Society: AGA

BACKGROUND: Perianal fistulizing complications (PFC) of Crohn’s disease (CD) are highly morbid and difficult to treat. Despite the advent of biologic medications, antibiotics remain a mainstay of PFC treatment. The optimal antibiotic regimen for PFC has not been determined, and extent of antibiotic use has not been quantified. To develop benchmarks for antibiotic stewardship, we aimed to characterize antibiotic utilization for PFC in a large multicenter cohort.
METHODS: We identified pediatric patients ≤21yr at CD diagnosis and enrolled in the ImproveCareNow Network (ICN) within 30 days of CD diagnosis (2007-2018), and with ≥3 outpatient visits. ICN is a multicenter, multidisciplinary pediatric IBD quality improvement collaborative. The ICN registry includes prospectively collected data from outpatient pediatric gastroenterology visits. Medications are recorded in the registry only at outpatient visits; therefore, use between visits including hospitalization are unavailable. Antibiotics in the registry include ciprofloxacin, metronidazole, and rifaximin. Duration of antibiotic use was approximated using the number of visits with an antibiotic use. We evaluated the presence of a PFC and timing of medication initiation. For patients with PFC, we only included antibiotics after PFC for maximum 2yr, for non-PFC patients all antibiotics were included. Descriptive statistics are presented.
RESULTS: We identified 5,720 patients from 80 ICN sites (median age 13.5yr, 40.5% female, 18.7% non-White), among whom 1,023 (30.7%) developed PFC. There were no differences in PFC rate by gender, age at diagnosis, race, or insurance. Patients with PFC were more than 2.5x as likely as those without PFC to receive antibiotics (27.5% vs. 11.1%, p<0.001). Most with PFC (58.9%) had multiple visits with antibiotics (up to 18 visits) within 2yr after PFC. Of those with PFC, males were more likely than females to receive antibiotics (29.7% vs 24.2%, p=0.010). There were no differences in antibiotic use by age at diagnosis or race. Patients with commercial insurance were less likely than those without commercial insurance to receive antibiotics after PFC (26.4% vs. 35.3%, p=0.003). Variation in antibiotic use after PFC ranged from 0 to 53.8% across ICN sites. Among patients treated with antibiotics, multiple simultaneous antibiotics (e.g., ciprofloxacin + metronidazole) were used more than 3x as often by those with PFC compared to those without PFC (75.6% vs. 24.4%, p<0.001).
CONCLUSIONS: One in four children with Crohn’s disease are prescribed antibiotics as outpatients after PFC development. Over half have multiple visits with antibiotics within the first 2 years after PFC, and multiple simultaneous antibiotics are used in ¾ of the patients with PFC. There is a large variation in antibiotic use between the centers and antibiotic stewardship in this population is sorely needed.

