108
Originalarbeit
Diagnostic performance of three serologic tests in childhood celiac disease
Authors
P. Bufler, G. Heilig, G. Ossiander, F. Freudenberg, V. Grote, S. Koletzko
Affiliation
Dr. von Hauner Children's Hospital, Ludwig Maximilians University Munich, München
Schlüsselwörter
Zusammenfassung
Abstract
"
!
!
Hintergrund: IgA- und IgG-Antikörper gegen deaminierte Gliadin-Peptide (DGP) binden das krankheitsauslösende Antigen und könnten deshalb bei Patienten mit Zöliakie besser zur Verlaufskontrolle unter Gluten-freier Diät (GFD) geeignet sein als Gewebs-Transglutaminase 2 (TG2)-IgA. Ziel dieser Studie war der Vergleich von DGP-IgG und DGPIgA mit TG2-IgA 4 verschiedener Hersteller zur Diagnostik und Verlaufskontrolle von Kindern mit Zöliakie. Patienten und Methodik: Es wurden 411 Seren von 91 IgA-kompetenten Kindern mit Biopsie-gesicherter Zöliakie zum Zeitpunkt der Diagnosestellung sowie unter GFD untersucht und mit Seren von 98 Kontrollpatienten verglichen. Die Tests (TheBindingSite, Euroimmun, Phadia/Thermo Fisher Scientific, INOVA) zur Detektion von TG2-IgA, DGP-IgG und DGP-IgA wurden entsprechend der Herstellerangaben durchgeführt. Ergebnisse: TG2-IgA und DGP-IgG hatten eine signifikant höhere Sensitivität (100 % und 90 − 100 %) für die Diagnose Zöliakie als DGP-IgA (67 − 86 %). Alle Tests hatten eine vergleichbar hohe Spezifität (97 – 100 %). Die Häufigkeit von TG2-IgA > 10 × des oberen Normwerts (ULN) zum Zeitpunkt der Diagnosestellung lag zwischen 47 und 90 %. Unter GFD normalisierte sich DGP-IgA früher als DGPIgG und TG2-IgA. Bei Diätfehlern hatten TG2-IgA die deutlichsten Titeranstiege. Schlussfolgerung: Die kombinierte Messung von TG2-IgA und DGP-IgG hat im Vergleich zur alleinigen Bestimmung von TG2-IgA bei IgA-kompetenten Kindern mit Zöliakie keinen diagnostischen Vorteil. Die hohe Variabilität der TG2-IgA > 10 × ULN zeigt, dass eine Harmonisierung der kommerziellen Tests anzustreben ist. Diätfehler unter GFD werden am besten durch einen Anstieg der TG2-IgA diagnostiziert.
Background: IgA- and IgG-antibodies against deamidated gliadin peptides (DGP) specifically bind the disease-inducing antigen and might be superior to transglutaminase type 2 (TG2) IgA in monitoring patients on a gluten-free diet (GFD). The aim of this study was to compare the performance of DGP-IgG and DGP-IgA with TG2-IgA of four manufacturers in pediatric celiac patients at diagnosis and during follow-up under a GFD. Patients and Methods: In total 411 sera of 91 IgA competent children with biopsy proven celiac disease were analyzed at diagnosis and during follow-up on a GFD. Ninety-eight children with normal duodenal histology served as controls. The tests (TheBindingSite, Euroimmun, Phadia, part of Thermo Fisher Scientific, INOVA) for detection of TG2-IgA, DGP-IgG and DGP-IgA were used according to the manufacturers’ instructions. Results: Sensitivity to diagnose CD was high for TG2-IgA (100 %) and DGP-IgG (90 − 100 %), but lower for DGP-IgA (67 − 86 %). Specificity was high for all tests (97 – 100 %). The frequency of TG2-IgA titers > 10 × upper limit of normal at diagnosis ranged from 47 − 90 %. Under a GFD DGPIgA became negative more rapidly than DGP-IgG and TG2-IgA. Non-adherence to GFD was best indicated by positive TG2-IgA. Conclusions: Combined testing for TG2-IgA and DGP-IgG does not increase the detection rate of CD in IgA competent children compared to TG2IgA only. There are significant differences with respect to proportions of celiac children with titers > 10 × ULN between the manufacturers. This calls for harmonization of tests. TG2-IgA showed the highest titer rise with non-adherence to the GFD, independent of the manufacturer.
● Sprue ● Nahrungsmittelintoleranz ● Malabsorption ● anti-deaminierte Gliadin" " "
Peptide
● anti-Gewebs-Trans"
glutaminase 2
● Gluten-freie Diät "
Key words
● celiac disease ● food intolerance ● malabsorption ● anti-deamidated gliadin" " " "
peptide antibody
● anti-tissue trans"
glutaminase 2
● gluten-free diet (GFD) "
received accepted
30.1.2014 10.11.2014
Bibliography DOI http://dx.doi.org/ 10.1055/s-0034-1385704 Z Gastroenterol 2015; 53: 108–114 © Georg Thieme Verlag KG Stuttgart · New York · ISSN 0044-2771 Correspondence Prof. Dr. Philip Bufler Dr. von Hauner Children's Hospital, Ludwig Maximilians University Munich Lindwurmstr. 4 80337 München Germany Tel.: ++ 49/89/4 40 05 28 11 Fax: ++ 49/89/4 40 05 78 98 [email protected]
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Downloaded by: National University of Singapore. Copyrighted material.
Diagnostischer Wert von drei serologischen Tests bei Kindern mit Zöliakie
Originalarbeit
Introduction
with newly diagnosed CD who had antibody concentrations of > 10 times ULN for the respective tests.
!
Celiac disease (CD) is an immune-mediated enteropathy induced by the exposure to grains such as wheat, rye and barley containing gluten or related prolamins in genetically predisposed individuals [1]. Symptoms of CD are non-specific and patients may show few or no symptoms. Therefore, CD still remains under-diagnosed despite the availability of serological screening tools with high sensitivity [2]. Serological testing against autoantibodies (EMA-IgA, TG2-IgA) in conjunction with total serum IgA is the cornerstone in the primary diagnostic work-up of patients suspected to have CD [1]. Recently, IgA- and IgG-antibodies against deamidated, synthetic gliadin peptides (DGP) have been described as valuable diagnostic parameters in childhood celiac disease [3 – 5]. Specifically, DGP-IgG antibodies were shown to be similarly accurate biomarkers as TG2-IgA in CD [4, 6, 7]. The unequivocal advantage of DGP-IgG testing is to circumvent the diagnostic gap in IgA-deficient individuals with suspected CD [1, 8]. The new ESPGHAN guidelines on the diagnostic criteria for CD suggest that all tests for celiac specific antibodies should be validated against duodenal biopsies in at least 50 children with active CD and 100 control children if they are applied in children for clinical decision making [1]. In a child with symptoms of CD and very high titers of celiac specific autoantibodies [> 10 times upper limit of normal (ULN)] a diagnosis without biopsies may be considered if the patient fulfills all other required criteria [1]. CD is not only a disease of the gut but also affects extra-intestinal organs. Early diagnosis and dietary treatment with a gluten-free diet (GFD) is mandatory to prevent acute and chronic adverse consequences of the disease. As such adherence to diet is important to maintain remission of disease and to reduce long-term morbidity secondary to CD. There are only sparse data available to show which antibody test is more suitable than the other to monitor adherence to GFD in pediatric CD patients [9]. DGP-antibodies specifically bind the disease-inducing antigen and might be superior in monitoring patients on a GFD. The performance of each antibody test might differ between various manufacturers [7]. In this study we retrospectively analyzed the sera of children with histologically proven CD at diagnosis and during follow-up under a GFD by using four commercial assays for detection of TG2-IgA, DGP-IgG and DGP-IgA. In addition, we calculated the proportions of children
Table 1
109
Patients and Methods !
We retrospectively tested 411 sera of 91 pediatric patients with clinical symptoms of suspected CD (n = 81) or at risk for CD (n = 10) before (59/91) and after (91/91) onset of the GFD. Initial CD serology was performed by TG2-IgA and/or EMA tests from various manufacturers either at our institution (n = 76) or by the referring physician (n = 15). Patients who were transferred to our hospital for endoscopy after initial positive CD screening were serologically reevaluated at the time of upper endoscopy. At risk patients included those with Down’s syndrome (n = 3), Turner syndrome (n = 2) and diabetes mellitus type 1 (n = 5). IgA-deficient patients were excluded. Time points and frequency of serological testing during follow-up was not pre-specified. Usually, outpatient visits were arranged in 3 − 6 month intervals after diagnosis of CD until seroconversion. Subsequently, patients were seen once a year. Adherence to the GFD was discussed at each outpatient visit. One patient had no follow-up serum sample. Fifty eight sera of 12 patients who admitted to be non-adherent to the GFD, were analyzed separately and were excluded from the follow-up analysis. Accordingly, a total of 78 patients, of these 22 without pre-GFD serology, were included to the follow-up analysis. Control sera were obtained from IgA competent children who consumed a normal gluten containing diet and underwent upper endoscopy because of gastro-esophageal reflux disease, inflammatory bowel disease, or Helicobacter pylori infections, and had normal duodenal histology [Marsh 0) (n = 98)]. The retrospective measurement of blinded serum samples was approved by the institutional ethical committee.
