Investigation of Tissue Transglutaminase Antibody Normalization in Response to Gluten-Free Diet in Children with Celiac Disease

Abstract A very limited amount of data are available regarding the follow-up of celiac disease (CD) treatment in Iran. The aim of this study is to investigate antitissue transglutaminase (atTG) normalization interval and the associated factors in CD patients. This retrospective study included CD patients enrolled in Children's Medical Center, Tehran University of Medical Sciences. The initial atTG titer and histological evaluation (with Marsh grade ≥2) were recorded. The atTG titer was assessed in each follow-up until the time of normalization where children were strictly on gluten-free diet. The age at the time of diagnosis, gender, Marsh grade at the time of diagnosis, other comorbidities, and family history of CD patients were recorded to determine the association of these factors with antibody normalization interval. In total, 71 patients were recruited in the study of which 34 (47.89%) subjects had atTG level below 20 U/mL at the average interval of 31.36 ( ±  2.89) months (95% confidence interval: 25.7–37.02). There was no significant difference between the antibody normalization interval and different age ranges and Marsh grade. Cox regression demonstrated that gender, age ranges, Marsh grade, positive family history of CD, and the presence of comorbidities did not significantly predict longer antibody normalization interval.

accordance with Marsh classification are the first two best practices for the diagnosis of CD. However, combination of TG2-IgA with antiendomysial (EMA) IgA antibody provides a very strong diagnostic accuracy. 8 No simple and reliable test has been reported to be associated with clinical activity and GFD compliance. 9,10 The guidelines suggest a serial antitissue transglutaminase (atTG) evaluation to investigate the response to the GFD. 11,12 Previous studies suggest that the undetectable titer of atTG-IgA is associated with mucosal healing during the follow-up. 13 For evaluating the atTG IgA, enzyme-linked immunosorbent assay has an acceptable sensitivity (90-98%) and specificity (95-97%) for CD. 5,14,15 These antibodies have an important role in diagnosing the newly presented cases with CD; however, their role for the determination of the mucosal recovery is vague. 14, 16 The positive predictive value for the EMA and TTG changes according to the level of the antibodies 17 ; where, higher levels have higher positive predictor values. Additionally, higher levels of antibodies are more sensitive and specific for villous atrophy. The aim of this study is to investigate the atTG normalization interval in CD patients and the associated factors as the predictor of the normalization interval.

Patients and Methods
This retrospective study was performed between March 2018 and September 2018 in Children's Medical Center, Tehran University of Medical Sciences, Tehran, Iran. All known cases of CD, diagnosed at Children's Medical center, were evaluated. The exclusion criteria were defined as the presence of IgA deficiency, less than three follow-ups for atTG titer, nonadherence to GFD (verbal assessment), and diagnosis of CD based on serology and not intestinal biopsy.The intestinal biopsies were performed at a maximum interval of 1 month following the abnormal serology. All biopsies were evaluated by the pathologists who were uninformed about the antibody titer, and the specimens were assessed based on the modified Marsh UEGW criteria. 18,19 Enzyme-linked immunoassay was used to measure the atTG titer by using commercial kit (Orgentec, Germany), where normal antibody titer was defined less than 20 units per milliliter (U/mL). 20 The initial atTG titer and histological evaluation (with Marsh grade of at least 2) were recorded. The atTG titer was assessed in each follow-up until the time of normalization. At the time of diagnosis, these patients initiated GFD. Age and Marsh grade at the time of diagnosis, gender, other medical anomalies, and family history of CD were recorded for each subject to determine the association of these factors with antibody normalization interval. This study was approved by the research ethics board of Tehran University of Medical Sciences.