INTRODUCTION: Eosinophilic esophagitis (EoE) is a chronic, allergen-mediated clinicopathologic disease affecting all ages. While the differences between pediatric and adult onset EoE has been well documented, little is known about the differences within the pediatric onset EoE. Herein, we investigated if clinical features are distinct for early childhood onset EoE (eo-EoE; < 5 years) when compared to late childhood onset EoE (lo-EoE; 5-18 years).
METHODS: We reviewed medical records of 269 children (≤18 years) newly diagnosed with EoE at Monroe Carell Jr. Children’s Hospital at Vanderbilt between May 2017 and October 2022. EoE was defined per the 2011 Consensus Guidelines. Their socio-demographic data, growth parameters, clinical presentation, allergic comorbidities, family history, esophagogastroduodenoscopy (EGD) findings rated per the endoscopic reference scoring system (EREFS), and esophageal histology (peak eosinophil count, and presence or absence of basal zone hyperplasia, eosinophilic microabscess, and lamina propria fibrosis) were extracted for analysis. The eo-EoE and lo-EoE groups were matched for gender, race, and ethnicity. Chi-squared and Fisher Exact tests were used for categorical data and paired t-tests for continuous variables.
RESULTS: In all, 50 children were in the eo-EoE group and 58 children were in the lo-EoE group. The mean (SD) age of the eo-EoE group was 1.86 (1.16) years and the lo-EoE group was 12.4 (3.33) years of age. Z-score for age (weight/length if age < 2 years or body mass index if age ≥ 2 years) was significantly lower for eo-EoE compared to lo-EoE [0.01 (1.48) vs. 0.80 (1.25); P < 0.004]. The eo-EoE group had significantly higher rates of eczema (54.0% vs. 17.2%; P<0.001), weight concerns (36.0% vs. 10.3%; P = 0.002), and feeding difficulties (30.0% vs. 0.0%; P < 0.001) compared to lo-EoE group. On the other hand, children in the lo-EoE group were more likely to present with abdominal pain (3.0% vs. 21.0%; P < 0.001) compared to children in the eo-EoE group. The total EREFS scores were lower in the eo-EoE group compared to the lo-EoE group [1.34 (0.96) vs 2.00 (1.28); P=0.006], and they were less likely to have edema (24.0% vs. 50.0%, P = 0.009). There were no differences between the groups with regards to family history of EoE or atopy and histologic findings.
CONCLUSIONS: Clinical features such as feeding difficulties, atopic co-morbidities, growth faltering, and EoE-relevant endoscopic abnormalities can distinguish children with eo-EoE from lo-EoE even after accounting for gender, race, and ethnicity. These results deepen our understanding of the pediatric EoE. Further investigation is needed to understand why the presentation of eo-EoE differs from lo-EoE and why the endoscopic findings are relatively mild in eo-EoE, despite no difference in histologic findings.

Background: Delayed diagnosis of inflammatory bowel disease (IBD) is associated with prolonged symptom burden and worse long-term outcomes. Racial and ethnic disparities and social barriers are commonplace in healthcare; however, the extent that these factors are associated with diagnostic delay are largely unknown.

Objective: To evaluate whether race, ethnicity, disease type, and social factors are associated with delayed diagnosis of pediatric IBD.

Methods: We recruited pediatric patients diagnosed with IBD in the preceding 6 months at 23 U.S. sites from 2019-2022. Parents/guardians reported race, ethnicity, birth country, gender, household income, insurance type, travel time to treatment center as well as the length of time between symptom onset and IBD diagnosis. Parents/guardians also completed a health literacy measure (the Newest Vital Sign) and the Health Care System Distrust scale. We conducted bivariate and multivariable analyses using generalized estimating equations to evaluate associations between these factors and time to diagnosis defined as ≤ 60 days, 61 to 180 days, 181-365 days and >365 days.

Results: We enrolled 843 patients (mean age at diagnosis 13.2 years, 51% male, 54% Crohn’s disease (CD), 33% ulcerative colitis (UC), 8% Hispanic, 71% non-White). Overall, the mean time to diagnosis was 263 days (median 180). Bivariate associations between measured variables and time to diagnosis are provided in Table 1. After adjustment, associations with longer time to diagnosis included CD vs UC (OR 2.5, 95% CI 1.9-3.3) and longer travel time to clinic [(> 1 hour (OR 1.7, 95% CI 1.2-2.3), > 2 hours (OR 1.6, 95% CI 1.1-2.4) each vs <30 minute]. Race, ethnicity, birth country, gender, parent education, household income, insurance type, health literacy, and health system distrust were not associated with time to diagnosis.

Conclusions: Consistent with prior literature, children with CD experienced longer time to diagnosis than those with UC. Distance to IBD center appears to be a barrier to timely diagnosis, suggesting an urgent need to evaluate innovative models to provide timely and efficient care for communities disproportionately affected by longer travel times. Reassuringly, our findings demonstrate equitable access to IBD diagnostic care across racial and ethnic groups.