Serologic testing All sera were stored at −20 °C until analysis and blinded prior to " Table 1. The analysis. The tests used for this study are listed in ● individual cut-off levels for each test were applied as recommended by the respective manufacturer. Test results above the lower level of the grey zone were declared positive. Serological analyses by ELISA-based tests of TheBindingSite (BS), Euroimmun (EU) and INOVA (IN) were performed by the same technician (GH) at
Cut-offs and statistical performance of serological tests 1.
TG2-IgA A
DGP-IgG B
C
D
A
DGP-IgA B
C
D
A
B
C
D
cutoff = positivity
≥4
≥ 20
≥ 20
≥7
≥ 10
≥ 25
≥ 20
≥7
≥ 10
≥ 25
≥ 20
≥7
titer > 10x ULN at diagnosis (%)
71
90
47
78
24
20
0
40
17
36
21
21
Sensitivity 1
100 (93.7 – 100)
100 (93.7 – 100)
100 (93.7 – 100)
100 (93.7 – 100)
89.5 (78.5 – 96)
98.2 (90.6 – 100)
93 (83 – 98.1)
100 (93.7 – 100)
66.7 (52.9 – 78.6)
86 (74.2 – 93.7)
84.2 (72.1 – 92.5)
78.9 (66.1 – 88.6)
Specificity 1
99 (94.4 – 100)
99 (94.4 – 100)
98 (92.8 – 99.9)
99 (94.4 – 100)
100 (96.3 – 100)
99 (94.4 – 100)
99 (94.4 – 100)
98 (92.8 – 99.8)
100 (96.3 – 100)
96.9 (91.3 – 99.4)
100 (96.3 – 100)
99 (94.4 – 100)
LR+
98
98
49
98
n. a.
96.3
91.1
49
n. a.
28
n. a.
77.4
LR-
0
0
0
0
0.11
0.02
0.07
0
0.33
0.15
0.16
0.21
A = TheBindingSite, B = Euroimmun, C = INOVA, D = Phadia, part of Thermo Fisher Scientific), LR = likelihood ratio, N. a. = not applicable. 1 Data are % (95 % confidence interval).
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Downloaded by: National University of Singapore. Copyrighted material.
Subjects and serum samples.
110
Originalarbeit
the author’s institution. Blinded and thawed, prediluted serum samples were sent to Phadia, part of Thermo Fisher Scientific (PH) for automated measurement of TG2-IgA, DGP-IgG and DGPIgA antibodies by FEIA.
Histology For histological evaluation at least 4 biopsies were taken from the second part of the duodenum and 2 biopsies were taken from the duodenal bulb (Jumbo forceps, Olympus FB-25 K, Hamburg). The biopsies were interpreted according to the modified Marsh classification [10]. Villous atrophy, crypt hyperplasia as well as the detection of intraepithelial lymphocytes (> 25/high power field) [11, 12] according to MARSH II or IIIa-c confirmed the diagnosis of CD according to ESPGHAN guidelines [13].
Fig. 1 TG2-IgA titers > 10 × ULN of children at diagnosis of CD (n = 59). Six patients showed TG2-IgA titers of < 10 × ULN at diagnosis in all tests. A: TheBindingSite, B: Euroimmun, C: INOVA, D: Phadia, part of Thermo Fisher Scientific.
Results are expressed as mean with standard deviation (SD) or median with interquartile range (IQR; 25th − 75th percentile) as appropriate. Age between cases and controls were compared by Kruskal-Wallis equality-of-populations rank test. Sensitivity, specificity, positive and negative likelihood ratios with exact binomial confidence intervals were calculated on the basis of the children’s true disease status defined by positive histological results and the ability of the serological test to define the disease status by the cut-off defined by the manufacturer. Serological test results in the follow-up were grouped into time intervals since diagnosis: 0 − 6, 7 − 12, 13 − 18 and > 18 months. A χ2 test was used to test for differences in the distribution of positive test results between companies and tests at different time periods after diagnosis. To compare the test characteristics of multiple serological assays we used Cochran's Q test. This statistical test considers the fact that the serological tests were applied in the same patients. Titers of antibodies at diagnosis of CD were calculated as times increase of ULN. These ratios were compared by a Wilcoxon matched-pairs signed-rank test.
Results !
Serum samples and patient characteristics A total of 411 sera of children with CD (n = 91, female:male 1.3:1) were available at the time of diagnosis (n = 59) and during followup under a GFD over a 94 months period [Jan. 2003 – Oct. 2009, median 22.0 (IQR 14.0 − 44.2)]. All control patients and 76/91 celiac patients underwent upper endoscopy at our unit. The remaining 15 celiac patients underwent upper endoscopy at the referring hospital. CD patients were younger than control patients (median age 5.6 years [IQR 3.1 – 8.9] vs. 12.4 years [IQR 7.4 – 15.6]; p < 0.001). There was a female predominance of 1.3:1 in the CD cohort and a male predominance of 1.3:1 in the control group.
Performance and titer-levels of TG2-IgA, DGP-IgG and DGP-IgA for diagnosis of CD
● Table 1 shows cut-off levels for the 12 different tests applied to "
this study which are identical with the value for the ULN. The given results are based on the 59 patients for whom serum was available at the time of diagnosis and the 98 control children. At diagnosis, titers of TG2-IgA, as calculated by times increase of ULN, showed significantly higher levels compared to DGP-IgG and DGP–IgA of each manufacturer (all p-values < 0.001). Depending on the applied serological test TG2-IgA concentrations
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Fig. 2 Celiac serology of children at diagnosis of CD. Venn diagram showing concordant positive test results of TG2-IgA, DGP-IgG and DGP-IgA as measured by different tests. A: TheBindingSite, B: Euroimmun, C: INOVA, D: Phadia, part of Thermo Fisher Scientific. One serum sample was not available for test C and D.
of > 10 × ULN ranged between 47 and 90 %. In 26/59 patients TG2-IgA concentrations were concordantly > 10 × ULN in all four " Fig. 1). Six patients had TG2-IgA titers of less than 10 × tests (● ULN at diagnosis in all tests. The proportions of children with values of DGP-IgG and DGP-IgA > 10 × ULN also varied between the " Table 1). four manufacturers (● Sensitivity of TG2-IgA and DGP-IgG to diagnose CD was 100 % and 89 – 100 %, respectively, for all applied tests. Sensitivity of DGPIgA was less (67 − 86 %) and showed significant differences " Table 1). (p < 0.001) between the four serological tests (● The specificity of TG2-IgA, DGP-IgG and DGP-IgA was > 96 % for " Table 1). One control patient with gastroall manufacturers (● esophageal reflux disease showed low positive titers of TG2-IgA in all 4 tests and positive DGP-IgG in 3/4 tests. A second control patient with Crohn’s disease was positive in only one TG2 assay (C), but negative in all DGP-IgG and DGP-IgA assays. " Fig. 2 illustrates that all CD patients at diThe Venn diagram of ● agnosis had positive TG2-IgA tests. Results for TG2-IgA, DGP-IgG and DGP-IgA were concordant positive in 39 − 51/59 patients (66 − 86 %) dependent on the applied serological test. 6 – 14/59 patients (11 − 24 %) were positive for TG2-IgA and DGP-IgG but neg" Fig. 2). ative for DGP-IgA (●
Downloaded by: National University of Singapore. Copyrighted material.
Statistical analysis
Fig. 3 Positivity of different tests for TG2-IgA, DGP-IgG and DGP-IgA at diagnosis of CD and after specified time periods under a GFD is defined as value above the individual borderline cut-off as indicated in " Fig. 5. The figure shows follow-up serology of 78 patients (only first measurement per interval) including 22 without pre-GFD serology in the follow-up analysis.
●
Patients non-adherent to diet are excluded. Number of patients total/without pre-GFD serum samples: time 0: 56/0; 1 − 6 months: 39/11; 7 − 12 months: 50/15; 13 − 18 months: 39/11; > 18 months 44/18. A: TheBindingSite, B: Euroimmun, C: INOVA, D: Phadia, part of Thermo Fisher Scientific. Graphs by assay with binomial 95 % confidence interval.
