
Immundiagnostik assay detected ADAs whereas
IFX was considered undetectable.
IFX-/ADAs-. Using the analytical perspective
(Table 4) only two or three double-negative samples were detected. Furthermore, Theradiag and
the other two assays never agreed. The in-house
and Immundiagnostik kit agreed in one out of
three samples.
The detection of samples with double-negative
status (IFX-/ADAs-) by clinical approach showed
similar levels of agreement between pairs of assays
(Table 5). When considering the in-house and
Immundiagnostik assay there was 62.5% agreement, but when considering Theradiag as the reference assay, there was a lower level of agreement
between methodologies (50–55.6%).
IFX-/ADAs+. From an analytical perspective,
Immundiagnostik and Theradiag had a good
agreement (75–90%) concerning IFX-/ADAs+.
A very good agreement was also obtained for this
status between the in-house and Immundiagnostik assays when considering in-house the standard test.
From the clinical point of view, IFX-/ADAs+ status showed good agreement (more than 84%)
between the in-house and Immundiagnostik assays.
The in-house and Theradiag assays showed a lower
level of agreement, although it was still considered
good at 90.5% and 79.2%, respectively.
Exogenous IFX in ADA-positive samples
To evaluate the effect of IFX on in-house,
Immundiagnostik and Theradiag assessment of
ADAs in sera, an experiment was designed that
involved spiking with exogenous IFX. The influence of exogenous IFX incubation time was
assessed previously and no differences were found
787
Therapeutic Advances in Gastroenterology 9(6)
Table 4. Agreement between the three assays for all combinations of IFX/ADAs status (analytical approach).
Agreement
In-house assay versus Theradiag assay(a)
IFX + ADA +
38.1% (n = 21)
IFX + ADA –
85% (n = 40)
IFX + ANTI +
50% (n = 16)
IFX - ADA –
0% (n = 2)
IFX – ADA +
75.0% (n = 16)
IFX - ADA –
0% (n = 2)
IFX – ADA+
57.1% (n = 21)
68%
0.502
[0.355–0.649]
72%
0.566
[0.425–0.707]
77%
0.562
[0.415–0.707]
Immundiagnostik assay versus Inhouse assay(d)
IFX + ADA –
85.0% (n = 40)
IFX + ADA +
35.0% (n = 20)
IFX + ADA –
85.0% (n = 40)
IFX + ADA +
58.3% (n = 12)
IFX - ADA –
33.3% (n = 3)
IFX – ADA +
93.8% (n = 16)
IFX - ADA –
33.3% (n = 3)
IFX – ADA +
62.5% (n = 24)
Immundiagnostik assay versus Theradiag
assay(e)
KAPPA
[95% CI]
Theradiag assay versus In-house
assay(b)
IFX + ADA–
85% (n = 40)
In-house assay versus Immundiagnostik(c)
Accuracy
Theradiag assay versus
Immundiagnostik(f)
IFX + ADA –
84.6% (n = 39)
IFX + ADA +
41.7% (n = 12)
IFX + ADA –
82.5% (n = 40)
IFX + ADA +
31.3% (n = 16)
IFX - ADA –
0.0% (n = 2)
IFX – ADA +
75.0% (n = 24)
IFX - ADA –
0.0% (n = 1)
IFX – ADA +
90.0% (n = 20)
(a)In-house
assay was considered the reference assay and it was evaluated whether Theradiag confirmed the double status result.
assay was considered the reference assay and it was evaluated whether in-house confirmed the double status result.
(c)In-house assay was considered the reference assay and it was evaluated whether Immundiagnostik confirmed the double status result.
(d)Immundiagnostik assay was considered the reference assay and it was evaluated whether in-house confirmed the double status result.
(e)Immundiagnostik assay was considered the reference assay and it was evaluated whether Theradiag confirmed the double status result.
(f)Theradiag assay was considered the reference assay and it was evaluated whether Immundiagnostik confirmed the double status result.
CI, confidence interval.
(b)Theradiag
(data not shown). An incubation time of 30 min
was selected for further experiments.