Introduction: Environmental enteropathy (EE), a condition associated with recurrent exposure to enteropathogens, is thought to be the primary driver of stunted growth and estimated to globally impact 22% and 7% of children under five years of age. To date, non-invasive EE biomarkers that maintain high accuracy across populations have yet to be identified.
Evidence of human mitochondrial DNA (mtDNA) in feces increasing in concentration during pathogenic (Zhu et al. 2022) and non-pathogenic (Boyapati et al. 2018) states of intestinal inflammation suggests its utility as a measure of intestinal inflammation. However, the extent to which fecal mtDNA may be a useful measure of EE is not known. To investigate this, we first quantified fecal mtDNA across geographic populations with varying prevalence of enteropathogen carriage. Second, we tested associations between fecal mtDNA and EE biomarkers to explore if fecal mtDNA captures novel information.
Methods: We quantified concentrations of fecal mtDNA using hCYTB484, an assay targeting the cytochrome b gene on the human mitochondrial genome, on the BioRad QX200 droplet digital polymerase chain reaction (PCR) and Qiagen QIAcuity digital PCR platforms. Following extraction for nucleic acids using Qiagen 96 Virus and Powerfecal QIAcube HT kits, we analyzed 447 child stools collected from the United States, 120 from Bangladesh, 66 from Mozambique, and 1,586 from Cambodia. Using the 62 stools from Mozambique, we also investigated correlations between concentrations of fecal mtDNA versus current widely used, non-invasive biomarkers of EE: neopterin, alpha-1-antitrypsin, myeloperoxidase, and calprotectin.
Results: We observed increases in concentrations of mtDNA per gram of stool in areas with higher prevalence of enteropathogen carriage in Bangladesh (median log₁₀ = 7.7, interquartile range, IQR = 0.66), Mozambique (median log₁₀ = 8.0, IQR = 0.91), and Cambodia (median log₁₀ = 9.9, IQR = 0.84) versus the US (median log₁₀ = 7.15, IQR = 0.58) (Figure 1). Using a principal component analysis on mtDNA and the EE biomarkers, we found the biomarkers captured similar information to each other, but the mtDNA eigenvector was roughly orthogonal to the biomarker eigenvectors suggesting that information from mtDNA may be distinct from the specific inflammatory responses of the biomarkers (Figure 2).
Discussion: Our results suggesting that fecal mtDNA increases with enteropathogen carriage align with prior work demonstrating increased EE severity with pathogen burden (Jamil et al. 2021). Comparison with current EE biomarkers suggests fecal mtDNA differs from current biomarkers, and may be more generalizable than pathogen-type specific inflammatory responses. Our work suggests value in further study of fecal mtDNA as a geographically agnostic indicator of intestinal inflammation that may provide a novel measurement of EE.

<b>Figure 1.</b> Box and whisker plots of concentrations of fecal mtDNA (hCYTB484) normalized to gram of stools for children living in the United States, Bangladesh, Mozambique, and Cambodia. Distributions show an increase in median concentrations from the US versus low- and middle-income countries.

Figure 1. Box and whisker plots of concentrations of fecal mtDNA (hCYTB484) normalized to gram of stools for children living in the United States, Bangladesh, Mozambique, and Cambodia. Distributions show an increase in median concentrations from the US versus low- and middle-income countries.

<b>Figure 2.</b> Principal component analysis of α-1-antitrypsin (A1AT), neopterin (NEO), myeloperoxidase (MPO), and fecal calprotectin (CAL) measured as mg per mass of stool with human fecal mtDNA (hCYTB484) measured as marker copies per mass of stool from 62 Mozambican children (1-48 months of age). The grouping of the eigenvectors for biomarkers of EE suggest similarity amongst them but the approximate orthogonality of the eigenvector for fecal mtDNA to those of the biomarkers suggests separation from the group.

