Auteur Topic: Artikel NVKC > Confirmation of high levels of transglutaminase-2 antibodies by  (gelezen 10437 keer)


  • Gold Member
  • *****
  • Berichten: 1157
Artikel van NVKC
Ned Tijdschr Klin Chem Labgeneesk 2013, vol. 38, no. 4

Confirmation of high levels of transglutaminase-2 antibodies by deamidated
gliadin antibodies in the diagnosis of celiac disease in children:
a laboratory perspective

Medical Immunology, Departments of Pathology1 and Pediatric Gastroenterology2, VU University Medical Center, Amsterdam, The Netherlands

According to the ESPGHAN guidelines for celiac disease diagnosis, biopsies in symptomatic children can be omitted when transglutaminase-2 antibody (TGA) levels exceed 10xULN, the child is HLA-DQ2.5/8 positive and TGA is confirmed by endomysium antibodies (EmA).

We prospectively explored if deamidated-gliadin-peptide antibodies (DGPA) could replace EmA-confirmation of >10xULN TGA, since this would be preferred in laboratories not performing EmA.
136 sera, with >10xULN TGA were received from Dutch labs that participate in the interlaboratory quality control program within the Netherlands (SKML, section HIM).
EmA confirmed strong positive TGA in 100% (n=136) of the sera.

DGPA were measured with three different tests (EliA; Thermofisher Scientific, QUANTA Flash; INOVA, GAF-3X ELISA; Euroimmun);
not all samples were DGPA-positive (IgG 89.7-97.1%; IgA 86.0-95.6%).
DGPA can be used to confirm TGA instead of EmA, but depending on the assay used, 1.5-15% of the 136 children would still have to be biopsied to diagnose celiac disease.
DGPA provide a good alternative for EmA in this diagnostic process, for laboratories that
do not perform EmA on a routine basis.

The updated ESPGHAN guidelines for diagnosis of celiac
disease (CD) recommend that in children with typical
symptoms of celiac disease and a serum concentration
of antibodies directed against transglutaminase-2
(TGA) exceeding 10 times the upper limit of normal
(ULN), the clinician follows a diagnostic pathway in
which a duodenal biopsy can be omitted (1). Confirmation
of such high TGA titers by a positive endomysium
antibody (EmA) test (in an independent sample) and
positivity for HLA-DQ2.5 or DQ8 is now sufficient to
establish the diagnosis CD, without the previous need
of confirmation by typical findings in duodenal biopsies
like villous atrophy and epithelial lymphocytosis.

Determination of EmA is not performed by all laboratories;
it is laborious and requires extensive experience
in the technique itself but especially in interpretation.
For laboratories not performing EmA tests, it would
be convenient to replace confirmation of high TGA
by EmA with confirmation by another ELISA-based
test. As an alternative for TGA confirmation by EmA,
deamidated gliadin peptide antibodies (DGPA) may
be used. This test is available for various automated
systems and as ELISA, like the TGA test. Another
advantage could be that the antigens recognized by
DGPA differ from those recognized by TGA, whereas
EmA recognize essentially the same antigen as TGA,
i.e. transglutaminase-2.
In this study, we compared sensitivity of EmA and
DGPA in 136 sera with >10xULN TGA and observed
that DGPA is a good candidate to replace EmA as confirmatory
test. DGPA are, however, less sensitive than
EmA. Replacing EmA with DGPA will thus result in
a slightly higher number of children that should be
This study may serve as a lead for laboratories
in conjunction with their clinic in deciding which
tests to use for diagnosing celiac disease in children
with high TGA levels.

All laboratories participating in the interlaboratory
quality control program within the Netherlands
(SKML, section HIM) were invited to participate in
this prospective study (n=58); seventeen labs joined
the study. From November 2011 untill May 2012, 136
eligible sera from 136 patients, were consecutively
collected in a prospective fashion. Sera were included
on basis of the TGA values as determined by the collaborating
labs, using their cut-off value. All subjects
were younger than 18 years. From 97 subjects it was
known if duodenal biopsy took place or not: this was
the case in 16 cases (17%), all of which were Marsh
IIIa-c. Of the TGA results, 79.7 % was obtained with
EliA on the immunocap (Phadia; Thermofisher Scientific),
the remaining 20.3% was obtained with Orgentec
TGA ELISA (7.2%), Phadia VarELISA (8.0%) or
an in-house ELISA (5.1%). The following cut-off levels
were used: for EliA 7 or 10 U/mL depending on the lab
(58% and 42%, respectively), for Orgentec ELISA 10
U/mL, for VarELISA 4 U/mL, for the VUmc in-house
ELISA 6 U/mL. The 10xULN used for inclusion, thus
were 70 or 100, 40 and 60 U/mL, respectively.

DGPA, TGA and EmA tests
Within the Dutch interlaboratory quality control program
for celiac disease, organized by SKML/VUmc,
three platforms were used for DGPA testing. We decided
to use these platforms to test all sera for DGPA IgG
and IgA, as well as retest all samples with the TGA tests
of these companies. The companies and tests involved
were Thermofisher Scientific (MA): EliA/Phadia;
diagnostics (CA): BIO-Flash/QUANTA Flash
and Euroimmun (Lübeck, Germany): GAF-3X ELISA.
The cut-offs as provided by the companies were used:
for DGPA IgG and IgA: Phadia EliA 10 U/mL, INOVA
QUANTA Flash 20 U/mL, Euroimmun 25 U/mL; for
TGA Phadia EliA 10 U/mL, INOVA QUANTA Flash
20 U/mL, Euroimmun 20 U/mL; equivocal results were
considered negative. In addition, all sera were tested
with the QUANTA Flash DGPA IgG/IgA combination
test, with a cut-off of 20 U/mL. EmA tests were
performed directly upon receipt on primate oesophagus,
using a two-fold serum dilution. Fluorescence was
evaluated qualitatively by at least two independent observers.
Sera were stored at -20oC until DGPA testing.

