Defining normal bowel wall thickness in children with inflammatory bowel disease in deep remission: A multicenter study on behalf of the pediatric committee of the International Bowel Ultrasound Group (IBUS)

Abstract

Objectives

Intestinal ultrasound (IUS) is a noninvasive tool for detecting and monitoring disease activity in children with inflammatory bowel disease (IBD). Remission values for bowel wall thickness (BWT) are extrapolated from adult data. We aimed to define normal BWT in children with IBD in sustained deep remission.

Methods

Multicenter, retrospective pediatric IUS database including children with IBD, without a history of surgery or complications, who underwent IUS after achieving sustained deep remission, defined as the absence of ulcerations on ileocolonoscopy and/or transmural healing on magnetic resonance enterography and steroid‐free clinical remission for ≥6 months. Univariate and multivariable analyses were performed using mixed‐effect modeling.

Results

Ninety‐eight children (40 [41%] female; 64 [5%] Crohn's disease [CD]: 29 (30%) ulcerative colitis [UC]: 5 (5%) IBD‐unspecified), median age 15.2 [13.8–16.8] years), and 484 bowel segments were included. Median BWT was 1.4 (1.1–1.6) mm, with no difference in BWT between bowel segments (p = 0.28). In multivariable analysis, a positive association remained between BWT and weight (univariate β = 0.004 [95% confidence interval, CI = 0.001–0.007], p = 0.016, multivariate β = 0.006 [95% CI = 0.001–0.011], p = 0.041). A negative association remained between BWT and disease duration (unadjusted β = −0.024 [95% CI = −0.047 to −0.001] p = 0.046, adjusted β = −0.028 [95% CI = −0.052 to −0.003], p = 0.039).

Conclusion

Previously inflamed BWT for children with IBD in sustained deep remission was less than 2.5 mm, which is less than 3 mm in adults, and unaffected by age, sex, and bowel segment in this population. BWT may be affected by weight and disease duration. These findings are crucial to the standardized assessment of transmural healing in children.

What is Known

• Intestinal ultrasound IUS) is a noninvasive, patient‐centric tool for monitoring disease activity in children with inflammatory bowel disease (IBD).

• Bowel wall thickness (BWT) is the most important parameter for assessing IBD activity.

What is New

• Previously inflamed BWT for children with IBD in sustained deep remission is less than 3 mm seen in adults, and unaffected by age, sex, and bowel segment in this population.

• BWT may be affected by weight and disease duration.

• These findings are crucial to the standardized assessment of transmural healing in children.

1. INTRODUCTION

Crohn's disease (CD) and ulcerative colitis (UC) are inflammatory bowel diseases (IBDs), resulting in inflammation that can extend beyond the mucosa. ¹ , ² Current IBD management strategies in children require accurate, noninvasive, monitoring of inflammation to guide treatment optimization with the goal of healing the bowel. While the gold‐standard assessment for evaluating healing is ileocolonoscopy, this alone fails to assess the transmural nature of IBD, is often not performed in a repeated manner by pediatric gastroenterologists, and when performed, is imperfect, with lower rates of ileal intubation than in adults. ³ , ⁴ Intestinal ultrasound (IUS) has emerged as a real‐time, noninvasive tool for detecting transmural healing in children with IBD. ⁵ Whereas there is a growing body of evidence for the accuracy of IUS to detect both mucosal and transmural inflammation in children with IBD, data on the ability of IUS to detect transmural healing remains limited. ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴

Healthy children without IBD have been found to have a similar segmental bowel wall thickness (BWT), the most important measure of transmural disease activity, as healthy adults. In addition, they have a lower BWT than children with CD, and a BWT that typically measures less than 2 mm. ¹⁵ However, there is a significant difference between colonic BWT in children ages 0–4 years compared to adolescents aged 15–19 years. ¹⁶ Compared to limited studies in healthy children, there have been no single or multicenter studies examining the IUS findings in children with IBD who achieved clinical remission and either mucosal and/or transmural healing on additional cross‐sectional imaging modalities. Therefore, we aimed to examine IUS findings in children with IBD in sustained deep remission, those who achieved steroid‐free clinical remission with objective evidence of complete healing based on the current gold‐standard assessments of ileocolonoscopy and/or magnetic resonance enterography (MRE) in an international, multicenter cohort.