Correlation of TG2-IgA, DGP-IgG and DGP-IgA In order to analyze the reliability of tests provided by different manufacturers we calculated the correlation coefficient between all tests. The individual tests for TG2-IgA or DGP-IgG showed a strong correlation of ≥ 0.9. The correlation between DGP-IgA tests was lower but still high (0.84 – 0.92). Correlation coefficients between all test for TG2-IgA and DGP-IgG, TG2-IgA and DGP-IgA as well as DGP-IgG and DGP-IgA are similarly low at 0.5 – 0.6 " Fig. 5). (●
Analysis of celiac patients non-adherent to the glutenfree diet
Fig. 4 Celiac serology (TG2-IgA, DGP-IgG and DGP-IgA) as measured by tests applied to this study in six patients with CD nonadherent to the GFD. Vertical arrows indicate intake of gluten. Values are expressed as the product of upper limit of normal values (× ULN). A: TheBindingSite, B: Euroimmun, C: INOVA, D: Phadia, part of Thermo Fisher Scientific.
DGP-IgA might be a better indicator for non-adherence to the GFD since titers normalized most rapidly after the onset of the " Fig. 3. We therefore separately analyzed GFD in CD as shown in ● celiac antibodies in twelve patients with CD known to be non-ad" Fig. 4 for six patients, incidental herent to the GFD. As shown in● or voluntary ingestion of gluten during an otherwise GFD induced low positive titers of DGP-IgA and DGP-IgG in CD patients. In only one patient was there a high titer rise of DGP-IgG and DGP-IgA tests of two manufactures. TG2-IgA, however, showed a rapid and high titer rise for all tests after deliberate gluten intake by CD patients.
Time course of antibodies under the gluten-free diet One to six months after starting a GFD 46 to 71 % of patients with pre-GFD serological results had still positive TG2-IgA or DGP-IgG titers. In contrast, only 7 to 14 % of patients were still positive for DGP-IgA within the same time frame. After 6 months > 91 % of patients had negative DGP-IgA results for all assays. The time to normalize TG2-IgA is longer compared to DGP-IgG and DGP-IgA (p < 0.001 for each difference at 7 − 12, 13 − 18 months and > 18 months). After 18 months under a GFD, 27 to 31 % of patients are still positive for TG2-IgA in 3 tests and 15 % were positive in one TG2-IgA test. Titers of DGP-IgG and DGP-IgA were within the normal range after 18 months of a GFD in > 85 % and > 92 % of patients with pre-GFD sera. Comparable results were obtained when sera of patients without pre-GFD serology (n = 22, total " Fig. 3). n = 78) were included to the analysis (●
Discussion !
In this study we compared the performance of DGP-IgG and DGPIgA with TG2-IgA tests of four manufacturers in a pediatric cohort of 59 active celiac and 98 control children. Duodenal biopsies were used as reference standard. Furthermore, we assessed the rate of positive TG2-IgA, DGP-IgG and DGP-IgA at different time points under GFD in 91 celiac children who had presented for follow-up to our department. We confirmed the high specificity of all tests, and a reduced sensitivity of DGP-IgA, compared to DGPIgG and TG2-IgA to diagnose CD [7]. Major differences were found between the 4 manufacturers with respect to the proportions of TG2-IgA values of > 10 × ULN. Under a GFD DGP-IgA normalized more rapidly than DGP-IgG and TG2-IgA. TG2-IgA showed the
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
111
Downloaded by: National University of Singapore. Copyrighted material.
Originalarbeit
Originalarbeit
Fig. 5 Correlation coefficient of all tests for serum measurement of TG2-IgA, DGP-IgG and DGP-IgA (p-value was calculated according to Spearman). BS: TheBindingSite, EU: Euroimmun, IN: INOVA, PH: Phadia, part of Thermo Fisher Scientific.
highest titer rise with non-adherence to the GFD, independent of the manufacturer. Current guidelines recommend the measurement of TG2-IgA and total IgA for initial screening of CD [1]. Various commercial providers and laboratories are offering combinations of TG2-IgA, DGPIgA, DGP-IgG, and EMA-IgA for serological screening of CD in symptomatic or at risk patients, although the costs of such multiple testing are high and the benefit of such panels has not been established. Specific antibody tests from different providers have been validated against histology but not evaluated against each other, and it is unclear whether different commercial tests perform similarly [7]. In addition, it is not sufficient to test for antibody positivity or negativity and only the most recent studies considered the antibody concentrations [1]. Hence, we compared quantitative results of three different tests of four manufacturers for the detection of celiac specific antibodies. In our cohort of IgA competent children, TG2-IgA testing detected all celiac patients with active disease, regardless of the manufacturer, while DGP-IgG testing alone or the combination of DGPIgG and DPG-IgA missed between 1 – 6 celiac patients, except for manufacturer D. Similar results have been described for other cohorts of adults and children [5, 8, 14, 15]. As previously shown, IgA antibodies against DGP showed a significantly lower sensitivity than TG2-IgA and DGP-IgG [5, 7, 15]. The excellent results for TG2-IgA compared to DGP-testing could be in part biased by the fact that the majority (53/59) of our active CD patients were initially tested with a TG2-IgA assay (A) at our institution. Of note, results of TG2-IgA measurements by three additional tests (B, C, D) showed a high correlation to our screening assay (A). With respect to specificity all 12 tests had an excellent performance. One control patient with Marsh 0 on histology was low titer positive in all performed test. Celiac specific antibodies can be induced by viral infections or other unknown environmental factors, and may show fluctuating positive levels in patients at genetic risk for CD [16]. Therefore we cannot exclude true CD in this patient with false negative histology due to patchy disease. Overall titers of TG2-IgA were significantly higher than titers of DGP-IgG or DGP-IgA as calculated by fold increase of ULN. Ac-
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
cording to the recent ESPGHAN guidelines TG2-IgA titers of > 10 × ULN, positive endomysial antibodies in a second blood sample, HLA-DQ2/8 positivity and clinical signs suggestive for CD are sufficient to establish the diagnosis of CD [1]. There were major differences between the four TG2-IgA tests regarding titer levels of > 10 × ULN. Depending on the test 47 – 90 % of our patients would be considered for further non-invasive testing to establish the diagnosis without biopsies. A high proportion of CD patients with TG2-IgA > 10 × ULN (90 − 91 %) was also reported by Vermeersch et al. in a mixed population of children and adults by using 4 different commercial assays [17]. The variability of values > 10 × ULN between single TG2-IgA tests points to the need of standardization between assays as pointed out by Egner et al. [18]. The cut-off of each test needs to be evaluated according to the ESPGHAN guidelines to circumvent intestinal biopsies. Since we did not test for HLA-DQ2/8 positivity or endomysial antibodies in all patients, we cannot extrapolate the proportion of patients to meet the current ESPGHAN recommendations for diagnosis of CD. This was not the purpose of this retrospective study. " Fig. 2 illustrates the different performThe Venn diagram of ● ances of the four manufacturers. The measurement of either DGP-IgG or DGP-IgA did not add any value to TG2-IgA results. These results are in concordance with a previous report [8] and underline the fact that the measurement of DGP-IgG or -IgA in combination with TG2-IgA does not improve celiac serology. DGP-antibodies bind the disease-inducing antigen and might be superior to monitor patients on a GFD. Liu and colleagues reported that titers of TG2-IgA and DGP antibodies (combined measurement of DGP-IgG and -IgA) paralleled in a cohort of 50 CD patients at diagnosis but DGP antibodies resolved sooner than TG2 IgA [19]. Consistent with this report and the study by Basso et al. [20] we found a significant drop of TG2-IgA and DGP-antibodies under GFD. DGP-IgA normalized most rapidly and already showed negative results in > 85 % of patients within the first 6 months of a GFD. Titers of DGP-IgA were similarly high to DGPIgG at diagnosis. Therefore, differences of the initial titer height cannot explain the more rapid normalization of DGP-IgA in comparison to DGP-IgG. In contrast, the late normalization of TG2-IgA in comparison to DGP-antibodies might be explained by initially
Downloaded by: National University of Singapore. Copyrighted material.
112
Originalarbeit
Abbreviations CD: DGP: DGP-IgA: DGP-IgG: EMA: GFD: TG2-IgA:
celiac disease deamidated gliadin peptide IgA antibodies against DGP IgG antibodies against DGP autoantibodies against endomysium gluten-free diet IgA antibodies against tissue transglutaminase
Acknowledgements !
The authors thank all staff physicians and nurses of the Division of Pediatric Gastroenterology, Hepatology and Nutrition for excellent patient care and support of the study. Conflict of interest: All tests were supplied free of charge by the manufacturing companies. The companies equally covered costs for laboratory supplies and the technician. The companies had no influence on study design or analysis of data. PB received lecture fees from MSD, GivenImaging and Abbvie. SK is advisory board member of Centocor, MSD, Danone, Merck, Vifor, Nestle Nutrition Institute, Abbvie and received lecture fees from MSD, Danone, Merck, Vifor, Nestle Nutrition Institute, Euroimmun, Thermo-Fisher (Phadia), Abbvie, Schär, Hipp and Falk.