Figure 2. Principal component analysis of α-1-antitrypsin (A1AT), neopterin (NEO), myeloperoxidase (MPO), and fecal calprotectin (CAL) measured as mg per mass of stool with human fecal mtDNA (hCYTB484) measured as marker copies per mass of stool from 62 Mozambican children (1-48 months of age). The grouping of the eigenvectors for biomarkers of EE suggest similarity amongst them but the approximate orthogonality of the eigenvector for fecal mtDNA to those of the biomarkers suggests separation from the group.

Background:
Between 15% and 40% of patients with Crohn's disease (CD) have perianal manifestations (PCD) at diagnosis. PCD are defined as perianal fistulas, abscesses, fissures, or inflammatory anal skin tags. Perianal fistulizing Crohn’s disease (PFCD): fistulas and/or abscesses, usually dictates a poor prognosis in children. Little is known about factors associated with PFCD in children. The primary aim of this study was to investigate the factors associated with PFCD at the diagnosis of pediatric CD. The secondary aim was to assess factors associated with the severity of PFCD according to the Van Assche score (VAS), a scoring system developed in 2003.

Methods:
We retrospectively collected data from patients (4 to 18 years) diagnosed with CD between 2009 and 2021 in our IBD center. Eligible patients had a magnetic resonance imaging (MRI) performed within 3 months of the diagnosis. PCD were assessed clinically (anal fissure, abscess, inflammatory anal skin tags) and on MRI (abscess, fistula [simple or multiple]). Two radiologists independently evaluated the MRI and graded the perianal disease with the VAS, disagreements were resolved by consensus. Multivariate analyses were performed to compare the patients’ characteristics at diagnosis and PFCD both with multivariate binary logistic regression and linear regression considering the VAS.

Results:
In total, 489 patients were included (57,9% males; median age 13,8 years [interquartile range (IQR) = 11,4-15,5]), of which 216 (44,2%) had any PCD. PFCD was identified with MRI in 115 patients (23,5%) [simple fistula without abscess (15,1%), multiple fistulas without abscess (3,3%), any fistula with abscesses (5,1%)]. The rate of PFCD was higher in patients with clinical PCD (76/177; 42,9%), as compared to those without any clinical PCD at diagnosis (39/312; 12,5%) (p<0,0001).
The median (IQR) VAS for patients with PFCD was 13,0 (9,0-15,0) as compared to 2,0 (0,0-2,0) for patients without PFCD (p<0,0001). Male sex (OR=3,00 [1,37-6,74]; p=0,0062), granulomas on intestinal biopsies (OR=3,30 [1,54-7,08]; p=0,0021) and anal fissures (OR=5,69 [2,13-15,17]; p=0,0005) were associated with PFCD. Granulomas on intestinal biopsies (adjusted beta regression coefficient [aβ]=3,16; p= 0,0188) and anal fissures ([aβ]=3,30; p=0,0310) were associated with higher VAS. The female sex ([aβ]=-2,87, p=0,037) was associated with lower VAS. Age at CD diagnosis, anal skin tags, endoscopic severity, CD location, fecal calprotectin levels and inflammatory blood markers were not associated with PFCD manifestations or severity.

Conclusion:
This study stresses the importance to perform perianal MRI early at diagnosis of CD, as PFCD may be discovered even in asymptomatic patients. Male sex, granulomas on intestinal biopsies and anal fissures were associated with PFCD occurrence and severity.

Background: Disorders of the gut-brain interaction (DGBI) result from a complex interplay of genetic, psychological, biological, and environmental factors, known as the biopsychosocial model of disease. Many patients develop severe disability, and the clinical factors that may be associated with it are not clear. It has been postulated that depression may be more prevalent in patients with DGBIs, but the role it may play in the development of disability has not been well studied.Furhermore the influence that subclinical depression may have on function has not been studied. We aim to establish the association of depression with functional disability in patients with DGBIs.