Evaluation of ADA levels with exogenous IFX spiking in four groups of sera samples. To evaluate
whether different intrinsic concentrations of
ADAs are important when assessing the influence
of exogenous IFX, we used stratified concentrations of ADA levels in sera (four different groups
as evaluated previously: 1.7–3, 3–10, 10–25 and
>25 µg/ml) and spiked them with exogenous
IFX. The evaluation was performed using the inhouse assay only. The results are presented in Figure 3. In-house assay lost the capability to detect
levels of antibodies in sera samples with ADA levels between 1.7 and 3 µg/ml at concentrations of
IFX over 5 µg/ml. The influence of exogenous
IFX was also evident in the 3–10 µg/ml group, but
788
loss of antibody detection capacity was only
observed beyond 100 µg/ml IFX, which generally
exceeds the therapeutic concentrations of infliximab in vivo even during induction phase [Adedokun et al. 2014]. A decreased ADA detection
capacity was observed in the two upper groups
(10–25 and >25 µg/ml) in a concentrationdependent manner, but in both groups the inhouse assay was able to detect antibodies even
with 300 µg/ml of exogenous IFX. Nevertheless, a
significant reduction of detection capacity of 75%
and 61% was observed in the 10–25 µg/ml and
>25 µg/ml groups, respectively.
Evaluation of ADA levels with exogenous IFX spiking
in two groups: high and intermediate levels of
ADAs. Taking into account the above-mentioned
results, we decided to explore the performance of
http://tag.sagepub.com
J Afonso, S Lopes et al.
Table 5. Agreement between the three assays for all combinations of IFX/ADAs status (clinical approach).
Agreement
In-house assay versus Theradiag
assay(a)
IFX + ADA +
0% (n = 11)
IFX + ADA –
77.8% (n = 45)
IFX + ANTI +
0% (n = 3)
IFX - ADA –
62.5% (n = 8)
IFX – ADA +
79.2% (n = 24)
IFX - ADA –
55.6% (n = 9)
IFX – ADA+
90.5% (n = 21)
76%
0.620
[0.489–0.751]
81%
0.716
[0.591–0.841]
76%
0.625
[0.491–0.758]
Immundiagnostik assay versus In-house
assay(d)
IFX + ADA –
83.8% (n = 37)
IFX + ADA +
54.5% (n = 11)
IFX + ADA –
88.6% (n = 35)
IFX + ADA +
60% (n = 10)
IFX - ADA –
62.5% (n = 8)
IFX – ADA +
95.7% (n = 23)
IFX - ADA –
62.5% (n = 8)
IFX – ADA +
84.6% (n = 26)
Immundiagnostik assay versus
Theradiag assay(e)
KAPPA
[95% CI]
Theradiag assay versus In-house assay(b)
IFX + ADA–
100% (n = 35)
In-house assay versus
Immundiagnostik(c)
Accuracy
Theradiag assay versus
Immundiagnostik(f)
IFX + ADA –
100% (n = 34)
IFX + ADA +
0% (n = 10)
IFX + ADA –
75.6% (n = 45)
IFX + ADA +
0% (n = 3)
IFX - ADA –
66.7% (n = 6)
IFX – ADA +
76.9% (n = 26)
IFX - ADA –
50% (n = 8)
IFX – ADA +
100% (n = 20)
(a)In-house
assay was considered the reference assay and it was evaluated whether Theradiag confirmed the double status result.
assay was considered the reference assay and it was evaluated whether in-house confirmed the double status result.
(c)In-house assay was considered the reference assay and it was evaluated whether Immundiagnostik confirmed the double status result.
(d)Immundiagnostik assay was considered the reference assay and it was evaluated whether in-house confirmed the double status result.
(e)Immundiagnostik assay was considered the reference assay and it was evaluated whether Theradiag confirmed the double status result.
(f)Theradiag assay was considered the reference assay and it was evaluated whether Immundiagnostik confirmed the double status result.
CI, confidence interval.
(b)Theradiag
the three assays under the interference of exogenous IFX. The two most representative groups
were used: samples with ADAs within 3–10 µg/ml
(intermediate) and those with >25 µg/ml (high
level). Figures 4 and 5 present the results of the
groups spiked with exogenous IFX at different concentrations (0, 5, 10, 15, 30, 100 and 300 µg/ml).
Table 6 shows the half maximal inhibitory concentrations (IC50).
