Hope, Benjamin C.*; Ameratunga, Rohan†; Austin, Paul M.‡; Evans, Helen M.§; MacFarlane, Jeannette||; Mouat, Stephen§; Chin, Simon E.§
Coeliac disease (CD) is a chronic inflammatory condition with a high prevalence in many countries (1,2). Its nonclassical manifestations in childhood have received widespread recognition in recent years (3,4). As a consequence, children with common complaints, such as recurrent abdominal pain, are now more likely to undergo serological screening for CD. Standard CD serology includes assays for anti-endomysial (EMA) and anti-tissue transglutaminase antibodies (TGA). In the research setting, these assays have sensitivities and specificities ranging from 90% to 95% (5–7); however, in the community setting, their performance may not be as impressive (8,9).
Circulating antibodies to wheat proteins were first found in patients with CD as long ago as the 1960s (10), but the early antigliadin assays that resulted from this discovery have poor diagnostic performance in comparison with TGA- and EMA-based tests (11). More recent research has shown that deamidation of the gliadin peptide is an important step in the immune-mediated pathogenesis of CD. Deamidation of gliadin by tissue transglutaminase (TG) creates peptide complexes, which activate T cells, leading in turn to mucosal inflammation and damage (12). It was hypothesised that assays based on deamidated gliadin may offer improved accuracy in the diagnosis of CD. Working with selectively deamidated gliadin peptides (DGPs) in 2001, Aleanzi et al (13) achieved encouraging results. Further research, using commercially available enzyme-linked immunosorbent assay (EIA) kits, has demonstrated that TG and DGP assays produce highly concordant results in both adults and children (14–16).
At Starship Children's Hospital, there have been considerable numbers of children with positive coeliac serology but subsequently normal duodenal histology. Aiming to improve the local screening process, we evaluated the performance of several modified anti-gliadin antibody (AGA) tests. They were compared with the standard TGA assays used by clinical laboratories in the Auckland region. Several other serological CD markers, including first-generation AGA tests, were also evaluated, although the EMA was not.
Coeliac serology–positive children, attending Starship Hospital for endoscopic small bowel biopsy, were recruited sequentially. Screening blood tests were processed at 1 of 8 laboratories around New Zealand, reflecting the large geographic area served by Starship Hospital. The referral tests always included a TG-based test, and sometimes EMA results were also available. Among the TG kits used were those made by Euroimmun (Lübeck, Germany), Biorad (Hercules, CA), and Inova (San Diego, CA). Children awaiting endoscopy for suspected CD were instructed to maintain normal dietary gluten intake before endoscopy.
Informed consent was obtained from the children and their families. The study was approved by the local regional ethics committee.
Repeat Serologic Testing
Blood samples were taken at the time of endoscopy. All of the sera were analysed at Auckland Hospital's Labplus laboratory by the same senior technician (P.A.) using commercial EIA according to the manufacturer's guidelines (Table 1). Assays included Inova Diagnostics anti-gliadin IgA (Inova GA) and IgG (Inova GG), anti-tissue transglutaminase IgA (Inova TGA) and IgG (Inova TGG), combined IgA and IgG anti-deamidated gliadin peptide (Inova DGPAG), and combined DGPTG screen (Inova DGPTG); Euroimmun GAF-3X (anti-modified gliadin IgG, Euro MGPG); Binding Site (Birmingham, UK) MGP IgA and IgG (Binding MGPA and Binding MGPG). The laboratory was blinded to the biopsy results at the time of serum analysis. Results were reported according to the limits set by the manufacturers. Total serum IgA was determined using rate nephelometry.
Minimum 6 biopsies were taken from the second part of the duodenum (Olympus [Tokyo, Japan] FB-25K endoscope, jumbo forceps). The histology was read or reviewed by a single pathologist (J.M.) according to the Marsh classification (17). After the first 54 subjects had been enrolled, there was a change in departmental policy, which recommended samples from the duodenal bulb, in addition to those from the second part. The pathologist was blinded to the result of the repeat serological testing.