References 01 Husby S, Koletzko S, Korponay-Szabo IR et al. European Society for Pediatric Gastroenterology, Hepatology, and Nutrition guidelines for the diagnosis of coeliac disease. J Pediatr Gastroenterol Nutr 2012; 54: 136 – 160 02 Leffler DA, Schuppan D. Update on serologic testing in celiac disease. Am J Gastroenterol 2010; 105: 2520 – 2524 03 Prince HE. Evaluation of the INOVA diagnostics enzyme-linked immunosorbent assay kits for measuring serum immunoglobulin G (IgG) and IgA to deamidated gliadin peptides. Clin Vaccine Immunol 2006; 13: 150 – 151 04 Sugai E, Vazquez H, Nachman F et al. Accuracy of testing for antibodies to synthetic gliadin-related peptides in celiac disease. Clin Gastroenterol Hepatol 2006; 4: 1112 – 1117 05 Prause C, Ritter M, Probst C et al. Antibodies against deamidated gliadin as new and accurate biomarkers of childhood coeliac disease. J Pediatr Gastroenterol Nutr 2009; 49: 52 – 58 06 Niveloni S, Sugai E, Cabanne A et al. Antibodies against synthetic deamidated gliadin peptides as predictors of celiac disease: prospective assessment in an adult population with a high pretest probability of disease. Clin Chem 2007; 53: 2186 – 2192 07 Giersiepen K, Lelgemann M, Stuhldreher N et al. Accuracy of diagnostic antibody tests for coeliac disease in children: summary of an evidence report. J Pediatr Gastroenterol Nutr 2012; 54: 229 – 241 08 Olen O, Gudjonsdottir AH, Browaldh L et al. Antibodies against deamidated gliadin peptides and tissue transglutaminase for diagnosis of pediatric celiac disease – diagnostic performance and cost in clinical practice. J Pediatr Gastroenterol Nutr 2012; 55: 695 – 700 09 Monzani A, Rapa A, Fonio P et al. Use of deamidated gliadin peptide antibodies to monitor diet compliance in childhood celiac disease. J Pediatr Gastroenterol Nutr 2011; 53: 55 – 60 10 Oberhuber G, Granditsch G, Vogelsang H. The histopathology of coeliac disease: time for a standardized report scheme for pathologists. Eur J Gastroenterol Hepatol 1999; 11: 1185 – 1194 11 Pellegrino S, Villanacci V, Sansotta N et al. Redefining the intraepithelial lymphocytes threshold to diagnose gluten sensitivity in patients with architecturally normal duodenal histology. Aliment Pharmacol Ther 2011; 33: 697 – 706 12 Walker MM, Murray JA, Ronkainen J et al. Detection of celiac disease and lymphocytic enteropathy by parallel serology and histopathology in a population-based study. Gastroenterology 2010; 139: 112 – 119 13 Revised criteria for diagnosis of coeliac disease. Report of Working Group of European Society of Paediatric Gastroenterology and Nutrition. Arch Dis Child 1990; 65: 909 – 911
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Downloaded by: National University of Singapore. Copyrighted material.
higher titers of TG2-IgA. A decrease of celiac specific antibodies was associated with better adherence to the GFD in adults and the combined measurement of DGP-IgA and TG2-IgA was best for follow-up [21]. Even if titers of DGP-IgA most rapidly normalized under the GFD we found that incidental or voluntary ingestion of gluten was best indicated by rising titers of TG2-IgA. All four TG2-IgA assays performed similarly, but test B showed the highest titer rise upon unforeseen gluten intake. The high titers in test B are also reflected in the high percentage of children with TG2 IgA > 10 × ULN by this assay. Under a GFD, we only found a low and transient titer rise of either TG2-IgA, DGP-IgG or DGP-IgA or combinations of tests in 7/91 patients. From gluten challenge studies it is well known that there is an individual response both in titer height and time interval to react with CD specific antibodies [22, 23]. Intermittent intake of small amounts of gluten is most likely not detected by serology, even if it shows adverse effects on the architecture of the mucosa [23]. Therefore, a TG2-IgA test with a high titer response will be advantageous to monitor compliance to GFD in patients with CD. Due to our study design we cannot calculate the sensitivity and specificity of DGP and TG2 antibodies to monitor adherence to the GFD as was done by other groups in children [9, 20, 24]. A recent prospective study showed that the negative likelihood ratio of EMA was best to rule out persistent mucosal injury during GFD in children with CD in comparison to TG2-IgA and DGP-antibodies [24]. In some contrast, Basso et al. reported in children that the sensitivity of TG2-IgA to detect non-adherence to the GFD was only 24 % and was exceeded by DGP-IgG and IgA (76 % and 60 %) [20]. In accordance, two studies in adults described that normal TG2-IgA and EMA were unable to predict recovery of villous atrophy in adult patients with CD under GFD [25, 26]. We found an excellent quantitative correlation of the individual antibody tests applied to this study. This is an important finding since there are no widely accepted gold standards for specific CD antibody assays [27]. However, the correlation between TG2-IgA, DGP-IgG and DGP-IgA titers was weak and confirmed a study by Liu [19]. Limitations of this study are the retrospective design and the screening of patients with CD by TG2-IgA of a single company (A) in the majority of patients. Therefore we can only indicate the sensitivity and specificity of the different serological tests for diagnosis of CD. However, the fact that all TG2-IgA assays showed a similar sensitivity, points to the diagnostic strength of TG2-IgA as the antibodies directed against the auto-antigen in CD. With the writing of this manuscript one test (A) is off market now. In summary, we show that TG2-IgA and DGP-IgG tests provided by four manufacturers are superior to DGP-IgA for initial diagnostic testing of children with suspected celiac disease. The rate of TG2-IgA titers > 10 × ULN at diagnosis differed significantly between the four TG2-IgA tests indicating that each test needs evaluation whether it fits the current ESPGHAN guidelines to circumvent intestinal biopsies. Non-adherence to a GFD was best indicated by a high titer rise of TG2-IgA.
113
Originalarbeit
14 Rashtak S, Ettore MW, Homburger HA et al. Comparative usefulness of deamidated gliadin antibodies in the diagnosis of celiac disease. Clin Gastroenterol Hepatol 2008; 6: 426 – 432; quiz 370 15 Vermeersch P, Geboes K, Marien G et al. Diagnostic performance of IgG anti-deamidated gliadin peptide antibody assays is comparable to IgA anti-tTG in celiac disease. Clin Chim Acta 2010; 411: 931 – 935 16 Simell S, Kupila A, Hoppu S et al. Natural history of transglutaminase autoantibodies and mucosal changes in children carrying HLA-conferred celiac disease susceptibility. Scand J Gastroenterol 2005; 40: 1182 – 1191 17 Vermeersch P, Geboes K, Marien G et al. Defining thresholds of antibody levels improves diagnosis of celiac disease. Clin Gastroenterol Hepatol 2013; 11: 398 – 403 18 Egner W, Shrimpton A, Sargur R et al. ESPGHAN guidance on coeliac disease 2012: multiples of ULN for decision making do not harmonise assay performance across centres. J Pediatr Gastroenterol Nutr 2012; 55: 733 – 735 19 Liu E, Li M, Emery L et al. Natural history of antibodies to deamidated gliadin peptides and transglutaminase in early childhood celiac disease. J Pediatr Gastroenterol Nutr 2007; 45: 293 – 300 20 Basso D, Guariso G, Fogar P et al. Antibodies against synthetic deamidated gliadin peptides for celiac disease diagnosis and follow-up in children. Clin Chem 2009; 55: 150 – 157
21 Nachman F, Sugai E, Vazquez H et al. Serological tests for celiac disease as indicators of long-term compliance with the gluten-free diet. Eur J Gastroenterol Hepatol 2011; 23: 473 – 480 22 Lahdeaho ML, Maki M, Laurila K et al. Small-bowel mucosal changes and antibody responses after low- and moderate-dose gluten challenge in celiac disease. BMC Gastroenterology 2011; 11: 129 23 Leffler D, Schuppan D, Pallav K et al. Kinetics of the histological, serological and symptomatic responses to gluten challenge in adults with coeliac disease. Gut 2012; DOI: 10.1136/gutjnl-2012-302196 24 Vecsei E, Steinwendner S, Kogler H et al. Follow-up of pediatric celiac disease: value of antibodies in predicting mucosal healing, a prospective cohort study. BMC Gastroenterology 2014; 14: 28 25 Hopper AD, Hadjivassiliou M, Hurlstone DP et al. What is the role of serologic testing in celiac disease? A prospective, biopsy-confirmed study with economic analysis. Clin Gastroenterol Hepatol 2008; 6: 314 – 320 26 Sharkey LM, Corbett G, Currie E et al. Optimising delivery of care in coeliac disease – comparison of the benefits of repeat biopsy and serological follow-up. Aliment Pharmacol Ther 2013; 38: 1278 – 1291 27 Li M, Yu L, Tiberti C et al. A report on the International Transglutaminase Autoantibody Workshop for Celiac Disease. Am J Gastroenterol 2009; 104: 154 – 163
Downloaded by: National University of Singapore. Copyrighted material.
