See “Role of Biopsies in Childhood Celiac Disease: Do We Need Less or More?” by Hill on page 173.
What Is Known
- More than 90% of pediatric patients with celiac disease achieve mucosal recovery after 1 year on a gluten-free diet.
- Current guidelines recommend follow-up serology both to assess dietary adherence and to use as a surrogate marker of mucosal recovery.
What Is New
- One in 5 children with celiac disease may have persistent enteropathy despite adherence to a gluten-free diet.
- Immunoglobulin A tissue transglutaminase may not be an accurate marker of mucosal recovery in these patients.
- We did not identify characteristics predictive of persistent enteropathy in this population.
Not long ago celiac disease (CD) was described as a childhood illness characterized by presentation with gastrointestinal symptoms, irritability, and poor growth (1). The development of a highly sensitive and specific diagnostic test for CD, the immunoglobulin A (IgA) tissue transglutaminase antibody (tTG) serology test, has allowed for minimally invasive and cost-effective screening of at-risk individuals and groups (2). This has transformed our clinical appreciation of CD and how dietary adherence and mucosal recovery are monitored. Currently, children diagnosed with CD are followed in close collaboration with dietetic colleagues for counseling to educate and promote adherence to the gluten-free diet (GFD). Growth and nutritional markers are monitored. In addition, current guidelines recommend follow-up serology both to assess dietary adherence and for use as a surrogate marker of mucosal recovery (3–6). In practice, antibody levels, usually the IgA tTG, are measured 6 months and then again 1 year after initiation of the GFD. When the level normalizes, it is presumed that mucosal recovery, attained through dietary adherence, has occurred. Adult patients with CD have similar follow-up with dietary counseling and the use of serologic markers. Studies evaluating mucosal healing in adults with CD on a GFD show that even after 2 years on a GFD, between one third and two thirds of patients have persistent mucosal damage, consistent with Marsh 3 criteria (7,8), irrespective of IgA tTG levels. Recent data evaluating pediatric mucosal recovery are scarce; however, early data have suggested a more complete and faster healing time in children compared with adults (9). Although not corroborated by evidence, anecdotally this has been explained by a decreased regenerative capacity of the adult intestine, resulting in slow and incomplete healing.
Although current guidelines endorse and recommend the use of serology as a marker of dietary adherence and mucosal recovery, these serology tests have not been validated for this purpose. Our objective was to determine the rate of mucosal recovery in the pediatric population and to evaluate whether tTG correlates with mucosal damage at the time of a repeat endoscopy with duodenal biopsy in children with CD on a GFD.
We performed a retrospective chart review at 2 hospitals and identified 103 pediatric patients who fulfilled our inclusion criteria described below. Study procedures were approved by the Partners Human Resource Committee and institutional review board at MassGeneral Hospital for Children (MGHfC) and Boston Children's Hospital (BCH).
Data for patients seen between January 2012 and March 2015 were extracted from medical records at MGHfC. Patients were identified by International Classification of Diseases, Ninth Revision code. Additional data were obtained for patients seen between January 2008 and December 2013 at BCH, identified using the BCH Celiac Database, a database of all children diagnosed with biopsy-proven CD at that institution. A total of 61 patients were identified at MGHfC and 42 patients were identified at BCH.
To be included in the study, patients were required to be 21 years or younger at the time of diagnosis of CD, must have had diagnostic endoscopy with Marsh 3 lesions, and have undergone a second endoscopy with duodenal biopsy at least 12 months after initiating a GFD. All patients meeting these criteria were included in the study.
The Celiac Research Program at Harvard Medical School is a collaboration between Beth Israel Deaconess Medical Center, BCH, and MGHfC, all located in Boston, MA. These 3 hospitals serve as quaternary care centers and referral centers for children and adults with CD and gluten-related disorders. Patients from Beth Israel Deaconess Medical Center were not included in the present study.
Data extracted from medical records included predominant clinical symptoms, serology tests, and duodenal histology at the time of the diagnostic and repeat endoscopy. Serology tests reviewed included IgA level, IgA tTG, and antiendomysial antibody (EMA) when available. Serological values collected within 4 months of the endoscopy were included in the analysis. We dichotomized serological values into positive/borderline or negative according to cut-off values defined by the laboratories, which performed the tests. Mucosal changes were scored by >1 pathologist at each institution using the Marsh criteria as modified by Oberhuber (0 = normal; 1 = increased intraepithelial lymphocytes [>25/100 epithelial cells], normal crypts and villi; 2 = increased intraepithelial lymphocytes, normal villi, crypt hyperplasia; 3 = increased intraepithelial lymphocytes, villous atrophy, crypt hyperplasia) (10,11). If endoscopic evaluation revealed multiple Marsh scores, the most severe was used.