Figure 4 (high levels of ADAs) illustrates that
exogenous IFX only interfered with ADA detection at concentrations higher than 100 µg/ml.
Spiking 100 µg/ml of exogenous IFX in sera samples with high levels of ADA resulted in a decline
in ADA detection of 87%, 46% and 16% in the
Theradiag, in-house and Immundiagnostik assays,
respectively. Nevertheless, in-house showed a
http://tag.sagepub.com
higher IC50 than Theradiag, at 143.6 µg/ml versus
74.9 µg/ml, respectively, indicating that it was less
affected than Theradiag by the presence of IFX.
In the Immundiagnostik assay the influence of
exogenous IFX showed a lower decline in ADA
detection. In fact, 300 µg/ml decreased ADA
detection by 17% (the IC50 was therefore impossible to determine), while in in-house and
Theradiag assays, ADA detection decreased by
75% and 93%, respectively. However, all methodologies were able to detect ADAs, in sera originally positive for high levels of ADAs, even with
300 µg/ml of exogenous IFX.
Figure 5 shows exogenous IFX interference in
sera samples with intermediate levels of ADAs.
The presence of exogenous IFX decreased ADA
detection in Theradiag and in-house with the
789
Therapeutic Advances in Gastroenterology 9(6)
Figure 3. Stratified anti-infliximab antibodies (ADA)
levels in the presence of exogenous infliximab
(IFX), evaluated by in-house assay. Therapeutic IFX
concentrations range was considered between 0 and
100 µg/ml [Adedokun et al. 2014].
lower concentration (5 µg/ml) of exogenous IFX.
For Theradiag this decrease led to the inability to
detect ADAs and consequently it was impossible to determine the IC50. In-house showed an
IC50 of 0.88 µg/ml while Immundiagnostik
showed a higher value of 140.1 µg/ml. At 300
µg/ml the interference was stronger in Theradiag
(97%) while in in-house detection decreased by
74%. In-house failed to detect ADAs at IFX
concentrations of 300 µg/ml, which are above
the therapeutic levels achieved in vivo. Above
30 µg/ml IFX the Immundiagnostik assay was
influenced by an additive concentration effect
but was able to detect ADAs even at the higher
concentration (300 µg/ml) of exogenous IFX
(total decrease of 60%).
Figure 4. High anti-infliximab antibodies (ADAs) levels in the presence of exogenous infliximab (IFX) evaluated
by in-house, Immundiagnostik and Theradiag assay. Therapeutic IFX concentrations range was considered
between 0 and 100 µg/ml [Adedokun et al. 2014].
Figure 5. Intermediate anti-infliximab antibodies (ADAs) levels in the presence of exogenous infliximab (IFX)
evaluated by in-house, Immundiagnostik and Theradiag assay. Therapeutic IFX concentrations range was
considered between 0 and 100 µg/ml [Adedokun et al. 2014].
790
http://tag.sagepub.com
J Afonso, S Lopes et al.
Table 6. Infliximab (IFX) inhibition effect on ADAs at intermediate and high levels in different assays.
IC50 (µg/ml)
In-house (AHLC)
Immundiagnostik (SPFE)
Theradiag (BE)
Intermediate levels
3 µg/ml > ADA <10 µg/ml
High levels
ADA >25 µg/ml
0.88
140.10
NA
143.60
NA
74.89
IC50, half maximal inhibitory concentration; NA, IC50 value not determined because either 50% inhibition fell outside the
highest concentration or because inhibition was 100% above the lowest concentration of IFX; ADA, anti-infliximab antibodies; AHLC, anti-human lambda chain; SPFE, semi-fluid phase enzyme; BE, bridging enzyme-linked immunosorbent
assay.
ADAs and to clarify the importance of the presence of detectable levels of IFX on the accuracy
of the ADA assays.
Figure 6. Levels of IFX and IgG4 in samples that were
ADA-negative in Theradiag and ADA-positive in both
the in-house and Immundiagnostik assays.
IgG4 ADAs
In order to evaluate the influence of IgG4 ADAs,
samples positive for ADAs were tested for the
presence of IgG4. IgG4 ADAs were found in
54.8% of the ADA-positive samples. Figure 6
shows the levels of IFX and IgG4 in samples that
were ADA-negative with Theradiag but ADApositive for both the in-house and Immundiagnostik
assays. Of the eight samples negative for antibodies with Theradiag, five presented positive levels
of IFX and three presented IgG4 ADAs.