114
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Originalarbeit
Diagnostic performance of three serologic tests in childhood celiac disease
Authors
P. Bufler, G. Heilig, G. Ossiander, F. Freudenberg, V. Grote, S. Koletzko
Affiliation
Dr. von Hauner Children's Hospital, Ludwig Maximilians University Munich, München
Schlüsselwörter
Zusammenfassung
Abstract
"
!
!
Hintergrund: IgA- und IgG-Antikörper gegen deaminierte Gliadin-Peptide (DGP) binden das krankheitsauslösende Antigen und könnten deshalb bei Patienten mit Zöliakie besser zur Verlaufskontrolle unter Gluten-freier Diät (GFD) geeignet sein als Gewebs-Transglutaminase 2 (TG2)-IgA. Ziel dieser Studie war der Vergleich von DGP-IgG und DGPIgA mit TG2-IgA 4 verschiedener Hersteller zur Diagnostik und Verlaufskontrolle von Kindern mit Zöliakie. Patienten und Methodik: Es wurden 411 Seren von 91 IgA-kompetenten Kindern mit Biopsie-gesicherter Zöliakie zum Zeitpunkt der Diagnosestellung sowie unter GFD untersucht und mit Seren von 98 Kontrollpatienten verglichen. Die Tests (TheBindingSite, Euroimmun, Phadia/Thermo Fisher Scientific, INOVA) zur Detektion von TG2-IgA, DGP-IgG und DGP-IgA wurden entsprechend der Herstellerangaben durchgeführt. Ergebnisse: TG2-IgA und DGP-IgG hatten eine signifikant höhere Sensitivität (100 % und 90 − 100 %) für die Diagnose Zöliakie als DGP-IgA (67 − 86 %). Alle Tests hatten eine vergleichbar hohe Spezifität (97 – 100 %). Die Häufigkeit von TG2-IgA > 10 × des oberen Normwerts (ULN) zum Zeitpunkt der Diagnosestellung lag zwischen 47 und 90 %. Unter GFD normalisierte sich DGP-IgA früher als DGPIgG und TG2-IgA. Bei Diätfehlern hatten TG2-IgA die deutlichsten Titeranstiege. Schlussfolgerung: Die kombinierte Messung von TG2-IgA und DGP-IgG hat im Vergleich zur alleinigen Bestimmung von TG2-IgA bei IgA-kompetenten Kindern mit Zöliakie keinen diagnostischen Vorteil. Die hohe Variabilität der TG2-IgA > 10 × ULN zeigt, dass eine Harmonisierung der kommerziellen Tests anzustreben ist. Diätfehler unter GFD werden am besten durch einen Anstieg der TG2-IgA diagnostiziert.
Background: IgA- and IgG-antibodies against deamidated gliadin peptides (DGP) specifically bind the disease-inducing antigen and might be superior to transglutaminase type 2 (TG2) IgA in monitoring patients on a gluten-free diet (GFD). The aim of this study was to compare the performance of DGP-IgG and DGP-IgA with TG2-IgA of four manufacturers in pediatric celiac patients at diagnosis and during follow-up under a GFD. Patients and Methods: In total 411 sera of 91 IgA competent children with biopsy proven celiac disease were analyzed at diagnosis and during follow-up on a GFD. Ninety-eight children with normal duodenal histology served as controls. The tests (TheBindingSite, Euroimmun, Phadia, part of Thermo Fisher Scientific, INOVA) for detection of TG2-IgA, DGP-IgG and DGP-IgA were used according to the manufacturers’ instructions. Results: Sensitivity to diagnose CD was high for TG2-IgA (100 %) and DGP-IgG (90 − 100 %), but lower for DGP-IgA (67 − 86 %). Specificity was high for all tests (97 – 100 %). The frequency of TG2-IgA titers > 10 × upper limit of normal at diagnosis ranged from 47 − 90 %. Under a GFD DGPIgA became negative more rapidly than DGP-IgG and TG2-IgA. Non-adherence to GFD was best indicated by positive TG2-IgA. Conclusions: Combined testing for TG2-IgA and DGP-IgG does not increase the detection rate of CD in IgA competent children compared to TG2IgA only. There are significant differences with respect to proportions of celiac children with titers > 10 × ULN between the manufacturers. This calls for harmonization of tests. TG2-IgA showed the highest titer rise with non-adherence to the GFD, independent of the manufacturer.
● Sprue ● Nahrungsmittelintoleranz ● Malabsorption ● anti-deaminierte Gliadin" " "
Peptide
● anti-Gewebs-Trans"
glutaminase 2
● Gluten-freie Diät "
Key words
● celiac disease ● food intolerance ● malabsorption ● anti-deamidated gliadin" " " "
peptide antibody
● anti-tissue trans"
glutaminase 2
● gluten-free diet (GFD) "
received accepted
30.1.2014 10.11.2014
Bibliography DOI http://dx.doi.org/ 10.1055/s-0034-1385704 Z Gastroenterol 2015; 53: 108–114 © Georg Thieme Verlag KG Stuttgart · New York · ISSN 0044-2771 Correspondence Prof. Dr. Philip Bufler Dr. von Hauner Children's Hospital, Ludwig Maximilians University Munich Lindwurmstr. 4 80337 München Germany Tel.: ++ 49/89/4 40 05 28 11 Fax: ++ 49/89/4 40 05 78 98 [email protected]
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Downloaded by: National University of Singapore. Copyrighted material.
Diagnostischer Wert von drei serologischen Tests bei Kindern mit Zöliakie
Originalarbeit
Introduction
with newly diagnosed CD who had antibody concentrations of > 10 times ULN for the respective tests.
!
Celiac disease (CD) is an immune-mediated enteropathy induced by the exposure to grains such as wheat, rye and barley containing gluten or related prolamins in genetically predisposed individuals [1]. Symptoms of CD are non-specific and patients may show few or no symptoms. Therefore, CD still remains under-diagnosed despite the availability of serological screening tools with high sensitivity [2]. Serological testing against autoantibodies (EMA-IgA, TG2-IgA) in conjunction with total serum IgA is the cornerstone in the primary diagnostic work-up of patients suspected to have CD [1]. Recently, IgA- and IgG-antibodies against deamidated, synthetic gliadin peptides (DGP) have been described as valuable diagnostic parameters in childhood celiac disease [3 – 5]. Specifically, DGP-IgG antibodies were shown to be similarly accurate biomarkers as TG2-IgA in CD [4, 6, 7]. The unequivocal advantage of DGP-IgG testing is to circumvent the diagnostic gap in IgA-deficient individuals with suspected CD [1, 8]. The new ESPGHAN guidelines on the diagnostic criteria for CD suggest that all tests for celiac specific antibodies should be validated against duodenal biopsies in at least 50 children with active CD and 100 control children if they are applied in children for clinical decision making [1]. In a child with symptoms of CD and very high titers of celiac specific autoantibodies [> 10 times upper limit of normal (ULN)] a diagnosis without biopsies may be considered if the patient fulfills all other required criteria [1]. CD is not only a disease of the gut but also affects extra-intestinal organs. Early diagnosis and dietary treatment with a gluten-free diet (GFD) is mandatory to prevent acute and chronic adverse consequences of the disease. As such adherence to diet is important to maintain remission of disease and to reduce long-term morbidity secondary to CD. There are only sparse data available to show which antibody test is more suitable than the other to monitor adherence to GFD in pediatric CD patients [9]. DGP-antibodies specifically bind the disease-inducing antigen and might be superior in monitoring patients on a GFD. The performance of each antibody test might differ between various manufacturers [7]. In this study we retrospectively analyzed the sera of children with histologically proven CD at diagnosis and during follow-up under a GFD by using four commercial assays for detection of TG2-IgA, DGP-IgG and DGP-IgA. In addition, we calculated the proportions of children
Table 1
109
Patients and Methods !
We retrospectively tested 411 sera of 91 pediatric patients with clinical symptoms of suspected CD (n = 81) or at risk for CD (n = 10) before (59/91) and after (91/91) onset of the GFD. Initial CD serology was performed by TG2-IgA and/or EMA tests from various manufacturers either at our institution (n = 76) or by the referring physician (n = 15). Patients who were transferred to our hospital for endoscopy after initial positive CD screening were serologically reevaluated at the time of upper endoscopy. At risk patients included those with Down’s syndrome (n = 3), Turner syndrome (n = 2) and diabetes mellitus type 1 (n = 5). IgA-deficient patients were excluded. Time points and frequency of serological testing during follow-up was not pre-specified. Usually, outpatient visits were arranged in 3 − 6 month intervals after diagnosis of CD until seroconversion. Subsequently, patients were seen once a year. Adherence to the GFD was discussed at each outpatient visit. One patient had no follow-up serum sample. Fifty eight sera of 12 patients who admitted to be non-adherent to the GFD, were analyzed separately and were excluded from the follow-up analysis. Accordingly, a total of 78 patients, of these 22 without pre-GFD serology, were included to the follow-up analysis. Control sera were obtained from IgA competent children who consumed a normal gluten containing diet and underwent upper endoscopy because of gastro-esophageal reflux disease, inflammatory bowel disease, or Helicobacter pylori infections, and had normal duodenal histology [Marsh 0) (n = 98)]. The retrospective measurement of blinded serum samples was approved by the institutional ethical committee.