The JMP version 5.1 (SAS Institute Inc, Cary, NC) and StatsDirect version 2.7.7 (StatsDirect Ltd, Cheshire, UK) software programs were used for statistical analysis. Assays were compared using an exact test of paired proportions (an alternative to the McNemar test). Confidence intervals (CIs) for the difference in proportions between diagnostic assays were calculated using Newcombe refinement of the Wilson score–based method. CIs for the percentages of true-positives and true-negatives were constructed using the Clopper and Pearson method.
The study period was November 2008 to December 2009. There were 79 consecutive participants, 42 of whom were girls (Table 2) (18,19). They were referred by paediatricians and general practitioners in the greater Auckland region. The age range was 18 months to 16 years (median 6 years). There were no exclusions, and none of the families declined enrollment. Reasons for performing initial CD serology included gastrointestinal symptoms, poor growth, refractory iron-deficiency anaemia, and family history. Eight had trisomy 21; 6 had type 1 diabetes mellitus (T1DM); and 1 had William syndrome. Two had low levels of IgA based on laboratory reference ranges for age.
At the time of endoscopy, 39 of 79 (49.4%) were Inova TGA positive and 35 of 79 (44.3%) were Inova DGPAG positive (Fig. 1). The other MGP assays showed similar positive predictive values (PPVs). Twenty-four of 39 (61.5%) who were Inova TGA-positive and 24 of 35 (68.6%) who were Inova DGPAG positive had confirmed CD on biopsy. Overall, 24 of the 79 participants (30%) had small bowel histology diagnostic of CD. All of these had Marsh 3 lesions. There were no discrepancies between duodenal bulb and more distal biopsies in those who had samples taken from the bulb. Of the proven CD cohort, 21 (88%) were negative for Inova TGG, 4 were negative for Binding MGPA, and 6 were negative for Inova GA. There were no other false-negative results from the other kits studied. Raising the cutoff to 30 U/mL for the 2 main Inova kits (TGA and DGPAG) resulted in an improved PPV and no false-negatives (Table 3). The Inova assays correlated well (Fig. 2).
The PPVs for the Inova TGG, Inova GG, and Binding MGPA were lower than the reference Inova TGA, with P values <0.05. Inova DGPTG had a lower PPV than either the Inova DGPAG or TGA alone, although the P value was 0.38.
The performance of Inova TGA was similar to Inova DGPAG in those children with symptoms. The proportion of biopsy-negative subjects was similar in both “symptomatic” and “screened” groups (69% and 73%, respectively), although the total number of screened cases (15) was small.
Higher titres of Inova TGA and DGP were associated with an increased likelihood of histologic CD, with this trend more marked for TGA (Figs. 3 and 4). There was also a statistically insignificant trend towards a greater likelihood of a serology-positive/biopsy-negative outcome, with increasing time interval from referral.
Both subjects with low total IgA (but not absolute deficiency) had normal biopsies. One had high titres of Inova TGA (>194) and was positive for Binding MGPA (19, cutoff <10 U/mL). The other was only positive for IgG-based tests (Inova GG and Binding MGPG).
None of the biopsies showed Marsh 2 changes. Of the 4 cases with Marsh 1 intraepithelial lymphocytosis (IEL), the same 2 were positive for Inova TGA and DGPAG, as well as the IgG MGP tests from the other 2 manufacturers.
The 2 trisomy 21 subjects who had CD were positive for all assays except Inova TGG. The same was true for the 2 patients with T1DM who had CD. One of the patients with T1DM was also 1 of the 2 cases of IEL with multiple positive assays, but was not given a diagnosis of CD.
Our results suggest that modified gliadin-based assays offer diagnostic performance, which is comparable with that of TGA-based tests, for a population of New Zealand children deemed to be at risk for CD. Using blood drawn at the time of endoscopy, 61.5% of TGA-positive and 68.6% of DGPAG-positive children had histologic CD. These PPVs are considerably lower than those often found in the existing literature; however, many of the published studies were carried out in a research setting on selected populations or with a retrospective design. One of the strengths of the present study is that the subjects represent the heterogeneous population, which now undergoes CD screening in the community.