Discussion
Several methodologies are available for use in drug
monitoring. However, different methodologies
return different results. There is still little information comparing different assays, in particular for
ADA detection, which is susceptible to interference in the presence of detectable levels of IFX.
The aim of this study was to evaluate and compare three different assays for detecting IFX and
http://tag.sagepub.com
Cutoffs for a qualitative evaluation of the assays
are difficult to establish, particularly for ADAs
assays. In fact, ADAs results are expressed in different units by the different assays. In addition,
Immundiagnostik defines a cutoff for positivity
that is different from the limit of detection and not
suggested by all commercial assays. The manufacturers’ instruction for Theradiag do not contain
any suggestion for a cutoff for ADAs positivity
besides the limit of detection of 0.01 µg/ml. Van
Schouwenburg and colleagues have shown that BE
assays (Theradiag methodology) performance is
depend not only of the amount of antibody in the
serum but also of the affinity of ADAs to the assay
[Van Schouwenburg et al. 2016]. This indicates a
possible source of intervariability of the assay,
namely for values near the limit of detection. In
fact, our results show that with Theradiag 70% of
the values considered to be positive, with analytical
approach, are between 0.01 and 0.04 µg/ml. This
might be a reflection of the intervariability of the
assay, rather than a faithful quantification of the
presence of antibodies. Taking this into consideration, we have decided to perform two different
approaches in the qualitative analyses of the results:
an analytical one, in which cutoffs were defined by
the manufacturers’ instructions; and a clinical one,
in which the cutoffs were defined according to the
literature.
From an analytical perspective, IFX assays
showed a substantial agreement and a good
accuracy, although the in-house and Theradiag
methods had only a moderate agreement. On the
other hand, upon applying a cutoff of 1 µg/ml,
the behavior of the three assays is very similar.
791
Therapeutic Advances in Gastroenterology 9(6)
This suggests that using the limit of detection as
a cutoff may decrease the quality agreement
between assays.
The behavior of the three assays in terms of IFX
determination was very similar. In fact, an accuracy of over 92% for IFX was observed when analyzing the agreement between the three assays. An
agreement of 100% (data not shown) was
observed
between
the
in-house
and
Immundiagnostik assays for IFX-negative samples, while the Theradiag assay identified four
IFX-positive samples that the other two assays
returned as negative. Casteele and colleagues also
evaluated three IFX assays based on the sandwich
ELISA principle and found good correlation
between them [Casteele et al. 2012]. The authors
used the same commercial kit as used here
(Theradiag). In fact, Theradiag returned 11 positive samples that were negative according to the
other two assays, showing that this assay can
detect nonspecific binding. Overall, high titers in
one assay were also high in the other two assays.
Nevertheless, the Immundiagnostik and in-house
assays were able to detect higher levels of IFX in
sera samples, which may indicate the use of a
more specific anti-human IgG antibody in the
detection step in these assays than in Theradiag.
The agreement between the three assays seemed
to be more qualitative than quantitative. The
three methodologies use the same ELISA sandwich principle, but there are some idiosyncratic
differences that can explain the small discrepancies in quantitative evaluation: different detection
limits and test sensitivity; reagent stability; different calibration standards and curve adjustment;
and different sample dilution, manipulation and
processing (e.g. different incubation times).
Nevertheless, from a clinical point of view, the
information given by the three assays (presence of
positive/negative levels; low/therapeutic/high levels of the drug) showed acceptable interassay
agreement. However, clinicians should be aware
that changing drug level assays during patient follow up could induce errors in the interpretation of
results and subsequent therapeutic strategies.
ADAs were evaluated using three methodologies:
in-house AHLC, Immundiagnostik SFPE and
Theradiag BE. These three assays have different
methodological principles. The in-house AHLC
is a sandwich ELISA that uses anti-human
lambda chain conjugate antibody in the detection
step, benefitting from IFX kappa chain exclusive
composition [Kopylov et al. 2012]. The Theradiag
792
BE uses a double-antigen bridge: ADAs create a
bridge between IFX immobilized on the plate
and IFX enzyme-linked conjugate. The
Immundiagnostik SFPE uses an initial acid buffer
treatment to dissociate the IFX–ADA immune
complexes and two IFX conjugates with different
types of conjugates. Acidified samples are incubated with an IFX–biotin conjugate, which
immobilizes ADAs to the plate, whilst an IFX–
peroxidase conjugate is used for detection.