2. METHODS

We conducted an international, multicenter, retrospective cohort study including children with IBD ≤ 18 years of age undergoing IUS after achieving sustained deep remission, defined as the absence of ulcerations and a simple endoscopic score (SES‐CD) ¹⁷ of 0 or Mayo Endoscopic score ¹⁸ of 0 on ileocolonoscopy and/or transmural healing on MRE and steroid‐free clinical remission from September 2020 to July 2023. Transmural healing on MRE was defined as the absence of bowel wall thickening, hyperenhancement, and associated features suggestive of inflammation (i.e., mesenteric inflammatory fat and lymphadenopathy). Sustained steroid‐free clinical remission was defined as a Pediatric Ulcerative Colitis Activity Index ¹⁹ or a Pediatric Crohn's Disease Activity Index ²⁰ <10 or a Physician Global Assessment of remission without steroid use (systemic and local formulations) for six or more consecutive months before IUS being performed. IUS was performed within 6 months of ileocolonoscopy and/or MRE. Data was collected from the electronic medical record, including patient demographics and disease‐specific factors such as disease duration, disease phenotype (Paris classification), ²¹ medications, and surgical history. Patients were excluded if they had a history of IBD‐related surgery, complications (stricture, abscess, fistula, stenosis, and inflammatory mass) and/or documented steroid use (topical or systemic) within 6 months of IUS. All patients had a noncomplicated, inflammatory phenotype for those with CD.

2.1. Ethics statement

Institutional review board approval was obtained at each site, including Mount Sinai Hospital, Alberta Children's Hospital, Children's Hospital of Los Angeles, and the Stollery Children's Hospital.

2.2. Outcomes

The primary outcome was median segmental BWT on IUS for: sigmoid colon (SC), descending colon (DC), transverse colon (TC), ascending colon (AC), and terminal ileum (TI). Measurements of the rectum were not included as measurement acquisition is not standardized and can be impacted by bladder filling. Secondary outcomes included: associations between BWT and age, sex, height, weight, disease duration, diagnosis, disease location, and medications. Furthermore, we assessed the difference in BWT between segments that had been previously inflamed and now healed compared to those that were never inflamed.

2.3. IUS

IUS examinations were performed on Samsung RS85 Prestige (South Korea), Canon Aplio 300 (Japan), or Philips EPIQ5 (USA). IUS examinations were performed as part of the standard of care by one of five certified International Bowel Ultrasound (IBUS)‐trained pediatric gastroenterologists. No bowel preparation or fasting was required before IUS. A standardized approach was employed for all examinations utilizing both low‐frequency transducers (ranging from 3 to 12 MHz) for global assessment and mid‐to‐high frequency linear transducers (2–14 MHz) for detailed assessment and BWT measurements. Segments of the large intestine and TI were examined using a gray scale and color Doppler evaluation. Maximum BWT was assessed in each segment in the longitudinal plane. Although not a parameter of interest for this study, bowel wall hyperemia (BWH) was assessed as part of standard assessment: confirmation of absence of color Doppler signal, with a velocity rate of 4–7 cm/s was documented. The absence of additional transmural characteristics of inflammation was documented; this was defined as the absence of mesenteric inflammatory fat, preservation of bowel wall stratification, absence of reactive mesenteric lymphadenopathy, and absence of complications. Healed segments that were previously inflamed were determined based on the Paris classification. ²¹ For example, in patients with CD with isolated ileitis (L1), the TI was classified as previously inflamed, while the entire colon was classified as normal. If the previous inflammatory status could not be ascertained, the previous inflammatory status was left as unknown and not analyzed. Therefore, patients with colonic CD were classified as having a normal TI, but colonic segments were marked unknown, and patients with ileocolonic CD were classified as unknown, as the degree of involvement of the colon and ileum could not be estimated.

2.4. Statistical analysis

Descriptive statistics were used to calculate median values and interquartile ranges (IQRs) for continuous variables and frequencies with percentages for categorical variables. The distribution of continuous variables was evaluated and described using histograms. The difference between median segmental BWT measurements was assessed using the Kruskal–Wallis test. We compared segmental BWT between previously inflamed segments and non‐inflamed segments using the Wilcoxon rank‐sum test.

