Understanding the evolution of the algorithms for the diagnosis of celiac disease (CD) may provide insight into the challenges we face trying to improve diagnostic accuracy with less invasive procedures. The diagnosis of CD has undergone significant transformation since the observations of Samuel Gee more than a century ago. The diagnosis was originally that of a clinical syndrome of chronic indigestion, loose, unformed, and malodorous stool affecting all ages but especially young children (1). Half a century later, Dicke noted that exclusion of prolamines led to clinical improvement of patients, leading to the exclusion of gluten as a diagnostic test as well as a therapeutic tool (2).
The advent of the Crosby-Kugler capsule, which enabled tissue sampling and histological examination of intestinal tissue, together with the later development of the Marsh criteria of mucosal damage in CD added another tier to the diagnostic workup of CD. The identification of gluten antibodies revolutionized the clinical approach to CD diagnosis, because it provided an accessible method to initiate work up of patient with suspected CD. Unfortunately, these antibodies were neither sensitive nor specific enough to be relied upon for diagnosis, and therefore clinical food challenges and rechallenges were still necessary for the diagnosis. In 1969 the European Society for Paediatric Gastroenterology met in Interlaken and formulated what could be considered the first European consensus statement—the Interlaken criteria—based on an informal expert response to a questionnaire (3). Following these guidelines the need for intestinal biopsies before diagnosis, following a gluten-free diet (GFD), and again after a gluten challenge became routine. Clinical symptoms and steatorrhea were generally considered part of the CD spectrum, although a silent form of CD was already recognized.
Three decades ago, anti-endomysium immunoglobulin A (EMA-IgA) was identified, and with a specificity of close to 100% provided a serological test, which, combined with histological examination, served as a basis for the description of both latent CD and the identification of atypical disease without overt gastrointestinal symptoms. In 1990, the European Society for Pediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) published revised criteria indicating that CD could be diagnosed with a single biopsy in symptomatic patients responding to a GFD (reserving gluten challenges for asymptomatic individuals, young children, and questionable cases) (4). As for use of serology as a diagnostic procedure, it was stressed that “the diagnosis of coeliac disease cannot be made on the presence of these antibodies alone,” because of both false positives and negatives.
In 1997, tissue transglutaminase was found to be the autoantigen in endomysium, and the identification of anti-tissue transglutaminase antibodies (TTG) would enable simpler and less expensive methods of screening for CD than EMA. More recently, antibodies against deaminated gliadin peptide (DGP) were added as a tool for diagnosing CD in IgA-deficient patients. Biochemical assays were improved and sensitivity and specificity were increased. Although EMA-IgA is highly specific, it is expensive and carries with it the need for expertise in the performance of the test. TTG-IgA also has high sensitivity and specificity, is less expensive, and has the advantage of being performed in an automated manner. Thus, it appears that TTG-IgA (and DGP-IgG in IgA-deficient subjects) are the serological tests of choice in the diagnosis of CD (5).
Genetic studies identified HLA-DQ2 and -DQ8 as the major determinants of CD susceptibility. Because these haplotypes are common in the general population, HLA determination is better suited for ruling out the presence of CD in suspicious cases rather than assisting the diagnosis of CD.
In 2012, ESPGHAN set forth to revise the diagnostic algorithms of CD in children, and to determine if a less invasive method, not involving intestinal biopsies could be practiced (5). The 2012 recommendations differentiate between several clinical scenarios., For symptomatic children with high levels of both TTG IgA (>10x upper limit of normal (ULN)) and EMA IgA, as well as compatible HLA, biopsy may be omitted, as the available evidence and expert consensus was that the combination of strong pretest probability, and test results would not overdiagnose CD. The guidelines recognize the high predictive value of accurate serology and the limitations of the biopsy resulting from the patchy nature of the disease (raising the possibility of false negative results). Patients not meeting these strict criteria, even if symptomatic, should still undergo histologic evaluation of duodenal biopsies (both bulb and second portion). Important pillars of the new guidelines are that diagnosis must be done on a gluten-containing diet and that repeated serological tests are needed when biopsy is omitted.