As expected, the older Inova GA and GG tests appear unhelpful in this group of children, although Lagerqvist et al (20) have argued in favour of their use, in conjunction with a TG assay, in children younger than 18 months.
All of the children enrolled in the study had positive coeliac antibodies (TGA) at the time of referral; just less than half remained TGA positive at the time of endoscopy. Thus, less than one-third (30%) of children referred with positive TGA were diagnosed as having CD after their first endoscopy. The finding is as perplexing as it is surprising. Comparison could be made to the data of Simell et al (21) on the natural history of TGA seropositivity in a cohort of prospectively studied Finnish children, in whom 35 of 1101 developed positive TGA and only 13 of 35 (37%) had CD changes at small bowel biopsy. Our study was not designed to investigate the accuracy of the referring tests, but the quantitative results were checked retrospectively and all of the subjects were confirmed to have had TGA antibody titres higher than the relevant manufacturer's negative cutoff.
Transient and fluctuating seropositivity has been observed not infrequently in both children (21,22) and adults (23). The cause of the phenomenon is not known. It may arise from low-grade autoreactivity, which resolves with immune maturity. Parents of a seropositive child may, consciously or otherwise, restrict his or her gluten intake and thereby remove the antigenic agent. Waiting times for endoscopy during the present study were longer than were desirable (mean 5.3 months, range 0.4–19), which may have made this effect more prominent; however, families were instructed not to lower their child's gluten intake before endoscopy, and the gluten-free diet is challenging, even for parents of children with a firm diagnosis of CD and multidisciplinary support. There may be some genuine false-positives, that is, TGA antibodies detected erroneously by inadequate tests or laboratory techniques; however, it seems unlikely that this could be the major mechanism in so many cases tested during the course of a year in several different laboratories.
Some of the serology-positive/biopsy-negative children may have potential CD, particularly those with IEL. In keeping with peer research and present guidelines, a diagnosis of CD in the present study was restricted to those with Marsh 3 mucosal changes (24,25). There were only 4 cases of IEL without villous atrophy, a small proportion of the serology-positive/biopsy-negative group. Accessible laboratory tests are not yet available to distinguish between nonspecific IEL and IEL as a harbinger of CD. Genome-wide association studies have identified nonhuman leukocyte antigen (HLA) loci that predispose to CD, but these are not sufficiently specific to assist in diagnosis (26). Highly sensitive immunohistochemical techniques, such as detection of TG antibodies in affected duodenal mucosa, are under investigation. They may prove valuable in determining which serology-positive/biopsy-negative children are at risk for progressing to CD enteropathy; however, this method is still restricted to the research domain (27). Because of the difficulty of ruling out potential CD, we have been repeating CD serology plus HLA-DQ2/8 testing 6 to 12 months after initial, normal duodenal biopsies. Absence of both HLA-DQ2 and DQ8 alleles makes potential CD highly unlikely, but does not appear to be sufficient to rule it out completely (28,29); however, children who are referred de novo with low titres of TGA, with minimal symptoms, and found to be HLA-DQ2 and HLA-DQ8 negative, are not offered endoscopy at our hospital routinely.
Subgroup analysis was not undertaken formally because of the small numbers involved. Of the 8 TGA-positive trisomy 21 patients referred for endoscopy, 2 remained positive when retested, and both had CD confirmed histologically. Five of the 6 patients with T1DM remained TGA positive at the time of biopsy: 1 had IEL, 2 had Marsh 3 changes, and 2 had normal biopsies.
Although the study design did not include long-term follow-up, additional data on some of the serology-positive/biopsy-negative group are now available. There were 15 patients who were Inova TGA positive and biopsy negative after their first repeat blood test and endoscopy. Of these, 7 had a further Inova TGA test at Labplus, with results available to the authors. Six patients remained TGA positive and 4 of them had undergone HLA typing and were positive for DQ2 or DQ8. Only 1 child had undergone repeat biopsies at Starship Hospital at the time of writing; these were positive for CD.