Statistical Analysis
The quantitative variables were expressed by mean and standard deviation, and number and percentage were used for the qualitative variable. The median time to at TG normalization was calculated by Kaplan-Meier survival analysis (50% patients having normalized results). Log-rank test was used to assess the significance of Kaplan-Meier survival analyses by determining the p-value. The association between the predictor variables and the normalization interval was determined by univariate regressions and the Cox regression model that included the independent variables, such as the age of diagnosis, gender, other medical comorbidities, family history of CD, and Marsh grade at time of diagnosis, to determine the association of predictor factors and atTG antibody normalization interval. A p-value <0.05 was considered statistically significant.

Results
Of the total 556 patients with the high level of atTG, 137 patients were presented with the Marsh score !2 based on the biopsy. Of the 137 patients, 66 subjects had less than three follow-ups for atTG titer or did not adhere to GFD; therefore, 71 patients were eligible to be recruited in the study. In total, 42 (59.15%) of these patients were females, with 25.56% of them with the age more than 10 years (►Table 1). The histological study of the patients revealed Marsh 2 lesion in 11 (15.94%) subjects and Marsh 3 lesion in 58 (84.06%) (►Table 1). In total, 18.3% of the patients in our study had type 1 diabetes, 4 (5.64%) patients had a positive family history of CD, 13 (18.31%) had type 1 diabetes mellitus (DM), and 1 (1.41%) had hepatitis.

Independent Predictors of the Antitissue Transglutaminase Normalization Interval
Cox regression demonstrated that gender (p-value ¼ 0.84), age ranges (p-value ¼ 0.94), Marsh grade (p-value ¼ 0.24), positive family history of CD (p-value ¼ 0.75), and presence of other comorbidities (p-value ¼ 0.43) did not predict longer antibody normalization interval.

Discussion
In this study, the atTG levels were reported as greater or lesser than 20 EU/mL and the exact titer was not determined by the laboratory investigations; therefore, the association between the antibody titers and atTG normalization interval could not be verified. From the patient population, a great   Patients with type 1 DM and CD have a higher atTG titer. 22 These patients have a more complex dietary program and consequently lower compliance. 23,24 In 2003, Liu et al indicated that the level of atTG fluctuated over time in DM 1 and CD-susceptible patients. They indicated that the antibody normalization was seen even with a gluten-containing diet and recommended a higher threshold of the antibody titer for diagnosis of CD. 25 In our study, the mean atTG normalization interval was less than nondiabetics, although it was not statistically significant (25.62 vs. 31.29 months). Parents' awareness and greater diet control in diabetic patients could be the cause of this finding.
Isaac et al reported the median time for normalization of the atTG to be 407 days for 80.5% of patients and 364 days for GFD patients. 22 Hogen Esch et al reported that the cumulative percent of the patients for whom atTG became normal at 6, 12, 18, and 24 months was 35, 55, 64, and 78%, respectively. 26 Current guidelines suggest serial atTG measurements and the normalization could be reached within the 12 months following the GFD. 11,12 The atTG normalization interval in the current study was 31.36 ( AE 2.89) months, which is higher than the previous studies. This is likely to reflect poorer GFD control in these patients.
The gold standard treatment of CD includes GFD, avoiding ingestion of food items that contain gluten. The goal of the treatment is both improving the symptoms and avoiding complications in the future. Avoiding gluten in the populations with gluten-rich diets is considerably difficult and it can affect the quality of life of the patients. Based on the study of Fabiani and Catassi, strict GFD results in declining of the atTG. 27 Ciacci et al found that the atTG level can indicate the GFD compliance and is associated with the presence and intensity of the intestinal damage. 28 The parents' knowledge about the gluten-containing food was considerably limited and they were not aware of the importance of being on GFD, which could be the reason of slower TTG normalization. The GFD compliance routinely was assessed by gastroenterologists by the means of verbal communication. It casts the light on the necessity of being referred to a dietitian to evaluate the GFD compliance tightly; besides giving practical recommendations about the foods containing gluten.

Conclusion
There was no association between Marsh score, age group, and atTG normalization interval, respectively. None of these factors could predict atTG normalization interval in this study. However, studies comparing the normalization time in GFD compliance group and noncompliance group can give better conclusion.

Note
All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.