Repeat Biopsy to Assess Duodenal Healing in Children With Celiac Disease and Eosinophilic Gastrointestinal Disorders

**Background:** Celiac disease (CD) is an immune-mediated enteropathy triggered by gluten, affecting up to 1% of children. The only treatment is a strict gluten-free diet (GFD). While duodenal biopsy is the gold standard for diagnosis, follow-up recommendations for pediatric CD are limited, and many clinicians rely on serologic markers such as tissue transglutaminase immunoglobulin A (tTG-IgA) to monitor mucosal healing. However, the reliability of tTG-IgA to reflect duodenal damage is unclear. This study aimed to assess the correlation between tTG-IgA levels and duodenal histology in children with both CD and eosinophilic gastrointestinal disorders (EGID), a population that routinely undergoes repeat endoscopy for EGID surveillance.

**Methods:** A retrospective chart review was conducted at Children's Hospital of Philadelphia (CHOP) from 2003 to 2018. Patients with International Classification of Diseases codes for both CD and EGID (eosinophilic esophagitis [EoE] or eosinophilic gastritis [EG]) were included. Inclusion required normal IgA levels, positive tTG-IgA at diagnosis, characteristic duodenal pathology, and at least one repeat EGD with duodenal biopsy after starting a GFD. Data collected included duodenal biopsy pathology (villous blunting graded as absent, mild, moderate, or severe; intraepithelial lymphocytes considered increased if >40 per 100 enterocytes), celiac serology (tTG-IgA, EMA-IgA, total IgA), and symptom reports. Duodenal healing was defined as normal villous architecture and no intraepithelial lymphocytes. Statistical analysis used Fisher exact test and t-test.

**Key Results:** Of 132 patients identified, 39 met all criteria (69% male, mean age at diagnosis 7 ± 3 years). At diagnosis, 92% (36/39) were symptomatic, most commonly with abdominal pain (51%) and malnutrition (38%). All 39 had positive tTG-IgA and duodenal villous blunting at initial EGD. At second EGD (mean 324 days after diagnosis), 44% (17/39) had normal duodenal biopsies, indicating no active CD. However, among those with normal biopsies and available tTG-IgA (15 patients), 60% (9/15) had abnormal tTG-IgA levels (P = 1.000). Conversely, 36% (14/39) had negative tTG-IgA at second biopsy, yet 57% (8/14) of these had active CD on biopsy. Among patients with both villous atrophy and intraepithelial lymphocytes (8/39), 25% (2/8) had normal tTG-IgA. Overall, there was no significant correlation between tTG-IgA status and duodenal pathology (P = 1.000). For patients with biopsies taken ≥2 years after diagnosis (18 patients), 94% (17/18) had normal duodenal biopsies and 83% (15/18) had normal tTG-IgA. Among patients with elevated tTG-IgA and normal duodenal biopsies (9 patients), 44% (4/9) had active EoE, 22% (2/9) had gastritis, and 33% (3/9) had no upper GI inflammation. EMA-IgA showed similar discordance: 60% (9/15) of patients with positive EMA-IgA had normal duodenal biopsies, and 12% (2/17) with negative EMA-IgA had abnormal biopsies (P = 0.106).

**Clinical Implications:** This study demonstrates that in children with CD and EGID, tTG-IgA levels do not reliably correlate with duodenal mucosal healing after GFD initiation. Abnormal tTG-IgA may persist despite histologic remission, and normal tTG-IgA may occur despite ongoing duodenal injury. These findings suggest that serologic surveillance alone is insufficient for monitoring CD intestinal healing in this patient cohort. Clinicians should consider repeat duodenal biopsy to confirm mucosal healing, especially in patients with persistent symptoms or abnormal serology. The data also raise the possibility that other inflammatory conditions (e.g., EoE, gastritis) may contribute to elevated tTG-IgA in the absence of active CD. Limitations include the retrospective design, small sample size, and confounding by concomitant EGID. Prospective studies are needed to validate these findings and explore the relationship between tTG-IgA and eosinophilic GI disorders.