Methods: This is a review of 334 patients seen in a multidisciplinary functional abdominal pain (FAP) program at a tertiary care center for initial evaluation. Only patients with pain-predominant p-DGBI were selected, including irritable bowel syndrome (IBS), functional abdominal pain (FAP), and functional dyspepsia (FD). Patients completed the Children’s Depression Inventory (CDI) as a measure of depression severity , the Abdominal Pain Index (API) as a measure of pain severity, and the Functional Disability Inventory (FDI) as a measure of functioning in daily life. FDI categories were selected as follows: mild (13-20), moderate (21-30), severe (>=30) disability. Criterion for severe clinical depression is reflected by a CDI score >64, and subclinical depression can be considered with scores of 60-64.

Results:
Patient demographics: 334patients were included. The mean patient age was 13.8 ± 2.3 years. and 51% were female.The mean CDI was 56 ± 11, with 36 (10.7%() having a CDI > 64, and 25 ((7.4%) a CDI >60 <64. The mean FDI was 19.6 ± 11.8 with 111 (33.2%) having modferate to severe disability. The mean API score was 2.7 ± 0.9.

There was a significant correlation between the presence of severe/moderate disability and significant depression (3=0.353; p< 0.001), and higher abdominal pain (r=0.304; p < 0.001. There was no correlation between depression and bdominal pain (p=0.6).

Presence and severity depression: Table 1 shows the prevalence of depression and abdominal pain according to the degree of disability.

Functional disability Across all p-DGBI diagnoses, depression was associated with severe functional disability, as was the presence of more severe abdominal pain

By using logistic regression both the presence of more abdominal pain, and depression were independently associated with moderate to severe disability

Conclusions:
We show that 18.1% of patients with p-DGBIs meet criteria for subclinical or clinical depression.The severity of the depression scores correlated with the presence of functional disability. Subcinical depression was also significanlty associated with disability, indicating that in those children further evaluation and treatment for depression may be necessary.

Background: Crohn’s disease (CD) is a chronic inflammatory disorder of the gastrointestinal tract with long-term detrimental consequences in children. Anti-tumor necrosis factor (aTNF) therapies are the only FDA-approved biologic therapy for pediatric CD, however, one-third of patients fail anti-TNF induction, and, of those who initially respond, 15% relapse every year. Thus, there is an urgency to elucidate the underlying immunogenetic mechanisms of anti-TNF refractory CD. Methods: A prospective cohort study was employed consisting of treatment-naïve (TN-CD; n= 12), aTNF refractory (REF-CD; n=11), and endoscopic remission CD (REM-CD; n=16) along with non-IBD control (n=13) children. High dimensional flow cytometric analyses (HD FACS) of the blood along with ileal and rectal biopsies were performed. Further ileal immune characterization was conducted via single-cell RNA sequencing (scRNA-seq) using the 10X Chromium platform and sequenced on Illumina’s Novaseq600. The Seurat v4.1 workflow was used for the UMAP cell clustering algorithm. To assess effects of conditioned media (CM) from SIRPα cells on non-IBD ileal organoids, blood SIRPα+ cells from REF-CD were enriched via magnetic activated cell-sorting and stimulated with LPS and flagellin (FLG; both at 10 ng/mL) for 96 hrs. The ileal organoids were treated with CM for 72 hrs followed by viability and intracellular cytokine staining. Descriptive statistics, one-way ANOVA with multiple comparisons testing, and univariate linear regression were applied for data analyses. Results: A significant enrichment of SIRPα+CD11c+HLA-DR+CD64+ (SIRPα) cells in the ileum and rectum of REF-CD was observed based on HD FACS (P< 0.05). In the blood, SIRPα cells expressing the gut-homing marker, CD103, were significantly increased in REF-CD (P < 0.05). In-depth characterization with scRNA-seq indicated a monocyte/macrophage lineage based on expression of transcriptional markers: CD14, FCGR3A, CD68, CD163, FCGR1A, and MRC1. Functionally, SIRPα macrophages expressed the pro-inflammatory cytokines, IL-1β, IL-6, and TNFα, detected by HD FACS as well as the pro-fibrotic cytokine, TGFB1, and the neutrophil chemokine, CXCL8. Univariate regression highlighted significant positive correlation of intestinal SIRPα macrophage proportions with fecal calprotectin (P<0.05) and C-reactive protein (P<0.05). The CM from SIRPα+ cells of REF-CD induced significant cell death and increased TNFα and IL-1β expression from ileal organoids. Conclusion: SIRPα cells are a distinct pro-inflammatory monocyte/macrophage population enriched in REF-CD patients that may mediate resistance to anti-TNF therapy via dysregulated secretion of pro-inflammatory cytokines and disrupting the intestinal epithelium. These findings provide insight into pathogenesis of anti-TNF refractory CD in children and highlight these cells to be a therapeutic target.