The relationship between BWT and patient variables of age, sex, height, weight, disease duration, diagnosis, disease location, and medications was assessed by univariable and multivariable analysis. Univariate analysis was performed using a mixed‐effect model linear regression, assessing single individuals as the intercept of the random effect to account for the clustering of each bowel segment measurement within individuals. This method was also used to assess for differences in BWT across bowel segments at the individual subject level. We performed a sensitivity analysis on the univariate analysis by adjusting each univariate model with the variable colonic segment as a fixed effect. This analysis was retained if it demonstrated a change (increase or decrease) of 10% or more in the beta coefficient in the variable of interest. Furthermore, a likelihood ratio test (LRT) was used to compare nested models from the original analysis to the sensitivity analysis model by evaluating whether the more complex model provided a significantly better fit to the data than the simpler model, based on the difference in their log‐likelihood values, assessed against a chi‐square distribution. The sensitivity analysis model was considered to provide a better fit if the p value was less than 0.05. Multivariable analysis was performed using multivariable linear regression and mixed‐effect modeling to assess if the associations noted in univariate analyses hold when accounting for all variables together. Similarly, a sensitivity analysis was performed, adding the colonic segment variable as a fixed‐effect variable in the multivariate model. Variable selection was not performed, as we assessed for associations rather than prediction modeling. Due to collinearity between IBD diagnosis and IBD disease location per Paris classification, ²¹ only the former was kept in the final association model. Model assumptions for the simple linear models, multilinear models, and mixed‐effect models were assessed using residual plots and Q–Q plots, where appropriate. Statistical significance was determined with an alpha of 0.05. All statistical analysis was performed using RStudio (version 2023.12.1+402; RStudio Team (2020). RStudio, Integrated Development for R. RStudio, PBC, Boston, MA, USA, http://www.rstudio.com).

3. RESULTS

3.1. Patient demographics

A total of 98 patients met the inclusion criteria. From their respective IUS examinations, 484 out of 490 (98.8%) bowel segments were available for analysis. The TI was the most commonly missed segment (4/6), while the AC and TC were each missed once. Forty (40.8%) patients were female with a median age of 15.2 (IQR = 13.8–16.8) years and a median disease duration of 2.6 (IQR = 1.2–4.3) years. Only four patients in the cohort were under age 10, with the youngest patient being 7 years old. Most patients had a diagnosis of CD; of the 64 (65.3%) with CD, 29 (47%) had ileal disease, 8(14%) had colonic, and 27 (39%) had ileocolonic disease phenotype. All patients had an inflammatory phenotype. For 34 children with UC, 21 (62%) had pancolitis, 7 (20%) had extensive colitis, and 6 (18%) had left‐sided colitis. Anti‐TNF monotherapy was the most common therapy in 41 (42%) patients, anti‐TNF combination therapy in 10 (10%), ustekinumab in 19 (19%), vedolizumab in 7 (7%), mesalamine in 11 (11%), immunomodulator monotherapy in 7 (7%), Upadacitinib in 1 (1%) and dual therapy in 1 (1%). One patient was on no therapy (1%). A summary of demographic and disease characteristics is provided in Table 1.

Table 1.

Baseline patient demographics.

Subjects	N (%), or median (IQR)	Range (min–max)
Female	40 (40.8)
Age (years)	15.2 (13.8–16.8)	7.0–20.0
Height (cm)	164.8 (157.0–172.7)	119.2–186.0
Weight (kg)	56.6 (46.4–66.4)	22.1–115.5
Disease duration at time of IUS (years)	2.6 (1.2–4.3)	0.5–11.9
CD:UC:IBD‐U	64 (65.3):29 (29.6):5(5.1)
CD lower gastrointestinal involvement (n = 64)
L1: ileal	29 (47)
L2: colonic	8 (14)
L3: ileocolonic	27 (39)
UC/IBD‐U disease location (n = 34)
E1: ulcerative proctitis	0 (0)
E2: left‐sided colitis	6 (17.6)
E3: extensive colitis	7 (20.6)
E4: pancolitis	21 (61.8)
Therapy
Anti‐TNF monotherapy	41 (41.8)
Anti‐TNF combination therapy	10 (10.2)
Ustekinumab	19 (19.4)
Vedolizumab	7 (7.1)
Mesalamine	11 (11.2)
Immunomodulator monotherapy	7 (7.1)
Upadacitinib	1 (1.0)
Dual biologic therapy	1 (1.0)