For analysis, subjects were evaluated by the presence or absence of symptoms and length of time on a GFD. We also recorded whether subjects received any dietetic counseling, defined as a nutrition consultation with a Registered Dietician, during the interval between the initial endoscopy and the follow-up endoscopy. We reviewed the physician's and dietician's notes commenting on the subjects adherence to the GFD at the clinic visit before the repeat endoscopy and scored adherence using criteria modified from Leffler et al (12): excellent = patient never eats gluten intentionally and/or has rare exposure, good = inadvertent exposure once per month, poor = exposure 1 to 2 times per week, noncompliant = not on a GFD, or unable to assess GFD adherence from medical record.
Categorical data are presented as frequency (percentage) and group comparisons made with either the Pearson chi-squared statistic or Fisher exact test when the expected cell count was <5. Continuous data are described as mean ± SD if normally distributed and median (interquartile range) otherwise. Most continuous outcomes were right-skewed and therefore group comparisons were made with the Wilcoxon rank-sum test. Two-group comparisons of normally distributed variables were evaluated by Student t test. All tests of significance were 2-sided with α = 0.05, and all analysis performed with SAS (Cary, NC).
At the time of data collection 103 subjects were identified for inclusion into the study. Subjects were 10.6 ± 5.0 years of age at diagnosis and 60% were girls (Table 1). Consistent with the literature, tTG performed well at diagnosis with 89% of subjects with a positive IgA tTG. Eleven subjects had a negative tTG at diagnosis. Two subjects with a normal IgA tTG were asymptomatic at diagnosis. One of these subjects was IgA deficient, homozygous for DQ2, and had a family history of CD. This patient underwent endoscopy to evaluate a gastric lesion and was found to have duodenal villous blunting. The second asymptomatic subject had a normal IgA tTG but an elevated IgA EMA and a family history of CD which prompted the endoscopy. The other 9 patients had gastrointestinal complaints which prompted an evaluation by a gastroenterologist. Six out of 9 individuals had human leukocyte antigen testing, which confirmed genetic susceptibility for CD. One of these 6 patients had an elevated IgA EMA in the presence of a normal IgA tTG and 1 patient was IgA deficient. Of the 3 patients who had gastrointestinal complaints, normal IgA tTG, and no human leukocyte antigen testing, 1 patient had an elevated IgA EMA, 1 patient was younger than 2 years of age, and all 3 had improvement on a GFD with subsequent resolution of villous atrophy upon repeat endoscopy.
Overall, the majority of subjects presented with gastrointestinal complaints such as abdominal pain and constipation. Approximately 14% of subjects were asymptomatic at diagnosis and identified via routine screening of high-risk populations. In addition, 12% of subjects in our population had a comorbid autoimmune disease such as type 1 diabetes or thyroiditis. In line with the inclusion criteria, all subjects had intestinal enteropathy consistent with a Marsh 3 lesion at diagnosis.
Patient Characteristics at Repeat Biopsy
At the time of the repeat endoscopy, subjects had been following a GFD for a median of 2.4 years (interquartile range 1.4–4.0) with a range of 1 to 12 years (Table 2). In most cases, patients had >1 indication for a repeat endoscopy. The most common indications for repeat endoscopy were due to persistent symptoms (43%) and new gastrointestinal symptoms (27%). Twenty-four subjects (34%) had persistently elevated serology at the time of the repeat biopsy. The majority of subjects (91%) had excellent adherence to the GFD. One-hundred one subjects (98%) saw a dietician. Only 1 subject had no record of seeing a dietician as the other subject had a sibling with CD and the family had previously seen a dietician. Thirty-four subjects (33%) were asymptomatic at the time of the repeat endoscopy. Twelve asymptomatic patients underwent a repeat endoscopy for follow-up of another disease (1 ulcer, 2 ulcerative colitis, 9 esophagitis). Eleven asymptomatic patients had a repeat endoscopy due to persistently elevated serology (2 had both eosinophilic esophagitis and elevated serology). Four asymptomatic children underwent a repeat endoscopy to assess for mucosal recovery and confirm the diagnosis of CD in the setting of seronegative CD. Finally seven asymptomatic patients underwent endoscopy to confirm that mucosal recovery was achieved. In total, 19% exhibited persistent enteropathy consistent with a Marsh 3 lesion at the time of the repeat endoscopy.