Serologic testing All sera were stored at −20 °C until analysis and blinded prior to " Table 1. The analysis. The tests used for this study are listed in ● individual cut-off levels for each test were applied as recommended by the respective manufacturer. Test results above the lower level of the grey zone were declared positive. Serological analyses by ELISA-based tests of TheBindingSite (BS), Euroimmun (EU) and INOVA (IN) were performed by the same technician (GH) at
Cut-offs and statistical performance of serological tests 1.
TG2-IgA A
DGP-IgG B
C
D
A
DGP-IgA B
C
D
A
B
C
D
cutoff = positivity
≥4
≥ 20
≥ 20
≥7
≥ 10
≥ 25
≥ 20
≥7
≥ 10
≥ 25
≥ 20
≥7
titer > 10x ULN at diagnosis (%)
71
90
47
78
24
20
0
40
17
36
21
21
Sensitivity 1
100 (93.7 – 100)
100 (93.7 – 100)
100 (93.7 – 100)
100 (93.7 – 100)
89.5 (78.5 – 96)
98.2 (90.6 – 100)
93 (83 – 98.1)
100 (93.7 – 100)
66.7 (52.9 – 78.6)
86 (74.2 – 93.7)
84.2 (72.1 – 92.5)
78.9 (66.1 – 88.6)
Specificity 1
99 (94.4 – 100)
99 (94.4 – 100)
98 (92.8 – 99.9)
99 (94.4 – 100)
100 (96.3 – 100)
99 (94.4 – 100)
99 (94.4 – 100)
98 (92.8 – 99.8)
100 (96.3 – 100)
96.9 (91.3 – 99.4)
100 (96.3 – 100)
99 (94.4 – 100)
LR+
98
98
49
98
n. a.
96.3
91.1
49
n. a.
28
n. a.
77.4
LR-
0
0
0
0
0.11
0.02
0.07
0
0.33
0.15
0.16
0.21
A = TheBindingSite, B = Euroimmun, C = INOVA, D = Phadia, part of Thermo Fisher Scientific), LR = likelihood ratio, N. a. = not applicable. 1 Data are % (95 % confidence interval).
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Downloaded by: National University of Singapore. Copyrighted material.
Subjects and serum samples.
110
Originalarbeit
the author’s institution. Blinded and thawed, prediluted serum samples were sent to Phadia, part of Thermo Fisher Scientific (PH) for automated measurement of TG2-IgA, DGP-IgG and DGPIgA antibodies by FEIA.
Histology For histological evaluation at least 4 biopsies were taken from the second part of the duodenum and 2 biopsies were taken from the duodenal bulb (Jumbo forceps, Olympus FB-25 K, Hamburg). The biopsies were interpreted according to the modified Marsh classification [10]. Villous atrophy, crypt hyperplasia as well as the detection of intraepithelial lymphocytes (> 25/high power field) [11, 12] according to MARSH II or IIIa-c confirmed the diagnosis of CD according to ESPGHAN guidelines [13].
Fig. 1 TG2-IgA titers > 10 × ULN of children at diagnosis of CD (n = 59). Six patients showed TG2-IgA titers of < 10 × ULN at diagnosis in all tests. A: TheBindingSite, B: Euroimmun, C: INOVA, D: Phadia, part of Thermo Fisher Scientific.
Results are expressed as mean with standard deviation (SD) or median with interquartile range (IQR; 25th − 75th percentile) as appropriate. Age between cases and controls were compared by Kruskal-Wallis equality-of-populations rank test. Sensitivity, specificity, positive and negative likelihood ratios with exact binomial confidence intervals were calculated on the basis of the children’s true disease status defined by positive histological results and the ability of the serological test to define the disease status by the cut-off defined by the manufacturer. Serological test results in the follow-up were grouped into time intervals since diagnosis: 0 − 6, 7 − 12, 13 − 18 and > 18 months. A χ2 test was used to test for differences in the distribution of positive test results between companies and tests at different time periods after diagnosis. To compare the test characteristics of multiple serological assays we used Cochran's Q test. This statistical test considers the fact that the serological tests were applied in the same patients. Titers of antibodies at diagnosis of CD were calculated as times increase of ULN. These ratios were compared by a Wilcoxon matched-pairs signed-rank test.
Results !
Serum samples and patient characteristics A total of 411 sera of children with CD (n = 91, female:male 1.3:1) were available at the time of diagnosis (n = 59) and during followup under a GFD over a 94 months period [Jan. 2003 – Oct. 2009, median 22.0 (IQR 14.0 − 44.2)]. All control patients and 76/91 celiac patients underwent upper endoscopy at our unit. The remaining 15 celiac patients underwent upper endoscopy at the referring hospital. CD patients were younger than control patients (median age 5.6 years [IQR 3.1 – 8.9] vs. 12.4 years [IQR 7.4 – 15.6]; p < 0.001). There was a female predominance of 1.3:1 in the CD cohort and a male predominance of 1.3:1 in the control group.
Performance and titer-levels of TG2-IgA, DGP-IgG and DGP-IgA for diagnosis of CD
● Table 1 shows cut-off levels for the 12 different tests applied to "
this study which are identical with the value for the ULN. The given results are based on the 59 patients for whom serum was available at the time of diagnosis and the 98 control children. At diagnosis, titers of TG2-IgA, as calculated by times increase of ULN, showed significantly higher levels compared to DGP-IgG and DGP–IgA of each manufacturer (all p-values < 0.001). Depending on the applied serological test TG2-IgA concentrations
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Fig. 2 Celiac serology of children at diagnosis of CD. Venn diagram showing concordant positive test results of TG2-IgA, DGP-IgG and DGP-IgA as measured by different tests. A: TheBindingSite, B: Euroimmun, C: INOVA, D: Phadia, part of Thermo Fisher Scientific. One serum sample was not available for test C and D.
of > 10 × ULN ranged between 47 and 90 %. In 26/59 patients TG2-IgA concentrations were concordantly > 10 × ULN in all four " Fig. 1). Six patients had TG2-IgA titers of less than 10 × tests (● ULN at diagnosis in all tests. The proportions of children with values of DGP-IgG and DGP-IgA > 10 × ULN also varied between the " Table 1). four manufacturers (● Sensitivity of TG2-IgA and DGP-IgG to diagnose CD was 100 % and 89 – 100 %, respectively, for all applied tests. Sensitivity of DGPIgA was less (67 − 86 %) and showed significant differences " Table 1). (p < 0.001) between the four serological tests (● The specificity of TG2-IgA, DGP-IgG and DGP-IgA was > 96 % for " Table 1). One control patient with gastroall manufacturers (● esophageal reflux disease showed low positive titers of TG2-IgA in all 4 tests and positive DGP-IgG in 3/4 tests. A second control patient with Crohn’s disease was positive in only one TG2 assay (C), but negative in all DGP-IgG and DGP-IgA assays. " Fig. 2 illustrates that all CD patients at diThe Venn diagram of ● agnosis had positive TG2-IgA tests. Results for TG2-IgA, DGP-IgG and DGP-IgA were concordant positive in 39 − 51/59 patients (66 − 86 %) dependent on the applied serological test. 6 – 14/59 patients (11 − 24 %) were positive for TG2-IgA and DGP-IgG but neg" Fig. 2). ative for DGP-IgA (●
Downloaded by: National University of Singapore. Copyrighted material.
Statistical analysis
Fig. 3 Positivity of different tests for TG2-IgA, DGP-IgG and DGP-IgA at diagnosis of CD and after specified time periods under a GFD is defined as value above the individual borderline cut-off as indicated in " Fig. 5. The figure shows follow-up serology of 78 patients (only first measurement per interval) including 22 without pre-GFD serology in the follow-up analysis.
●
Patients non-adherent to diet are excluded. Number of patients total/without pre-GFD serum samples: time 0: 56/0; 1 − 6 months: 39/11; 7 − 12 months: 50/15; 13 − 18 months: 39/11; > 18 months 44/18. A: TheBindingSite, B: Euroimmun, C: INOVA, D: Phadia, part of Thermo Fisher Scientific. Graphs by assay with binomial 95 % confidence interval.
Correlation of TG2-IgA, DGP-IgG and DGP-IgA In order to analyze the reliability of tests provided by different manufacturers we calculated the correlation coefficient between all tests. The individual tests for TG2-IgA or DGP-IgG showed a strong correlation of ≥ 0.9. The correlation between DGP-IgA tests was lower but still high (0.84 – 0.92). Correlation coefficients between all test for TG2-IgA and DGP-IgG, TG2-IgA and DGP-IgA as well as DGP-IgG and DGP-IgA are similarly low at 0.5 – 0.6 " Fig. 5). (●
Analysis of celiac patients non-adherent to the glutenfree diet
Fig. 4 Celiac serology (TG2-IgA, DGP-IgG and DGP-IgA) as measured by tests applied to this study in six patients with CD nonadherent to the GFD. Vertical arrows indicate intake of gluten. Values are expressed as the product of upper limit of normal values (× ULN). A: TheBindingSite, B: Euroimmun, C: INOVA, D: Phadia, part of Thermo Fisher Scientific.