Abbreviations: CD, Crohn's disease; IBD‐U, inflammatory bowel disease‐unclassified; IQR, interquartile range; IUS, intestinal ultrasound; TNF, tumor necrosis factor; UC, ulcerative colitis.

3.2. BWT

Overall median segmental BWT was 1.3 mm (IQR = 1.1–1.6 mm). Median BWT was 1.4 mm (IQR = 1.1–1.8 mm) in the SC, 1.4 mm (IQR = 1.1–1.8 mm) in the DC, 1.3 mm (IQR = 1.0–1.6 mm) in the TC, 1.3 mm (IQR = 1.1–1.5 mm) in the AC and 1.3 mm (IQR = 1.0–1.6 mm) in the TI. The overall range of BWT across all bowel segments was 0.5–3.5 mm. There were no significant differences in median segmental BWT between bowel segments (p = 0.28) (Figure 1). The distribution of BWT measurements across all segments was positively skewed on the histogram, with a peak between 1.0 and 1.5 mm (Figure S1). Histograms for segmental BWT measurements had a similar peak and distribution. While 433 out of 484 (90%) BWT measurements were under 2 mm, 477 out of 484 (99%) bowel segments measured <2.5 mm. The mean BWT in this sample was 1.37 mm, with the population mean within the 95% confidence interval (CI) of 1.33–1.40 mm.

Figure 1.

Segmental bowel wall thickness. Boxplot represented in the style of Tukey, with the horizontal line representing the median, the end of the box representing IQR, upper whisker representing the upper fence (Q3 − 1.5*IQR), and lower whisker representing the lower fence (Q1 − 1.5*IQR), with outliers noted as single points above or below these parameters. IQR, interquartile range.

There were 322 out of 490 (66%) bowel segments that could be ascertained as previously inflamed segments or uninvolved segments from disease distribution data, with 146 out of 322 (45%) as previously inflamed. In the previously inflamed segments, 29 out of 146 (20%) were from patients with CD, 98 (67%) were from UC, and 13% were from those with IBD‐U. In the uninvolved segments, 124 out of 176 (70%) were from patients with CD, 47 (27%) from patients with UC, and 3% from patients with IBD‐U. 319/322 bowel segments had BWT measurements available for summary analysis. When evaluating bowel segments considered previously inflamed, there was no significant difference in BWT between previously inflamed segments (median = 1.35 mm, IQR = 1.10–1.60 mm) and uninvolved segments (median = 1.30 mm, IQR = 1.01–1.60 mm, p = 0.625). However, previously inflamed TI had higher BWT (median BWT = 1.45 mm (IQR = 1.23–1.85 mm) compared to uninvolved TI (median = 1.13 mm, IQR = 1.00–1.42 mm, p = 0.003). There was no significant difference in segmental BWT between previously inflamed and uninvolved SC, DC, TC, and AC, summarized in Table 2.

Table 2.

Comparison between segmental BWT in previously inflamed segments and uninvolved bowel segments.

Bowel segments (N = 319)	Previously inflamed segments median BWT (IQR) (N = 145)	Uninvolved segments median BWT (IQR) (N = 174)	p
All segments (mm) (N = 319)	1.35 (1.10–1.60)	1.30 (1.01–1.60)	0.625
Terminal ileum (mm) (N = 69 [28:41])	1.45 (1.28–1.85)	1.13 (1.00–1.42)	0.003
Ascending colon (mm) (N = 62 ([21:41])	1.35 (1.1–1.50)	1.30 (1.10–1.50)	0.682
Transverse colon (mm) (N = 62 [28:34])	1.21 (1.00–1.68)	1.35 (1.03–1.60)	0.810
Descending colon (mm) (N = 63 [34:29])	1.31 (1.11–1.56)	1.60 (1.10–1.90)	0.053
Sigmoid colon (mm) (N = 63 [34:29])	1.30 (1.03–1.49)	1.40 (1.10–1.77)	0.605

Note: Total segments are provided as N, with number of segments affected previously: number of segments not affected previously.