The new guidelines set forth by ESPGHAN have challenged the necessity for the use of the gold standard for diagnosis (serology + histology), stating that in a subset of cases a new standard may perform just as well (clinical scenario + genetic background + histology). It remains to be determined whether the new clinical/laboratory standard functions as well as the gold standard of serology + biopsy, whether different cutoffs values for serology kits should be used, and whether including HLA typing is necessary.
Currently there are still no published prospective studies evaluating the new diagnostic guidelines. Nevoral et al retrospectively evaluated 345 children evaluated for suspected CD (6). Their findings were in line with the guidelines including the insistence that all asymptomatic patients undergo endoscopy since highly elevated TTG (>10ULN) and positive EMA had only 67% specificity for Marsh grade 2–3 lesions in the asymptomatic group, compared to 99% for symptomatic patients. Of note, this study did not evaluate patient HLAs which may significantly alter the results in the asymptomatic patients. Recently Klapp et al reported a retrospective analysis of 153 suspected CD patients with results of TTG, EMA, HLA and endoscopy (7). They found that the 2012 ESPGHAN guidelines had a 97.4% PPV for a Marsh 2–3 lesion at initial endoscopy and a 100% PPV for a diagnosis of CD upon further follow-up. One symptomatic patient with positive EMA, TTG>10xULN and negative HLA was found to have Marsh 0–1 at initial endoscopy. He was subsequently diagnosed with cow's-milk allergy. EMA and TTG disappeared on milk elimination. Although rare, this seems to indicate that the inclusion of HLA as an integral part of the criteria may be justified. This highlights the need for prospective evaluation of the guidelines. An ESPGHAN multi-center study coordinated by Koletzko and colleagues is currently evaluating the recommendations with a prospective multinational study (PROCEDE) in which patients being evaluated endoscopically for suspected CD are followed clinically and serologically in order to understand whether the proposed guidelines are appropriate.
When implementing the new ESPGHAN guidelines clinicians must take into account that not all serologic tests are created equal. Significant differences exist in so far as assay methodology and reporting. TTG results are reported in relative units and therefore continuous participation in quality control programs is necessary to maintain relative uniformity in reporting. Furthermore, care providers must be aware that some commercial kits use logarithmic reporting scales as opposed to linear results of other assays. EMA immunofluorescence assessment carries an inherent interobserver variability that must also be considered and laboratory standards maintained.
Although the 2012 guidelines have simplified the diagnosis of CD for a large fraction of patients, there remain several issues unresolved concerning the diagnosis of the disease and its ramifications. Will patients diagnosed based on blood tests and symptoms be as compliant to a GFD as those diagnosed by biopsy, which is often considered by patients as “hard evidence” of disease? Also, lower thresholds for omitting the biopsy may perform as accurately as the current ones. Regardless, we still do not know whether or not there are subgroups of patients with CD who may be categorized based on their risk of future complications. While evidence builds up as to the morbidity and mortality risks of CD on gluten-containing diets in general, it remains to be seen whether there are subgroups who may not need to be immediately placed on such diets or if there are some patients who may be taken off the diet after a certain time period. For example, if we omit biopsies for asymptomatic patients, this will mean that potential CD is treated with GFD. Intriguing and controversial is the report by Matysiak-Budnik et al (8) who described a group of patients with CD diagnosed as children, who, following an initial GFD, returned to normal diets for prolonged periods of time (mean 14 years). A portion of these patients seemed to maintain latency in so far as symptoms, mucosal damage and even osteopenia.
Currently, only HLA is used for defining risk for CD; however, the minor genetic determinants of CD may have an impact on phenotype or natural history of the disease more than simply inducing disease. In the future, we may need to look at a panel of HLA and non-HLA genes to predict disease behavior. Future research may delineate genotype–phenotype relationships which could predict poor outcome on gluten-containing diets.