DGP-IgA might be a better indicator for non-adherence to the GFD since titers normalized most rapidly after the onset of the " Fig. 3. We therefore separately analyzed GFD in CD as shown in ● celiac antibodies in twelve patients with CD known to be non-ad" Fig. 4 for six patients, incidental herent to the GFD. As shown in● or voluntary ingestion of gluten during an otherwise GFD induced low positive titers of DGP-IgA and DGP-IgG in CD patients. In only one patient was there a high titer rise of DGP-IgG and DGP-IgA tests of two manufactures. TG2-IgA, however, showed a rapid and high titer rise for all tests after deliberate gluten intake by CD patients.
Time course of antibodies under the gluten-free diet One to six months after starting a GFD 46 to 71 % of patients with pre-GFD serological results had still positive TG2-IgA or DGP-IgG titers. In contrast, only 7 to 14 % of patients were still positive for DGP-IgA within the same time frame. After 6 months > 91 % of patients had negative DGP-IgA results for all assays. The time to normalize TG2-IgA is longer compared to DGP-IgG and DGP-IgA (p < 0.001 for each difference at 7 − 12, 13 − 18 months and > 18 months). After 18 months under a GFD, 27 to 31 % of patients are still positive for TG2-IgA in 3 tests and 15 % were positive in one TG2-IgA test. Titers of DGP-IgG and DGP-IgA were within the normal range after 18 months of a GFD in > 85 % and > 92 % of patients with pre-GFD sera. Comparable results were obtained when sera of patients without pre-GFD serology (n = 22, total " Fig. 3). n = 78) were included to the analysis (●
Discussion !
In this study we compared the performance of DGP-IgG and DGPIgA with TG2-IgA tests of four manufacturers in a pediatric cohort of 59 active celiac and 98 control children. Duodenal biopsies were used as reference standard. Furthermore, we assessed the rate of positive TG2-IgA, DGP-IgG and DGP-IgA at different time points under GFD in 91 celiac children who had presented for follow-up to our department. We confirmed the high specificity of all tests, and a reduced sensitivity of DGP-IgA, compared to DGPIgG and TG2-IgA to diagnose CD [7]. Major differences were found between the 4 manufacturers with respect to the proportions of TG2-IgA values of > 10 × ULN. Under a GFD DGP-IgA normalized more rapidly than DGP-IgG and TG2-IgA. TG2-IgA showed the
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
111
Downloaded by: National University of Singapore. Copyrighted material.
Originalarbeit
Originalarbeit
Fig. 5 Correlation coefficient of all tests for serum measurement of TG2-IgA, DGP-IgG and DGP-IgA (p-value was calculated according to Spearman). BS: TheBindingSite, EU: Euroimmun, IN: INOVA, PH: Phadia, part of Thermo Fisher Scientific.
highest titer rise with non-adherence to the GFD, independent of the manufacturer. Current guidelines recommend the measurement of TG2-IgA and total IgA for initial screening of CD [1]. Various commercial providers and laboratories are offering combinations of TG2-IgA, DGPIgA, DGP-IgG, and EMA-IgA for serological screening of CD in symptomatic or at risk patients, although the costs of such multiple testing are high and the benefit of such panels has not been established. Specific antibody tests from different providers have been validated against histology but not evaluated against each other, and it is unclear whether different commercial tests perform similarly [7]. In addition, it is not sufficient to test for antibody positivity or negativity and only the most recent studies considered the antibody concentrations [1]. Hence, we compared quantitative results of three different tests of four manufacturers for the detection of celiac specific antibodies. In our cohort of IgA competent children, TG2-IgA testing detected all celiac patients with active disease, regardless of the manufacturer, while DGP-IgG testing alone or the combination of DGPIgG and DPG-IgA missed between 1 – 6 celiac patients, except for manufacturer D. Similar results have been described for other cohorts of adults and children [5, 8, 14, 15]. As previously shown, IgA antibodies against DGP showed a significantly lower sensitivity than TG2-IgA and DGP-IgG [5, 7, 15]. The excellent results for TG2-IgA compared to DGP-testing could be in part biased by the fact that the majority (53/59) of our active CD patients were initially tested with a TG2-IgA assay (A) at our institution. Of note, results of TG2-IgA measurements by three additional tests (B, C, D) showed a high correlation to our screening assay (A). With respect to specificity all 12 tests had an excellent performance. One control patient with Marsh 0 on histology was low titer positive in all performed test. Celiac specific antibodies can be induced by viral infections or other unknown environmental factors, and may show fluctuating positive levels in patients at genetic risk for CD [16]. Therefore we cannot exclude true CD in this patient with false negative histology due to patchy disease. Overall titers of TG2-IgA were significantly higher than titers of DGP-IgG or DGP-IgA as calculated by fold increase of ULN. Ac-
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
cording to the recent ESPGHAN guidelines TG2-IgA titers of > 10 × ULN, positive endomysial antibodies in a second blood sample, HLA-DQ2/8 positivity and clinical signs suggestive for CD are sufficient to establish the diagnosis of CD [1]. There were major differences between the four TG2-IgA tests regarding titer levels of > 10 × ULN. Depending on the test 47 – 90 % of our patients would be considered for further non-invasive testing to establish the diagnosis without biopsies. A high proportion of CD patients with TG2-IgA > 10 × ULN (90 − 91 %) was also reported by Vermeersch et al. in a mixed population of children and adults by using 4 different commercial assays [17]. The variability of values > 10 × ULN between single TG2-IgA tests points to the need of standardization between assays as pointed out by Egner et al. [18]. The cut-off of each test needs to be evaluated according to the ESPGHAN guidelines to circumvent intestinal biopsies. Since we did not test for HLA-DQ2/8 positivity or endomysial antibodies in all patients, we cannot extrapolate the proportion of patients to meet the current ESPGHAN recommendations for diagnosis of CD. This was not the purpose of this retrospective study. " Fig. 2 illustrates the different performThe Venn diagram of ● ances of the four manufacturers. The measurement of either DGP-IgG or DGP-IgA did not add any value to TG2-IgA results. These results are in concordance with a previous report [8] and underline the fact that the measurement of DGP-IgG or -IgA in combination with TG2-IgA does not improve celiac serology. DGP-antibodies bind the disease-inducing antigen and might be superior to monitor patients on a GFD. Liu and colleagues reported that titers of TG2-IgA and DGP antibodies (combined measurement of DGP-IgG and -IgA) paralleled in a cohort of 50 CD patients at diagnosis but DGP antibodies resolved sooner than TG2 IgA [19]. Consistent with this report and the study by Basso et al. [20] we found a significant drop of TG2-IgA and DGP-antibodies under GFD. DGP-IgA normalized most rapidly and already showed negative results in > 85 % of patients within the first 6 months of a GFD. Titers of DGP-IgA were similarly high to DGPIgG at diagnosis. Therefore, differences of the initial titer height cannot explain the more rapid normalization of DGP-IgA in comparison to DGP-IgG. In contrast, the late normalization of TG2-IgA in comparison to DGP-antibodies might be explained by initially
Downloaded by: National University of Singapore. Copyrighted material.
112
Originalarbeit
Abbreviations CD: DGP: DGP-IgA: DGP-IgG: EMA: GFD: TG2-IgA:
celiac disease deamidated gliadin peptide IgA antibodies against DGP IgG antibodies against DGP autoantibodies against endomysium gluten-free diet IgA antibodies against tissue transglutaminase
Acknowledgements !
The authors thank all staff physicians and nurses of the Division of Pediatric Gastroenterology, Hepatology and Nutrition for excellent patient care and support of the study. Conflict of interest: All tests were supplied free of charge by the manufacturing companies. The companies equally covered costs for laboratory supplies and the technician. The companies had no influence on study design or analysis of data. PB received lecture fees from MSD, GivenImaging and Abbvie. SK is advisory board member of Centocor, MSD, Danone, Merck, Vifor, Nestle Nutrition Institute, Abbvie and received lecture fees from MSD, Danone, Merck, Vifor, Nestle Nutrition Institute, Euroimmun, Thermo-Fisher (Phadia), Abbvie, Schär, Hipp and Falk.