Abbreviations: BWT, bowel wall thickness; IQR, interquartile range.

When comparing BWT between bowel segments at the subject level, the DC was, on average, 0.12 mm thicker than the AC in each individual, a difference that was statistically significant (95% CI = 0.018–0.226 mm, p = 0.023).

3.3. Relationship between BWT and clinical features

In univariate analysis (mixed‐effect model), there was a significant positive association between BWT and age (unadjusted β = 0.023 [95% CI = 0.001–0.045], p = 0.039), and weight (unadjusted β = 0.004 [95% CI = 0.001–0.008], p = 0.014). There was a significant negative association between BWT and disease duration (unadjusted β = −0.024 [95% CI = −0.047 to −0.001], p = 0.046]. No other parameters were significantly associated with BWT in the univariate analysis (Table S1). In the sensitivity analysis, accounting for the colonic segment as a fixed effect, the results demonstrate that adjusting for colonic segments had a minimal effect on the associations between the predetermined variables and BWT (Table S2). An LRT was also performed to compare models fit when analyzing the colonic segments. The results from the LRT indicated only a marginal difference in model fit when including colonic segments versus not.

In the multivariable analysis (mixed‐effect model) (Figure 2), a significant positive association remained between BWT and weight (adjusted β = 0.006 [95% CI = 0.001–0.011], p = 0.038] and a signifcant negative association between BWT and disease duration (adjusted β = −0.028 [95% CI = −0.052 to −0.003], p = 0.038). There was no significant association between BWT and any other parameters in the multivariable analysis. Results of univariate and multivariable analyses are summarized in Tables S1 and S2. Similarly, adding colonic segments as a fixed effect had a minimal effect on the associations between the model parameters and BWT. An LRT indicated only a marginal difference in model fit when including colonic segments versus not.

Figure 2.

Association between bowel wall thickness and patient parameters. Values are beta coefficients that summarize the estimates from univariate (red) mixed model multivariable linear regression (blue) for the relationship between bowel wall thickness and patient parameters. Significant variables are highlighted with an asterisk. IBD‐U, inflammatory bowel disease‐unclassified.

4. DISCUSSION

Our international multi‐center, retrospective study demonstrated that healed BWT on IUS for children with IBD in sustained deep remission is less than 3 mm, which is the consensus definition of transmural healing in adults and is unaffected by age and sex. Our findings suggest that weight and disease duration may affect BWT. Larger‐scale studies are needed to better understand this relationship, considering that only the association with weight remained significant in the sensitivity analysis.

In our cohort, median BWT was 1.4 mm with no difference in BWT between bowel segments, which is less than the 3 mm cutoff utilized in adult IBD patients. ²² Healthy children without a diagnosis of IBD have been found to have a segmental BWT less than 2 mm. ¹⁵ In a systematic review examining the diagnostic accuracy of IUS in children with IBD, BWT values ranged from 1.5 to 4 mm, and many studies described different values for small bowel versus large bowel. ¹⁰ This study was also not limited to patients in deep remission. Our study aligns with the literature, in that 99% of bowel segments analyzed in children with IBD in sustained deep remission were less than 2.5 mm. Our study highlights that in children with UC and IBD‐U, the median BWT of previously inflamed segments is no different than those that were never inflamed. However, the TI of children with small intestinal CD is significantly thicker than that of patients who never had TI disease, although this 0.3 mm is unlikely to be clinically significant. Although median BWT in our cohort was 1.4 mm (IQR = 1.1–1.6 mm), it is important to note that these children were in deep remission and further studies are needed to evaluate the role of patient and disease characteristics on bowel wall thickness, however, it is important to highlight that pediatric patients with IBD in deep remission had BWT 1.4 mm, and thus, utilizing the adult cutoff of 3 mm is not appropriate for the pediatric population. BWT is a continuous variable, and a cutoff is determined to optimize the sensitivity and specificity of determining the presence of information. To our knowledge, this is the largest cohort to report BWT values in children with IBD who have sustained deep remission, and it provides a baseline as to how we monitor children with IBD, standardizing our assessment of transmural healing.