A total of 79 samples of sera were compared using
the three assays. The analytical approach showed
the same moderate agreement for the three assays
than the clinical approach. However, and from
the analytical perspective, Theradiag was able to
detect 37 positive ADAs, whereas from a clinical
perspective only 24 patients were considered to
be positive for ADAs. Conversely, the in-house
assay detected almost the same number of ADApositive patients with the two approaches (34 for
analytical versus 37 for clinical). The results
obtained using a clinical perspective with the
Theradiag kit suggest a high rate of false-negative
outcomes. This high prevalence of false-negative
outcomes with Theradiag was reported previously by other authors [Imaeda et al. 2012;
Steenholdt et al. 2013]. Kopylov and colleagues
reported that BE was unable to detect ADAs in
the presence of IFX [Kopylov et al. 2012]. Our
results showed that 22 out of 24 ADA-positive
samples were IFX-negative. Only three patients
showed a double-positive status with Theradiag,
however neither was confirmed as IFX-positive in
the in-house or Immundiagnostik assay. These
samples were probably false positive for IFX. In
fact, the disagreement of the in-house and
Immundiagnostik assay with Theradiag for double-positive status (IFX+/ADAs+) was 100%.
Since Theradiag uses labeled IFX as the detection antibody, the presence of IFX may compete
with the detection of the ADAs. On the other
hand, the BE is unable to detect monovalent
IgG4 ADAs, contributing to the false-negative
results. Our results were consistent with these two
statements and showed that samples positive for
ADAs with the in-house and Immundiagnostik
assays, but negative with Theradiag, one of two
situations occurred: (i) sera samples were IFXpositive; or (ii) ADAs were IgG4 antibodies. The
inability to detect antibodies in the presence of
the drug is important as two recent studies have
shown that patients with IFX levels >3 µg/ml but
with positive ADA have significantly higher levels
of C-reactive protein (CRP) and less mucosal
http://tag.sagepub.com
J Afonso, S Lopes et al.
healing [Casteele et al. 2015; Yanai et al. 2015],
indicating a reduced control of inflammation
mediated by these antibodies even when drug levels are adequate.
In order to understand the influence of drug presence on ADA detection, we designed an experiment using IFX-negative sera that were incubated
with different concentrations of exogenous IFX.
The in-house assay was slightly affected by 5 µg/ml
exogenous infliximab but was able to detect antibodies up to 100 µg/ml of IFX. Kopylov and colleagues described the same drug concentration
dependency for the AHLC assay [Kopylov et al.
2012]. However, our results permit us to define
which concentration of IFX decrease the assay
capacity to determine ADAs. The Immundiagnostik
assay showed the best behavior in the presence of
exogenous IFX. Interference was observed above
100 µg/ml of exogenous IFX but the assay was able
to detect ADAs even at the higher concentration of
300 µg/ml of IFX. However, the results obtained in
sera with high levels of ADAs were surprising and
to the best of the authors’ knowledge have not
been reported previously. Even the Theradiag BE,
which is widely described as being unable to detect
antibodies in the presence of the drug, was able to
identify ADA-positive samples in the presence of
high levels of exogenous IFX. The results indicated that assays are not only limited by the levels
of drug in the sera but also by the concentration of
ADAs. These phenomena could be explained as
follows: (i) in Theradiag, ADAs bind to all free
drug in serum but there is still enough ADAs to
bind to the plated IFX and bridge it to the conjugated one; (ii) in Immundiagnostik high levels of
ADAs are most likely to diminish free IFX
interference.