Predictive Value of Tissue Transglutaminase in Identifying Marsh 3 Histology at Repeat Biopsy
In practice, IgA tTG was a poor predictor of Marsh 3 histology at repeat biopsy as sensitivity was 43%, specificity was 68%, the positive predictive value (PPV) was 25%, and the negative predictive value (NPV) was 83% (Table 3). The poor predictability of IgA tTG was consistent regardless of whether subjects reported symptoms at the time of the repeat endoscopy or the duration of the GFD. Of note, repeat serology within 4 months of the repeat biopsy was available for 71 of the subjects. Subjects with serology obtained >4 months from the time of the repeat biopsy were not included in this section of the analysis.
Concordance of Immunoglobulin A Tissue Transglutaminase With Histology at Repeat Biopsy
At the time of the follow-up biopsy, tTG was elevated in 43% of subjects with persistent enteropathy and 32% of subjects with mucosal healing. The overall concordance of IgA tTG with histology at the time of the repeat biopsy is shown in Table 4. These numbers are based on the 71 subjects that had an IgA tTG measured within 4 months of the repeat endoscopy.
Concordance of Symptoms With Histology at Repeat Scope
We examined the concordance between symptoms and mucosal recovery in the entire cohort (n = 103) strictly by the presence or absence of symptoms at the time of the repeat endoscopy (Table 5). We found that the majority of patients with complaints that may be related to CD had mucosal recovery 84%) (Marsh 0–2). Forty-five percent of patients with persistent enteropathy were asymptomatic. Twenty-six percent of the 23 asymptomatic patients who underwent repeat endoscopy to follow-up another disease (12), to assess for mucosal recovery due to seronegativity at diagnosis (4), or to assess for mucosal recovery (7), had persistent enteropathy. Twenty-seven percent of the 11 asymptomatic patients who underwent repeat endoscopy due to an elevated tTG had persistent enteropathy.
Comparison of Select Characteristics by Marsh Severity at Repeat Endoscopy
We found no significant characteristics predictive of persistent enteropathy. Sex, age at diagnosis, presence of a positive tTG within 4 months of the repeat endoscopy, and self-reported adherence to a GFD were not significantly different in subjects with persistent enteropathy compared with those in remission. In addition, there was no significant difference in frequency of persistent enteropathy when comparing subjects maintaining a GFD for >2 years compared with those maintaining a GFD for <2 years. Finally, there was no difference in Marsh severity between subjects that reported symptoms at the time of the repeat endoscopy compared with those who did not.
Historically, making the diagnosis of CD required 3 phases and 3 biopsies (13). The current protocol, which requires only 1 biopsy, relies upon the assumption that the vast majority of children with CD experience complete mucosal recovery after 1 year of a GFD (14,15). These guidelines, drafted more than 25 years ago were based on clinical evidence. Over the past several decades there has, however, been a shift to an older age at onset of CD and milder enteropathy at diagnosis. These findings along with the transformation in clinical presentation from almost exclusively gastrointestinal symptoms to a more diverse, systemic presentation provide further evidence that pediatric CD has changed (16–20). Despite this, the need for a follow-up biopsy has not been revisited and the need for an initial biopsy to confirm the disease in pediatric patients has been questioned. Data from adult patients with CD suggest the persistence of enteropathy in >33% of patients on a GFD, irrespective of symptoms or positive serology (7,8). This persistent enteropathy in adults has been associated with an increased risk of lymphoma, low bone density, and fracture (21–23). To date there is no clear understanding of the risk of mortality. In children, the long-term consequences of suboptimal healing are also unclear. Malabsorption and ongoing inflammation in children may have negative repercussions on physical and cognitive development.