EmA confirmation
136 serum samples with TGA exceeding 10xULN,
according to locally used cut offs, were included. All
sera were tested in the VUmc in-house EmA IF test
and all 136 confirmed as positive. For the purpose of
this study, for all 136 patients, these test results were
considered sufficient proof of celiac disease.

DGPA and TGA tests
After receipt and inclusion of all sera, they were collectively
tested for DGPA and TGA with the different
tests. For the EliA and QUANTA Flash platforms,
DGPA IgG test was most sensitive, yielding 92.6% and
97.1% positivity, whereas IgA DGPA were positive in
84.6% and 94.1% of the sera, respectively (table 1).
The Euroimmun DGPA IgG ELISA was positive for
89.7% of the sera whereas a higher level of positivity
(95.6%) was reached for IgA. The QUANTA Flash
screen (DGPA IgG/A combination) gave in 97.8% of
the cases a positive result. Highest sensitivity was
reached when positivity of IgG and/or IgA was observed
(EliA 94.1%, QUANTA Flash 98.5% and GAF-
3X 97.8% positive samples), so when DGPA IgG and
IgA were both performed.
Additionally, TGA were determined in all sera by Euroimmun
ELISA and INOVA QUANTA Flash. For all
samples not yet tested for EliA TGA by the laboratory
of origin, EliA TGA was determined as well (Fig 1).
None of the sera yielded a lower than 10xULN level
of TGA for INOVA and Euroimmun TGA tests. For
EliA, 10 sera yielded levels between 70 and 100 U/mL
TGA, thus depending on the ULN used, these samples
do or do not fulfill the criteria. Of these 10 samples,
6 (60%) were negative for one or more DGPA tests.
These 6 sera represent 27% (6/22) of all sera negative
for one or more of the seven DGPA tests. More specifically,
sera with an EliA TGA concentration between
70 and 100 U/mL, had a higher percentage of DGPA
IgG negativity (as being the most sensitive DGPA
test) as compared with EliA TGA levels >100 U/mL
(table 2). Sera with the lowest TGA concentrations in
the EliA test were also among the sera with the lowest
TGA levels in the INOVA and Euroimmun tests (data
not shown).

We found that out of 136 children’s sera with a TGA
concentration higher than 10xULN (ULN as chosen
by the collaborating lab) between 84.6 and 98.5% were
positive for DGPA, depending on the test platform and
whether IgG, IgA or both antibodies were measured.
EmA, on the other hand, were positive in all samples,
using a 1 in 2 dilution. DGPA can therefore indeed
safely be used to confirm TGA in stead of EmA. However,
even when testing for both IgG and IgA DGPA
(either test positive) 8 (for EliA), 2 (for Quantaflash) or
3 (for GAF-3-X) out of 136 children would require a
biopsy for diagnosis or EmA confirmation by another
laboratory, whereas none needed biopsy to diagnose
CD when EmA would be used as confirmatory test (2).
We support the conclusion made in the paper of Egner
et al (3), that results yielded with different TGA
tests may not be comparable and that using 10xULN
may not be suitable for all TGA tests. Accordingly,
we show here, that the choice of ULN for the TGA
test determines if a child will be biopsied or not. For
example, a child with EliA TGA levels of 80 U/mL
will be biopsied when 7 U/mL is used as ULN, but not
when 10 U/mL is used as ULN, although in both cases
EmA may confirm the high TGA levels.
Interestingly, we show here for EliA TGA that sera with levels between
70 and 100 U/mL were all positive for EmA, but
that the percentage of sera negative for one or more
DGPA tests was significantly higher in this group as
compared to the group with a TGA concentration >100
U/mL. Remarkably, TGA levels obtained with Orgentec
and Euroimmun assay all exceeded the 10xULN
far more than with the Elia TGA assay. Future
studies should reveal if confirmation of >10xULN TGA by
DGPA IgG is more specific than confirmation by
EmA. Although a previous study showed that TGA is
superior to DGPA for diagnosis (4), specificity of the
combination of these two tests being positive was not
We conclude that DGPA can be used as confirmatory
test when TGA exceeds 10xULN, in laboratories
not performing EmA. It would, however, result in a
increased number of biopsies to be performed.
The choice of TGA test and the choice of ULN (lower
or upper limit of equivocal area) critically determine
if biopsies are taken or if diagnosis is made without
biopsy. These tests thus need extensive validation in a
relevant population of children.

Table 1. Numbers (#) and percentages of 136 transglutaminase
antibody (TGA)high sera with positive deamidated gliadin peptide
antibody (DGPA) IgG and/or IgA for the performed tests.

Figure 1. Results of different TGA tests. Dashed lines represent
10xULN as provided by the companies. 80% of the EliA
results were results obtained by the collaborating labs, 20%
were obtained with other TGA tests and re-measured with EliA
at VUMC. Some labs used 70 U/mL as 10xULN for EliA TGA
test (dotted line). All values exceeding the upper test limit due
to dilution by the collaborating lab are shown as the upper test
limit value. GAF-3X results were extrapolated by Euroimmun.

Table 2. Relation between EliA TGA levels and DGPA IgG
outcome. c2 test was used to calculate significant differences
between the groups.