In summary, for IFX drug level determination,
the three methodologies are equivalent, however
the agreement between them seems to be more
qualitative than quantitative. Regarding ADAs,
despite being equally effective, the three assays
show significant differences. Disagreement
increased when samples had a double-positive or
double-negative status, probably related to the
specific limitations of each assay. The choice of
cutoff to discriminate positive versus negative also
enhances disagreement. This led us to define two
different approaches with different cutoffs and
present both evaluations. Clinical approach seems
to highlight inabilities of the assays. In fact,
Theradiag showed inability to detect positive
ADA levels in samples with IFX (<5 µg/ml) or a
http://tag.sagepub.com
high percentage of IgG4 ADAs. IgG4 ADAs are
monovalent and unable to create a bridge with the
labeled IFX used in detection step. In contrast,
both the in-house assay and the Immundiagnostik
assay are able to detect ADAs in the presence of
IFX up to levels of >100 µg/ml which conform
with the therapeutic concentrations observed in
patients receiving IFX. Unexpectedly, in samples
with high levels of ADAs (>25 µg/ml), the presence of IFX was not as important and even
Theradiag was able to detect antibodies.
Nevertheless, Immundiagnostik was least affected
by IFX, followed by the in-house assay. In conclusion, our results indicated that assays are not only
limited by the levels of drug in the sera but also by
the concentration of ADAs.
Clinicians must be aware when optimizing treatment that binary (IFX/ADAs) stratification of
results as positive and negative can differ according to the assay used. There is a strong agreement
between assays for IFX+/ADAs- and IFX-/
ADAs+ status. The choice of assay will probably
have little influence on therapeutic decisions to
change the class of drug (IFX+/ADA-) or change
anti-TNFα antibody (IFX-/ADAs+). However,
erroneous therapeutic decisions may occur when
patients show double-negative or double-positive
status, since agreement between assays is significantly lower in these circumstances. Treatment
intensification (IFX-/ADAs-) and change of drug
class or concomitant use of immunomodulators
(IFX+/ADAs+) should take into account the fact
that the results are assay dependent.
Acknowledgements
We thank Sandra Dias for her contribution as the
GEDII coordinator.
Joana Afonso performed all laboratory experiments and participated in the data analysis, interpretation and drafting of the manuscript. Susana
Lopes, Raquel Gonçalves, Paulo Caldeira, Paula
Lago, Helena Tavares de Sousa, Jaime Ramos,
Ana Rita Gonçalves, Paula Ministro, Isadora Rosa,
Ana Isabel Vieira, Rosa Coelho, Patrícia Tavares,
João-Bruno Soares, Diana Carvalho, Paula Sousa
and Tânia Meira participated in patient recruitment and data collection. Filipa Silva Ferreira participated in data analysis and interpretation.
Claudia Camila Dias performed the statistical
analysis and participated in data analysis and interpretation. Yehuda Chowers and Shomron BenHorin revised the manuscript critically for
important intellectual content. Fernando Magro
conceived and designed the study, participated in
793
Therapeutic Advances in Gastroenterology 9(6)
data analysis, interpretation and drafting of the
manuscript, and revised it critically for important
intellectual content. All authors read and approved
the final version of the manuscript.
Funding
This work was supported by the Portuguese
Group of Studies in Inflammatory Bowel Disease
(GEDII), Portugal, and partially by the European
Regional
Development
Fund
(ERDF);
COMPETE (Operational Programme ‘Thematic
Factors of Competitiveness) and Fundação para
a Ciência e a Tecnologia, Portugal (FCT) [grant
number EXPL/IVC-PEC/1302/2013].
Conflict of interest statement
The authors declare that there is no conflict of
interest.
References
Adedokun, O., Sandborn, W., Feagan, B., Rutgeerts,
P., Xu, Z., Marano, C. et al. (2014) Association
between serum concentration of infliximab and
efficacy in adult patients with ulcerative colitis.
Gastroenterology 147: 1296.
Baert, F., Noman, M., Vermeire, S., Van Assche, G.,
D’haens, G., Carbonez, A. et al. (2003) Influence of
immunogenicity on the long-term efficacy of infliximab
in Crohn’s disease. N Engl J Med 348: 601–608.
Ben-Horin, S., Yavzori, M., Katz, L., Kopylov, U.,
Picard, O., Fudim, E. et al. (2011) The immunogenic
part of infliximab is the F(Ab ‘)(2), but measuring
antibodies to the intact infliximab molecule is more
clinically useful. Gut 60: 41–48.