In our population, 19% of pediatric patients on a GFD had persistent enteropathy despite treatment on a GFD for at least 1 year. Forty-five percent of patients with persistent enteropathy were asymptomatic at the time of the repeat endoscopy. The majority of asymptomatic patients with CD who underwent a repeat endoscopy did so to assess mucosal recovery due to a different gastrointestinal disorder (ulcer, eosinophilic esophagitis, ulcerative colitis) and duodenal biopsies were taken for completeness. Four patients with seronegative CD at diagnosis underwent a repeat endoscopy to confirm CD remission, 7 others were evaluated to confirm remission of CD and the remaining asymptomatic patients underwent a repeat endoscopy due to persistently elevated tTG. The frequency of persistent enteropathy noted is similar to findings from Ghazzawi et al (24) who found that 15% of pediatric patients with CD had persistent enteropathy on a GFD. The frequency of persistent enteropathy in our work is higher than Bannister et al (25) and Vécsei et al (26) who reported persistent enteropathy in 5% and 9% of pediatric patients with CD on a GFD, respectively. We found that 15.9% of patients with symptoms at the time of the repeat endoscopy had persistent enteropathy, whereas 26% of asymptomatic patients had persistent enteropathy at repeat endoscopy. Characteristics including presence of symptoms at the time of the repeat endoscopy, persistently elevated tTG, and following the GFD for <2 years were not predictive of persistent enteropathy. Poor adherence, particularly in teenagers could be a potential reason for persistent enteropathy especially in patients who do not report symptoms. In this group of patients who underwent repeat endoscopy, adherence to the GFD was, however, evaluated by a dietician or gastroenterologist and 91% of subjects were found to have excellent adherence to the GFD. This high level of adherence may not reflect our general patient population, however. Patients with symptoms or persistently elevated IgA tTG and clearly identified nonadherence/ongoing gluten exposure are unlikely to undergo repeat endoscopy as often as children without clear gluten exposure and would therefore not be included in this analysis.
IgA tTG was not an accurate measure of mucosal recovery in this population of pediatric patients with CD on a GFD. The NPV of IgA tTG was highest at 87% in patients treated with a GFD for >2 years; however, PPV was poor at 36%. In addition, the NPV of tTG was not improved when correlated with the presence or absence of symptoms at the repeat endoscopy. Overall at the time of the follow-up biopsy, tTG was elevated in 43% of subjects with persistent enteropathy and 32% of subjects with mucosal healing. Specifically in subjects that were symptomatic at the time of the repeat endoscopy, 84% had mucosal recovery upon repeat biopsy. Only 55% of patients with persistent enteropathy at the time of the repeat endoscopy were symptomatic. Therefore, in our population, neither the presence of symptoms nor a positive tTG could be relied upon as a measure of mucosal recovery in patients with CD on a GFD.
Despite the clinical practice and endorsement of using serological tests as markers of dietary adherence and mucosal recovery in pediatric patients with CD on a GFD, these serology tests have not been validated for this purpose. Our findings raise concerns about this monitoring approach. Identifying the frequency of persistent enteropathy in children is important to begin to understand term consequences that may be associated with it. Furthermore, identifying minimally invasive accurate surrogate endpoints for patients is of utmost importance to identify patients that may benefit from potential for new therapeutic agents that may serve as adjuvant medications to the GFD (27). Our work is in agreement with previous research that supports a lower accuracy of serological studies at the time of follow-up endoscopy compared to at diagnosis in patients with CD on a GFD. Bannister et al (25) evaluated 150 pediatric patients (mean age 7.5 years) diagnosed with CD who had been on a GFD for a mean of 1.4 years. They found that 5% of patients had persistent mucosal damage defined as Marsh 3 histology. In the present study, IgA tTG and IgG deamidated gliadin peptide (DGP) were measured and used in combination at the time of the repeat endoscopy resulting in a sensitivity of 65% and specificity of 85%, a PPV of 22% and NPV of 98% (25). When the statistical approach was repeated with any equivocal values regarded as “positive,” the NPV improved to 100% but the specificity dropped to 68%. Although this combined approach resulted in a higher NPV, accuracy overall compared with diagnosis was suboptimal. A separate study evaluated 53 children with CD (mean age 11.3 years) who underwent repeat endoscopy after approximately 2.2 years on a GFD (26). The present study found that 9% of patients had persistent mucosal damage defined as Marsh 3 histology at the time of follow-up. In the present study, the sensitivity of IgA tTG was 83% and specificity was 87% (26). These studies provide further evidence that current serological tests used as surrogate endpoints for mucosal recovery need further investigation. IgA tTG in combination with IgG DGP may provide a more accurate measure of mucosal recovery at the time of the repeat endoscopy. The NPV of these tests when used in combination may be quite powerful.