In our cohort, BWT was associated with weight. The relationship between weight, age, and pubertal stage is complex. The ranges for normal BWT in children vary in the literature and have been shown to increase slightly with age; however, age in isolation does not necessarily reflect growth and pubertal development. BWT on IUS was not significantly affected by age in our cohort, although it is important to note that the median age of our cohort was 15.2 years, with no children less than 6 years of age. Thus, these findings may not be generalizable to younger patients and are an area for future research. The role of pubertal stage and BWT has yet to be investigated and given there was an association between weight, but not age or height, this may suggest the relationship is more nuanced. Additionally, disease duration was negatively associated with BWT. BWT is increased by active inflammation but can also be impacted by fibrosis and muscular hypertrophy, ²³ , ²⁴ which are significant considerations if the disease duration is extensive. The duration of the median disease in our cohort was only 2.6 years, so it is unknown whether significant scar tissue may have accumulated during this time. The negative association between BWT and disease duration may be attributable to early achievement and maintenance of transmural healing.

Our study has several limitations. Although multicenter, it is retrospective, and ultrasonographers were not blinded to patients' history or symptoms when performing ultrasound assessments. We allowed for a maximum of 6 months between IUS and endoscopy or MRE, which can introduce potential for change in disease status, however, there were no change in therapy, symptoms or biomarkers in those patients and if this were to occur it would likely bias toward the null effect. Without comparison to patients with active disease, or subgroup analysis by disease phenotype, it is not possible to define the upper range of normal, and the predictive value of bowel wall thickness is beyond the scope of this study. The majority of patients were older than 10 years of age, and there was no equal representation of children in each age group to facilitate subgroup analysis. Given the retrospective nature, we did not have information regarding pubertal status, and this will be an important parameter to evaluate in future studies. Given that all patients with Crohn's disease had an inflammatory phenotype, we cannot generalize the data to patients with stricturing/penetrating disease. Additionally, our cohort consisted predominately of CD patients in comparison to UC and several studies have demonstrated variance in BWT based on disease phenotype. ¹⁵ Although there was no significant difference in BWT values by bowel segment in pediatric patients with sustained clinical and transmural healing, larger studies are needed to assess if there are differences in baseline values on the basis of disease distribution and phenotype.

5. CONCLUSION

In summary, in our cohort of children with IBD in sustained deep remission, summary data demonstrated that previously inflamed BWT was less than 2.5 mm in 99% of bowel segments. This is less than the 3 mm cut‐off for adults, and is unaffected by age, sex, and bowel segment. BWT may be affected by weight, and disease duration, however, further studies are needed to assess the complex interplay between anthropometric parameters, pubertal status and BWT as well as potential differences in BWT values in relation to disease distribution, phenotype and duration. With the evolution of IBD monitoring strategies and the emphasis on noninvasive modalities to accurately assess disease activity, these findings are key in setting a new standard for assessment of transmural healing in children with IBD.

CONFLICT OF INTEREST STATEMENT

Mallory Chavannes: Reports research funding from the BMS foundation. Hien Q. Huynh: Speaker fees for Janssen, Organon, Sanofi, and Abbvie. Advisory board member for Sanofi, AbbVie, and Celltrion. Jennifer DeBruyn: Consulting fees are from AbbVie and Celltrion; conference funds are from Janssen. Michael T. Dolinger: Consulting fees from Neurologica, a subsidiary of Samsung Electronics Co., Ltd, Pfizer, and BMS. The remaining authors declare no conflicts of interest.

Supporting information

Supplemental Figure 1.

Supplemental Table 1.

Supplemental Table 2.

Kellar A, Chavannes M, Huynh HQ, et al. Defining normal bowel wall thickness in children with inflammatory bowel disease in deep remission: a multicenter study on behalf of the pediatric committee of the International Bowel Ultrasound Group (IBUS). J Pediatr Gastroenterol Nutr. 2025;81:53‐61. 10.1002/jpn3.70049