Intriguing studies have identified CD4+ gluten-DQ2 tetramers in the peripheral blood of CD patients following a short gluten challenge (9). These cells were not present in controls, or in the patients with CD while on GFDs. More recently, Galatola et al reported that a small gene expression panel from peripheral blood monocytes could discriminate between active CD and healthy controls (10). As we advance towards biopsy-free diagnosis of CD, such analyses of peripheral blood cell parameters may add further security to the diagnosis of CD, and perform as a proxy to the histological status of the intestine. This may be more appropriate since CD is a multi-organ disease that is not limited to the intestine. Furthermore, now that clinicians encounter patients on “low gluten” diets more often than in the past, either because they live in CD households (eg, siblings), because of supposed general health benefit to non-CD populations, or suspected nonceliac gluten sensitivity, surrogate markers that are not dependent on a gluten-containing diet are desirable.
As we improve our understanding of the pathogenesis of CD, the interplay of genetic, epigenetic, and environmental factors may need to be considered as a part of the diagnostic process. An example of such could be the influence of an individual's microbiota on CD development or course. Gastrointestinal microbiota has been reported, in some studies, to differ between patients with CD and healthy controls (11–13). Although still unclear as to cause and effect, the findings beckon the possibility that analysis of one's duodenal microbiota footprint, or the presence of certain cell surface receptors allowing bacterial adhesion in patients with CD may play a role in the diagnosis, prognostication, or treatment of CD.
In summary, the 2012 ESPGHAN guidelines have taken us a step forward towards confidently diagnosing certain patients with CD noninvasively. At the same time they have kept the door open to further advances, both as finesses of the protocol itself, as well as into new and unexplored directions to aid in the diagnosis of the disease. As we tread forward into the age of personalized medicine, predictive outcome measures will most likely gradually become integral in the diagnosis of a CD spectrum of diseases.
1. Losowsky MS. A history of coeliac disease. Dig Dis
2. Tommasini A, Not T, Ventura A. Ages of celiac disease: from changing environment to improved diagnostics. World J Gastroenterol
3. European Society for Paediatric Gastroenterology. Acta Pædiatrica
4. Revised criteria for diagnosis of coeliac disease. Report of Working Group of European Society of Paediatric Gastroenterology and Nutrition. Arch Dis Child
5. 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
6. Nevoral J, Kotalova R, Hradsky O, et al. Symptom positivity is essential for omitting biopsy in children with suspected celiac disease according to the new ESPGHAN guidelines. Eur J Pediatr
7. Klapp G, Masip E, Bolonio M, et al. Celiac disease: the new proposed ESPGHAN diagnostic criteria do work well in a selected population. J Pediatr Gastroenterol Nutr
8. Matysiak-Budnik T, Malamut G, de Serre NP-M, et al. Long-term follow-up of 61 coeliac patients diagnosed in childhood: evolution toward latency is possible on a normal diet. Gut
9. Raki M, Fallang L-E, Brottveit M, et al. Tetramer visualization of gut-homing gluten-specific T cells in the peripheral blood of celiac disease patients. Proc Natl Acad Sci
10. Galatola M, Izzo V, Cielo D, et al. Gene Expression Profile of Peripheral Blood Monocytes: A Step towards the Molecular Diagnosis of Celiac Disease? PLoS ONE
11. Cheng J, Kalliomaki M, Heilig H, et al. Duodenal microbiota composition and mucosal homeostasis in pediatric celiac disease. BMC Gastroenterology
12. Sellitto M, Bai G, Serena G, et al. Proof of concept of microbiome-metabolome analysis and delayed gluten exposure on celiac disease autoimmunity in genetically at-risk infants. PLoS ONE
13. Sanchez E, Donat E, Ribes-Koninckx C, et al. Duodenal-mucosal bacteria associated with celiac disease in children. Appl Environ Microbiol