Bendtzen, K. (2013) Personalized medicine:
theranostics (therapeutics diagnostics) essential
for rational use of tumor necrosis factor-alpha
antagonists. Discov Med 83: 201–211.
Bendtzen, K. and Svenson, M. (2011) Detection and
Quantification of Antibodies to Bipharmaceuticals: Practical
and Applied. Chichester: John Wiley & Sons Ltd.
Casteele, N., Buurman, D., Sturkenboom, M.,
Kleibeuker, J., Vermeire, S., Rispens, T. et al. (2012)
Detection of infliximab levels and anti-infliximab
antibodies: a comparison of three different assays.
Aliment Pharmacol Therapeut 36: 765–771.
Visit SAGE journals online
http://tag.sagepub.com
SAGE journals
794
Casteele, N., Khanna, R., Levesque, B., Stitt, L.,
Zou, G. Singh, S. et al. (2015) The relationship
between infliximab concentrations, antibodies to
infliximab and disease activity in Crohn’s disease. Gut
64: 1539–1545.
Imaeda, H., Andoh, A. and Fujiyama, Y. (2012)
Development of a new immunoassay for the accurate
determination of anti-infliximab antibodies in
inflammatory bowel disease. J Gastroenterol 47: 136–143.
Kopylov, U., Mazor, Y., Yavzori, M., Fudim, E.,
Katz, L. Coscas, D. et al. (2012) Clinical utility
of antihuman lambda chain-based enzyme-linked
immunosorbent assay (ELISA) versus double antigen
ELISA for the detection of anti-infliximab antibodies.
Inflamm Bowel Dis 18: 1628–1633.
Landis, J. and Koch, G. (1977) Measurement of
observer agreement for categorical data. Biometrics 33:
159–174.
Roblin, X., Rinaudo, M., Del Tedesco, E., Phelip,
J., Genin, C., Peyrin-Biroulet, L. et al. (2014)
Development of an algorithm incorporating
pharmacokinetics of adalimumab in inflammatory
bowel diseases. Am J Gastroenterol 109: 1250–1256.
Steenholdt, C. (2013) Use of infliximab and antiinfliximab antibody measurements to evaluate and
optimize efficacy and safety of infliximab maintenance
therapy in Crohn’s disease. Danish Med J 60(4):
B4616.
Steenholdt, C., Ainsworth, M., Tovey, M., Klausen,
T., Thomsen, O., Brynskov, J. et al. (2013)
Comparison of techniques for monitoring infliximab
and antibodies against infliximab in Crohn’s disease.
Ther Drug Monitor 35: 530–538.
Steenholdt, C., Bendtzen, K., Brynskov, J., Thomsen,
O. and Ainsworth, M. (2014a) Clinical implications of
measuring drug and anti-drug antibodies by different
assays when optimizing infliximab treatment failure
in Crohn’s disease: post hoc analysis of a randomized
controlled trial. Am J Gastroenterol 109: 1055–1064.
Steenholdt, C., Brynskov, J., Thomsen, O., Munck,
L., Fallingborg, J., Christensen, L. et al. (2014b)
Individualised therapy is more cost-effective than dose
intensification in patients with Crohn’s Disease who
lose response to anti-TNF treatment: a randomised,
controlled trial. Gut 63: 919–927.
Ungar, B., Anafy, A., Yanai, H., Ron, Y., Yavzori,
M., Picard, O. et al. (2015) Significance of low level
infliximab in the absence of anti-infliximab antibodies.
World J Gastroenterol 21: 1907–1914.
Van Schouwenburg, P., Kruithof, S., Wolbink,
G., Wouters, D. and Rispens, T. (2016) Using
monoclonal antibodies as an international standard
for the measurement of anti-adalimumab antibodies. J
Pharm Biomed Anal 120: 198–201.
Yanai, H. and Hanauer, S. (2011) Assessing response
and loss of response to biological therapies in IBD.
Am J Gastroenterol 106: 685–698.
Yanai, H., Lichtenstein, L., Assa, A., Mazor, Y.,
Weiss, B., Levine, A. et al. (2015) Levels of drug and
antidrug antibodies are associated with outcome of
interventions after loss of response to infliximab or
adalimumab. Clin Gastroenterol Hepatol 13: 522–530.
http://tag.sagepub.com