$Figure 1. Enrichment of pro-inflammatory SIRPa macrophages in children with anti-TNF refractory CD. A) t-SNE clustering analyses of innate immune cells in the ileum and expression of intensity for cell surface proteins and intracellular cytokines. Cells pre-gated on live, MHC-II+CD3-, CD19-, NK1.1- cells. B) Cytokine expression profile of SIRPa cells based on t-SNE clustering. C) Quantitation of SIRPa phagocytes expressing pro-inflammatory cytokines in the ileum (left panel), rectum (middle panel) and blood (right panel). One way ANOVA with Tukey’s multiple comparisons test. D) Single-cell transcriptomic profiles of SIRPa phagocytes based on UMAP projections of gene expression for macrophage transcriptional markers, cytokines, and chemokine.$

Figure 1. Enrichment of pro-inflammatory SIRPa macrophages in children with anti-TNF refractory CD. A) t-SNE clustering analyses of innate immune cells in the ileum and expression of intensity for cell surface proteins and intracellular cytokines. Cells pre-gated on live, MHC-II+CD3-, CD19-, NK1.1- cells. B) Cytokine expression profile of SIRPa cells based on t-SNE clustering. C) Quantitation of SIRPa phagocytes expressing pro-inflammatory cytokines in the ileum (left panel), rectum (middle panel) and blood (right panel). One way ANOVA with Tukey’s multiple comparisons test. D) Single-cell transcriptomic profiles of SIRPa phagocytes based on UMAP projections of gene expression for macrophage transcriptional markers, cytokines, and chemokine.

Figure 2. Intestinal SIRPa+ cells positively associated with inflammation and conditioned-media from SIRPa+ cells induces intestinal epithelial cell death and pro-inflammatory cytokine secretion. A) Univariate regression of fecal calprotectin (top panel) and CRP (bottom panel) with SIRPa macrophage frequencies based on HD FACS. B) Dot plots for cell viability and intracellular cytokine staining for TNFa and IL-1b of trypsinized ileal organoids restimulated with LPS+FLG after no treatment (negative control), or 72-hr incubation with LPS+FLG or conditioned media with or without activated SIRPa+ cells. C) Representative histograms for % cell viability along with % TNFa+ cells and % IL-1b+ cells pre-gated on live cells.

Figure 2. Intestinal SIRPa+ cells positively associated with inflammation and conditioned-media from SIRPa+ cells induces intestinal epithelial cell death and pro-inflammatory cytokine secretion. A) Univariate regression of fecal calprotectin (top panel) and CRP (bottom panel) with SIRPa macrophage frequencies based on HD FACS. B) Dot plots for cell viability and intracellular cytokine staining for TNFa and IL-1b of trypsinized ileal organoids restimulated with LPS+FLG after no treatment (negative control), or 72-hr incubation with LPS+FLG or conditioned media with or without activated SIRPa+ cells. C) Representative histograms for % cell viability along with % TNFa+ cells and % IL-1b+ cells pre-gated on live cells.

Subra Kugathasan

Emory University

Appearances