References 01 Husby S, Koletzko S, Korponay-Szabo IR et al. European Society for Pediatric Gastroenterology, Hepatology, and Nutrition guidelines for the diagnosis of coeliac disease. J Pediatr Gastroenterol Nutr 2012; 54: 136 – 160 02 Leffler DA, Schuppan D. Update on serologic testing in celiac disease. Am J Gastroenterol 2010; 105: 2520 – 2524 03 Prince HE. Evaluation of the INOVA diagnostics enzyme-linked immunosorbent assay kits for measuring serum immunoglobulin G (IgG) and IgA to deamidated gliadin peptides. Clin Vaccine Immunol 2006; 13: 150 – 151 04 Sugai E, Vazquez H, Nachman F et al. Accuracy of testing for antibodies to synthetic gliadin-related peptides in celiac disease. Clin Gastroenterol Hepatol 2006; 4: 1112 – 1117 05 Prause C, Ritter M, Probst C et al. Antibodies against deamidated gliadin as new and accurate biomarkers of childhood coeliac disease. J Pediatr Gastroenterol Nutr 2009; 49: 52 – 58 06 Niveloni S, Sugai E, Cabanne A et al. Antibodies against synthetic deamidated gliadin peptides as predictors of celiac disease: prospective assessment in an adult population with a high pretest probability of disease. Clin Chem 2007; 53: 2186 – 2192 07 Giersiepen K, Lelgemann M, Stuhldreher N et al. Accuracy of diagnostic antibody tests for coeliac disease in children: summary of an evidence report. J Pediatr Gastroenterol Nutr 2012; 54: 229 – 241 08 Olen O, Gudjonsdottir AH, Browaldh L et al. Antibodies against deamidated gliadin peptides and tissue transglutaminase for diagnosis of pediatric celiac disease – diagnostic performance and cost in clinical practice. J Pediatr Gastroenterol Nutr 2012; 55: 695 – 700 09 Monzani A, Rapa A, Fonio P et al. Use of deamidated gliadin peptide antibodies to monitor diet compliance in childhood celiac disease. J Pediatr Gastroenterol Nutr 2011; 53: 55 – 60 10 Oberhuber G, Granditsch G, Vogelsang H. The histopathology of coeliac disease: time for a standardized report scheme for pathologists. Eur J Gastroenterol Hepatol 1999; 11: 1185 – 1194 11 Pellegrino S, Villanacci V, Sansotta N et al. Redefining the intraepithelial lymphocytes threshold to diagnose gluten sensitivity in patients with architecturally normal duodenal histology. Aliment Pharmacol Ther 2011; 33: 697 – 706 12 Walker MM, Murray JA, Ronkainen J et al. Detection of celiac disease and lymphocytic enteropathy by parallel serology and histopathology in a population-based study. Gastroenterology 2010; 139: 112 – 119 13 Revised criteria for diagnosis of coeliac disease. Report of Working Group of European Society of Paediatric Gastroenterology and Nutrition. Arch Dis Child 1990; 65: 909 – 911
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
Downloaded by: National University of Singapore. Copyrighted material.
higher titers of TG2-IgA. A decrease of celiac specific antibodies was associated with better adherence to the GFD in adults and the combined measurement of DGP-IgA and TG2-IgA was best for follow-up [21]. Even if titers of DGP-IgA most rapidly normalized under the GFD we found that incidental or voluntary ingestion of gluten was best indicated by rising titers of TG2-IgA. All four TG2-IgA assays performed similarly, but test B showed the highest titer rise upon unforeseen gluten intake. The high titers in test B are also reflected in the high percentage of children with TG2 IgA > 10 × ULN by this assay. Under a GFD, we only found a low and transient titer rise of either TG2-IgA, DGP-IgG or DGP-IgA or combinations of tests in 7/91 patients. From gluten challenge studies it is well known that there is an individual response both in titer height and time interval to react with CD specific antibodies [22, 23]. Intermittent intake of small amounts of gluten is most likely not detected by serology, even if it shows adverse effects on the architecture of the mucosa [23]. Therefore, a TG2-IgA test with a high titer response will be advantageous to monitor compliance to GFD in patients with CD. Due to our study design we cannot calculate the sensitivity and specificity of DGP and TG2 antibodies to monitor adherence to the GFD as was done by other groups in children [9, 20, 24]. A recent prospective study showed that the negative likelihood ratio of EMA was best to rule out persistent mucosal injury during GFD in children with CD in comparison to TG2-IgA and DGP-antibodies [24]. In some contrast, Basso et al. reported in children that the sensitivity of TG2-IgA to detect non-adherence to the GFD was only 24 % and was exceeded by DGP-IgG and IgA (76 % and 60 %) [20]. In accordance, two studies in adults described that normal TG2-IgA and EMA were unable to predict recovery of villous atrophy in adult patients with CD under GFD [25, 26]. We found an excellent quantitative correlation of the individual antibody tests applied to this study. This is an important finding since there are no widely accepted gold standards for specific CD antibody assays [27]. However, the correlation between TG2-IgA, DGP-IgG and DGP-IgA titers was weak and confirmed a study by Liu [19]. Limitations of this study are the retrospective design and the screening of patients with CD by TG2-IgA of a single company (A) in the majority of patients. Therefore we can only indicate the sensitivity and specificity of the different serological tests for diagnosis of CD. However, the fact that all TG2-IgA assays showed a similar sensitivity, points to the diagnostic strength of TG2-IgA as the antibodies directed against the auto-antigen in CD. With the writing of this manuscript one test (A) is off market now. In summary, we show that TG2-IgA and DGP-IgG tests provided by four manufacturers are superior to DGP-IgA for initial diagnostic testing of children with suspected celiac disease. The rate of TG2-IgA titers > 10 × ULN at diagnosis differed significantly between the four TG2-IgA tests indicating that each test needs evaluation whether it fits the current ESPGHAN guidelines to circumvent intestinal biopsies. Non-adherence to a GFD was best indicated by a high titer rise of TG2-IgA.
113
Originalarbeit
14 Rashtak S, Ettore MW, Homburger HA et al. Comparative usefulness of deamidated gliadin antibodies in the diagnosis of celiac disease. Clin Gastroenterol Hepatol 2008; 6: 426 – 432; quiz 370 15 Vermeersch P, Geboes K, Marien G et al. Diagnostic performance of IgG anti-deamidated gliadin peptide antibody assays is comparable to IgA anti-tTG in celiac disease. Clin Chim Acta 2010; 411: 931 – 935 16 Simell S, Kupila A, Hoppu S et al. Natural history of transglutaminase autoantibodies and mucosal changes in children carrying HLA-conferred celiac disease susceptibility. Scand J Gastroenterol 2005; 40: 1182 – 1191 17 Vermeersch P, Geboes K, Marien G et al. Defining thresholds of antibody levels improves diagnosis of celiac disease. Clin Gastroenterol Hepatol 2013; 11: 398 – 403 18 Egner W, Shrimpton A, Sargur R et al. ESPGHAN guidance on coeliac disease 2012: multiples of ULN for decision making do not harmonise assay performance across centres. J Pediatr Gastroenterol Nutr 2012; 55: 733 – 735 19 Liu E, Li M, Emery L et al. Natural history of antibodies to deamidated gliadin peptides and transglutaminase in early childhood celiac disease. J Pediatr Gastroenterol Nutr 2007; 45: 293 – 300 20 Basso D, Guariso G, Fogar P et al. Antibodies against synthetic deamidated gliadin peptides for celiac disease diagnosis and follow-up in children. Clin Chem 2009; 55: 150 – 157
21 Nachman F, Sugai E, Vazquez H et al. Serological tests for celiac disease as indicators of long-term compliance with the gluten-free diet. Eur J Gastroenterol Hepatol 2011; 23: 473 – 480 22 Lahdeaho ML, Maki M, Laurila K et al. Small-bowel mucosal changes and antibody responses after low- and moderate-dose gluten challenge in celiac disease. BMC Gastroenterology 2011; 11: 129 23 Leffler D, Schuppan D, Pallav K et al. Kinetics of the histological, serological and symptomatic responses to gluten challenge in adults with coeliac disease. Gut 2012; DOI: 10.1136/gutjnl-2012-302196 24 Vecsei E, Steinwendner S, Kogler H et al. Follow-up of pediatric celiac disease: value of antibodies in predicting mucosal healing, a prospective cohort study. BMC Gastroenterology 2014; 14: 28 25 Hopper AD, Hadjivassiliou M, Hurlstone DP et al. What is the role of serologic testing in celiac disease? A prospective, biopsy-confirmed study with economic analysis. Clin Gastroenterol Hepatol 2008; 6: 314 – 320 26 Sharkey LM, Corbett G, Currie E et al. Optimising delivery of care in coeliac disease – comparison of the benefits of repeat biopsy and serological follow-up. Aliment Pharmacol Ther 2013; 38: 1278 – 1291 27 Li M, Yu L, Tiberti C et al. A report on the International Transglutaminase Autoantibody Workshop for Celiac Disease. Am J Gastroenterol 2009; 104: 154 – 163
Downloaded by: National University of Singapore. Copyrighted material.
114
Bufler P et al. Diagnostic performance of … Z Gastroenterol 2015; 53: 108–114