The utility of IgG DGP as a marker of mucosal disease in patients with CD on a GFD was corroborated in a recent comprehensive study evaluating 12 serological tests at the time of follow-up endoscopy in adult patients with CD on a GFD (28). The present study used retrospective samples to examine 100 adult patients with CD who underwent a repeat biopsy after an average of 4.5 years on a GFD. In addition to finding that the manufacturer cutoff levels created for the diagnosis of CD are suboptimal for patients with CD on a GFD, the present study also showed that the IgG DGP correlated with residual intestinal damage as measured by a follow-up small intestinal biopsy in adult patients (28).
Although we did not identify predictors of persistent mucosal damage in this pediatric population, Lebwohl et al evaluated 7648 children and adults over an 8-year period to determine what factors may be predictive of persistent villous atrophy. In the present study, 31% of patients had persistent villous atrophy at the time of the follow-up endoscopy which occurred at a median time of 1.3 years (29). Persistent villous blunting was more common in older patients and males, and less common in patients with higher educational attainment (29).
The limitations of our study include that the IgA tTG assay was run in multiple laboratories due to the retrospective nature. Although not ideal, all laboratories were clinical laboratory improvement amendments certified. For this reason and the purposes of the study, IgA tTG was dichotomized as positive or negative according to the cut off levels supplied by the laboratory performing the test. In addition, we included only serology tests run within 4 months of the repeat endoscopy, which reduced our sample size to 71 subjects for part of the analysis. Because both BCH and MGHfC are quaternary care centers, the results may not be generalizable to other institutions. Studies have suggested that patients referred to specialized centers were 3 times more likely to have nonresponsive celiac disease than those initially followed at the quaternary care center (30). Furthermore, the use of centralized pathology reading, standard biopsy procedures including standardized locations and numbers of biopsies, and evaluating the mucosa for villous height to crypt depth ratio would provide a more accurate measure of mucosal recovery. In addition, there was a wide range of time during which the repeat endoscopy was performed which does not allow us to comment on the natural history of mucosal recovery in pediatric CD. Finally, although all patients were advised about the GFD by a knowledgeable dietician, their adherence assessment was based on physician and or dietician report at the time of the clinic visit in the present study.
We found that 19% of pediatric patients with CD on a GFD may have persistent enteropathy. Although the long-term effects are not known, persistent enteropathy may predispose pediatric patients with CD to future complications and suboptimal growth. Evidence in adult patients with CD is accumulating that not only is persistent enteropathy common but that the presence of symptoms is not predictive of enteropathy. In addition, IgA tTG, whereas accurate at diagnosis, is a poor predictor of persistent enteropathy in children with CD on a GFD irrespective of whether symptoms are present. Although current guidelines do not suggest a repeat endoscopy to assess for mucosal recovery, it is the only way to confirm it. Given the mounting evidence in adults, our concern for persistent enteropathy, particularly in growing children led to this retrospective study. Because of the retrospective nature, we cannot comment on the natural history of mucosal recovery in patients with CD on a GFD as patients did not undergo repeat endoscopy at a predetermined time point. Therefore, these findings must be confirmed prospectively and systematically to evaluate the natural history of mucosal recovery in pediatric patients with CD on a GFD. Long-term complications including growth trajectory must be monitored and further study evaluating available treatment options beyond the GFD and timing of subsequent endoscopies in patients found to have persistent enteropathy must be established. Our findings show that nearly 1 in 5 children with CD may have persistent enteropathy despite maintaining a strict GFD (based on self-report and provider assessment) and that IgA tTG is not an accurate marker of mucosal healing. These findings suggest the need not only for a baseline endoscopy to confirm the diagnosis of CD, but consideration of a repeat endoscopy to confirm remission. These findings suggest that prospective studies, which include objective measurements of dietary transgression are needed to establish whether the management criteria of CD in childhood should be revisited.