It is known that CD can have patchy mucosal involvement (30,31). This, in theory, could lead to a missed diagnosis at endoscopy, if only unaffected areas were sampled. At the start of the study, to minimise this risk, 6 biopsies were taken from the second and third parts of the duodenum of children with suspected CD. Attempts to further reduce the risk of false-negative endoscopy results by routinely sampling the duodenal bulb did not change serology-biopsy correlations, although a large study found that 2.5% of 665 children had histological CD at this proximal location only (32).
Donaldson et al (7) have demonstrated a correlation between higher TGA titres and the likelihood of having biopsy-proven CD. This trend was also observed here, more convincingly for Inova TGA than the same manufacturer's DGPAG. Interestingly, there were Inova TGA-positive, biopsy-negative cases across the range of titres, whereas there were no negative biopsies in children with Inova DGPAG titres >100. There were no proven cases with Inova TGA or DGPAG titres <30, raising the possibility of adapting the manufacturer's cutoff for the paediatric population, without having an unacceptable effect on the false-negative rate; however, much larger community studies are required to validate this. Another question that arose was whether children with lower TGA titres at referral were more likely to have negative TGA titres at biopsy. Again, the study was not designed to answer this, but post hoc analysis showed that the median referral TGA titres were higher (178) in those who remained TGA positive at the time of biopsy than in those who became TGA negative (62), P < 0.0001.
The main limitation of the present study is that it lacks the power to show statistically significant differences between the main assays. Given the similarities in performance of the TGA and MGP assays, many more subjects would have been needed to reach statistical significance. It has been suggested that DGP positivity precedes that of TGA in younger children (33). Because of the small number of infants and young children in our study, we were not able to investigate this. We did not use the EMA assay, which is more costly, labour intensive, and subject to observer variability. Published results for EMA are excellent, however, and it would have been valuable to assess its performance in this setting.
In summary, the diagnostic performance of a traditional TGA-based test was matched by that of the MGP tests evaluated in the present study. The proportion of serology-positive/biopsy-negative children is higher than has been reported elsewhere, regardless of which assay was used. These children are being studied because some may develop CD later. Large, long-term prospective studies are needed to determine the use of modified gliadin peptide tests in the management of suspected CD in children.
The authors thank Children's Research Centre, Starship Hospital (Gail Gillies and Peter Reed), and Labplus personnel.
1. Hoffenberg EJ, MacKenzie T, Barriga KJ, et al. A prospective study of the incidence of childhood celiac disease. J Pediatr
2. Steens RF, Csizmadia CG, George EK, et al. A national prospective study on childhood celiac disease in the Netherlands 1993–2000: an increasing recognition and a changing clinical picture. J Pediatr
3. Ravikumara M, Tuthill DP, Jenkins HR. The changing clinical presentation of coeliac disease. Arch Dis Child
4. Lurz E, Scheidegger U, Spalinger J, et al. Clinical presentation of celiac disease and the diagnostic accuracy of serologic markers in children. Eur J Pediatr
5. Rostom A, Dube C, Cranney A, et al. The diagnostic accuracy of serologic tests for celiac disease: a systematic review. Gastroenterology
2005; 128 (4 suppl 1):S38–S46.
6. Ladinser B, Rossipal E, Pittschieler K. Endomysium antibodies in coeliac disease: an improved method. Gut
7. Donaldson MR, Firth SD, Wimpee H, et al. Correlation of duodenal histology with tissue transglutaminase and endomysial antibody levels in pediatric celiac disease. Clin Gastroenterol Hepatol
8. Hill ID. What are the sensitivity and specificity of serologic tests for celiac disease? Do sensitivity and specificity vary in different populations? Gastroenterology
2005; 128 (4 suppl 1):S25–S32.