1. Farrell RJ, Kelly CP. Celiac sprue. N Engl J Med
2. Hill ID. What are the sensitivity and specificity of serologic tests for celiac disease? Do sensitivity and specificity vary in different populations? Gastroenterology
3. Hill ID, Dirks MH, Liptak GS, et al Guideline for the diagnosis and treatment of celiac disease in children: recommendations of the North American Society for Pediatric
Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr
4. 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
5. Rostom A, Murray JA, Kagnoff MF. American Gastroenterological Association (AGA) Institute technical review on the diagnosis and management of celiac disease. Gastroenterology
6. Rubio-Tapia A, Hill ID, Kelly CP, et al ACG clinical guidelines: diagnosis and management of celiac disease. Am J Gastroenterol
7. Ciacci C, Cirillo M, Cavallaro R, et al Long-term follow-up of celiac adults on gluten-free diet: prevalence and correlates of intestinal damage. Digestion
8. Rubio-Tapia A, Rahim MW, See JA, et al Mucosal recovery and mortality in adults with celiac disease after treatment with a gluten-free diet. Am J Gastroenterol
9. Wahab PJ, Meijer JWR, Mulder CJJ. Histologic follow-up of people with celiac disease on a gluten-free diet: slow and incomplete recovery. Am J Clin Pathol
10. Marsh MN. Gluten, major histocompatibility complex, and the small intestine. A molecular and immunobiologic approach to the spectrum of gluten sensitivity (‘celiac sprue’). Gastroenterology
11. Oberhuber G. Histopathology of celiac disease. Bioméd Pharmacothér
12. Leffler DA, Dennis M, George JB, et al A simple validated gluten-free diet adherence survey for adults with celiac disease. Clin Gastroenterol Hepatol
13. Meeuwisse GW. Diagnostic criteria in coeliac
disease. Acta Paediatr Scand
14. Guandalini S, Ventura N, Ansaldi A, et al Diagnosis of coeliac
disease: time for a change? Arch Dis Child
15. Walker-Smith JA, Guandalini S, Schmitz J, et al Revised criteria for diagnosis of coeliac
disease. Report of Working Group of European Society of Paediatric Gastroenterology and Nutrition. Arch Dis Child
16. Fasano A, Berti I, Gerarduzzi T, et al Prevalence of celiac disease in at-risk and not-at-risk groups in the United States: a large multicenter study. Arch Intern Med
17. McGowan KE, Castiglione DA, Butzner JD. The changing face of childhood celiac disease in North America: impact of serological testing. Pediatrics
18. Roma E, Panayiotou J, Karantana H, et al Changing pattern in the clinical presentation of pediatric
celiac disease: a 30-year study. Digestion
19. Maki M, Kallonen K, Lahdeaho ML, et al Changing pattern of childhood coeliac
disease in Finland. Acta Paediatr Scand
20. Kivelä L, Kaukinen K, Lähdeaho ML, et al Presentation of celiac disease in Finnish children is no longer changing: a 50-year perspective. J Pediatr
21. Catassi C, Fabiani E, Corrao G, et al Risk of non-Hodgkin lymphoma in celiac disease. JAMA
22. West J, Logan RF, Card TR, et al Fracture risk in people with celiac disease: a population-based cohort study. Gastroenterology
23. Walters JRF. Analysis of the absolute risks in coeliac
disease indicates the importance of the prevention of osteoporosis. Gut
24. Ghazzawi Y, Rubio-Tapia A, Murray JA. Mucosal healing in children with treated celiac disease. J Pediatr Gastroenterol Nutr
25. Bannister EG, Cameron DJ, Ng J, et al Can celiac serology
alone be used as a marker of duodenal
mucosal recovery in children with celiac disease on a gluten-free diet. Am J Gastroenterol
26. Vécsei 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 Gastroenterol
27. Gottlieb K, Dawson J, Hussain F, et al Development of drugs for celiac disease: review of endpoints for phase 2 and 3 trials. Gastroenterol Rep
28. De Chaisemartin L, Meatchi T, Malamut G, et al Application of deamidated gliadin antibodies in the follow-up of treated celiac disease. PLoS One
29. Lebwohl B, Murray JA, Rubio-Tapia A, et al Predictors of persistent villous atrophy in coeliac
disease: a population-based study. Aliment Pharmacol Ther
30. Leffler DA, Dennis M, Hyett B, et al Etiologies and predictors of diagnosis in nonresponsive celiac disease. Clin Gastroenterol Hepatol