9. Abrams JA, Brar P, Diamond B, et al. Utility in clinical practice of immunoglobulin a anti-tissue transglutaminase antibody for the diagnosis of celiac disease. Clin Gastroenterol Hepatol
10. Taylor KB, Truelove SC, Thomson DL, et al. An immunological study of coeliac disease and idiopathic steatorrhoea. Serological reactions to gluten and milk proteins. Br Med J
11. Reeves GE, Squance ML, Duggan AE, et al. Diagnostic accuracy of coeliac serological tests: a prospective study. Eur J Gastroenterol Hepatol
12. Molberg O, McAdam SN, Korner R, et al. Tissue transglutaminase selectively modifies gliadin peptides that are recognized by gut-derived T cells in celiac disease. Nat Med
13. Aleanzi M, Demonte AM, Esper C, et al. Celiac disease: antibody recognition against native and selectively deamidated gliadin peptides. Clin Chem
14. Sugai E, Vazquez H, Nachman F, et al. Accuracy of testing for antibodies to synthetic gliadin-related peptides in celiac disease. Clin Gastroenterol Hepatol
15. Jaskowski TD, Donaldson MR, Hull CM, et al. Novel screening assay performance in pediatric celiac disease and adult dermatitis herpetiformis. J Pediatr Gastroenterol Nutr
16. 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
17. Marsh MN. Gluten, major histocompatibility complex, and the small intestine. A molecular and immunobiologic approach to the spectrum of gluten sensitivity (‘celiac sprue’). Gastroenterology
18. Cronk C, Crocker AC, Pueschel SM, et al. Growth charts for children with Down syndrome: 1 month to 18 years of age. Pediatrics
19. Martin ND, Smith WR, Cole TJ, et al. New height, weight and head circumference charts for British children with Williams syndrome. Arch Dis Child
20. Lagerqvist C, Dahlbom I, Hansson T, et al. Antigliadin immunoglobulin A best in finding celiac disease in children younger than 18 months of age. J Pediatr Gastroenterol Nutr
21. Simell S, Hoppu S, Hekkala A, et al. Fate of five celiac disease-associated antibodies during normal diet in genetically at-risk children observed from birth in a natural history study. Am J Gastroenterol
22. Tosco A, Salvati VM, Auricchio R, et al. Natural history of potential celiac disease in children. Clin Gastroenterol Hepatol
23. Johnston SD, Watson RG, McMillan SA, et al. Serological markers for coeliac disease: changes with time and relationship to enteropathy. Eur J Gastroenterol Hepatol
24. 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
25. Kurppa K, Salminiemi J, Ukkola A, et al. Utility of the new ESPGHAN criteria for the diagnosis of celiac disease in at-risk groups. J Pediatr Gastroenterol Nutr
26. Trynka G, Wijmenga C, Heel DA, et al. A genetic perspective on coeliac disease. Trends Mol Med
27. Tosco A, Salvati VM, Auricchio R, et al. Natural history of potential celiac disease in children. Clin Gastroenterol Hepatol
28. Karell K, Louka AS, Moodie SJ, et al. HLA types in celiac disease patients not carrying the DQA1*05-DQB1*02 (DQ2) heterodimer: results from the European Genetics Cluster on Celiac Disease. Hum Immunol
29. Bonamico M, Ferri M, Mariani P, et al. Serologic and genetic markers of celiac disease: a sequential study in the screening of first degree relatives. J Pediatr Gastroenterol Nutr
30. Bonamico M, Mariani P, Thanasi E, et al. Patchy villous atrophy of the duodenum in childhood celiac disease. J Pediatr Gastroenterol Nutr
31. Scott BB, Losowsky MS. Patchiness and duodenal-jejunal variation of the mucosal abnormality in coeliac disease and dermatitis herpetiformis. Gut
32. Bonamico M, Thanasi E, Mariani P, et al. Duodenal bulb biopsies in celiac disease: a multicenter study. J Pediatr Gastroenterol Nutr
33. 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
© 2013 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,