Materials and methods
Search strategy
We conducted a systematic search with help of a clinical librarian (F.S.E.-J.) in the PubMed, Embase (Ovid), Cochrane Library, and CINAHL (EBSCO) databases for studies describing bowel wall thickness measured by transabdominal US in healthy subjects aged 0–18 years. We excluded studies using contrast agent, studies only describing the appendix and studies whose full text was unavailable. Additionally we excluded articles not written in English, French, German, Spanish, Italian or Dutch. We did not restrict our search to a certain period of time.
The search terms are shown in the supplementary material. The titles and abstracts of the articles retrieved using the search strategy were screened independently by two reviewers (E.A.W., F.A.E.V., each with 4 years of experience in bowel ultrasound) to identify potentially eligible studies. The same reviewers then retrieved full texts of these potentially eligible studies and independently assessed them for eligibility. Any disagreements were resolved through discussion with a third reviewer (B.G.P.K. with 20 years of experience in pediatric gastroenterology).
We used a standardized piloted form to extract data from the included studies and to assess methodological quality. Extracted information included number of patients, demographic details, study design, location of participant recruitment, definition of “healthy” as defined by authors, US technique (brand, probe, bowel preparation, method of bowel wall measurement) and bowel wall thickness per segment (jejunum; ileum; cecum; ascending, transverse and descending colon; rectum) per age category in millimeters (mm). Age was categorized as follows: 0–4 years, 5–9 years, 10–14 years and 15–18 years, based on an earlier study [
3].
Methodological quality
To assess methodological quality we used the Checklist for Cross-Sectional/Prevalence Studies from the Agency for Healthcare Research and Quality Methodology [
4] and added three questions deemed relevant by the reviewers, based on recommendations from the
Cochrane Handbook for Systematic Reviews on Diagnostic Test Accuracy [
5]: “Did test operators have appropriate training?” and “Was ultrasound technique described properly?” and “Was definition of healthy clearly described?” Methodological quality was independently assessed by two reviewers (E.A.W. and F.A.E.V.).
Data analysis
To calculate the mean bowel wall thickness per segment per age category over studies, we performed a meta-analysis. Studies that used the standard way of measuring bowel wall thickness (from the serosa/muscularis propria interface to the mucosa/lumen interface) were included in the meta-analysis. The sample-size weighted pooled mean and pooled standard deviation (SD) scores were calculated with Excel version 2016 (Microsoft, Redmond, WA). We first assessed the differences among the age categories with ANOVA. If a significant difference was found, we further investigated with subsequent Student’s t-tests for independent samples, corrected for multiple testing with the Bonferroni method using GraphPad Prism® version 7 (GraphPad, San Diego, CA). First, we investigated differences between consecutive age categories, and if no significant differences between consecutive age groups were found, differences between other age categories were analyzed. In the Results sections, only the corrected P-values are presented.
Discussion
In this systematic review we assessed the bowel wall thickness as measured with ultrasound in healthy children. Obtaining reference values from a healthy population is of great importance because the role of bowel US in children is rapidly increasing [
12]. We found that the reported values of bowel wall thickness in healthy children range from 0.8 mm to 1.9 mm in small bowel and from 1.0 mm to 1.9 mm in the colon, when measuring from the serosa/muscularis propria interface to the mucosa/lumen interface. Although all included studies had some methodological flaws, these values can be used as guidance in clinical practice when screening children suspected of having bowel pathology, especially inflammatory bowel disease.
In this systematic review, we also found a difference in bowel wall thickness among pediatric age categories: colonic bowel wall thickness was larger in older compared to younger children. The differences between the youngest age groups (0–4 years) and the oldest age groups (15–19 years) ranged from 0.3 mm to 0.5 mm. This is in line with a study in 122 healthy adults aged 23–79 years [
11] that also found a positive correlation between age and bowel wall thickness (r=0.069,
P=0.003). This raises the question whether ultrasonographers should use different cut-off values for different age categories.
A study in children who were newly diagnosed with Crohn's disease (aged 9–18 years) reported an ileal bowel wall thickness of 5.6±1.8 mm [
13]. In addition, a study in children aged 2–18 with active ulcerative colitis reported colonic bowel wall thickness values of >3 mm [
14]. Hence, the small difference in bowel wall thickness between older and younger children is probably not clinically relevant in the diagnosis and follow-up of children with inflammatory bowel disease, also because children with inflammatory bowel disease are usually in their teens [
15]. However, for children with early onset inflammatory bowel disease this needs to be confirmed because there are no data on US findings in this patient group.
The relevance of this age-related bowel wall thickness difference for the diagnostics in other causes of enterocolitis, like allergic or infectious causes, is unclear because there is a scarcity of data on US findings in these disorders. Interestingly, bowel wall thickness in the small bowel was larger in children aged 0–4 years compared to children aged 5–9 years, in both the jejunum and the ileum. Also, the weighted pooled SD in the ileum was quite high (0.6 mm) in the children aged 0–4 years. Two of the included studies also reported the presence of mesenteric lymph nodes, free fluid and increased mesenteric vascularity in the ileal and jejunal regions in children aged 0–1 years [
7,
10]. This implies that among infants the small bowel wall is variable because of changes in lymphoid tissue in the Peyer patch associated with immunologic maturation and that reference values for small bowel wall thickness in this group of children have a wide range, possibly affecting the accuracy for individual patients. However, the small study populations of the included studies should be taken into account when interpreting these results.
The studies included in this systematic review used different approaches for bowel preparation. Some experts advise not to use any preparation, or merely to take in non-carbonated fluid 30 min before the US examination [
16], while a recent consensus statement of the European Society of Paediatric Radiology (ESPR) and European Society of Gastrointestinal and Abdominal Radiology (ESGAR) states that children should not eat any solid food or drink carbonated fluid or milk for 2–6 h before bowel US exam, based on expert opinion [
17]. Nylund et al. [
11] compared bowel wall thickness measured after overnight fasting to bowel wall thickness measured 30 min after eating a 300 Kcal meal in 23 healthy adults and reported a small increase of bowel wall thickness in the terminal ileum (change from 1.1±0.2 mm to 1.2±0.2 mm,
P<0.05) and sigmoid colon (change from 1.2±0.3 mm to 1.4±0.4 mm,
P<0.05) [
11]. Although the second measurement was not blinded and this is a small difference, it seems advisable to standardize bowel preparation protocols, especially in research settings. The same study compared bowel wall thickness measured with 8-MHz transducers to 12-MHz transducers using mixed linear model analysis and found a small influence of transducer type, with lower bowel wall thickness measurements when using the 12-MHz transducer (−0.05 mm,
P<0.001). In this systematic review the included studies used different types of transducers, which is most likely explained by the year in which the studies were conducted; older studies used lower-frequency transducers. We do not think that the currently presented results are influenced by this small difference of 0.05 mm, but for future studies on bowel wall thickness, it would be advisable to uniformly use high-frequency transducers to minimize measurement variation.
This systematic review shows that all studies on bowel US in healthy children have some methodological flaws. First, the methodological quality of most included studies was unclear or low on important features of the methodological quality assessment. Examples of this are the unclear protocols for missing data and unclear or absent measures for quality assurance, such as intra-operator reliability analyses. Another limitation of the included studies is the small number of patients in the age categories 5–9 years, 10–14 years and 15–19 years. Only one of the included studies reported the bowel wall thickness in these age categories [
3], and the others either included only infants [
7,
10] or presented only the results for all participants together, regardless of their age [
2,
6,
8,
9]. In addition, not all studies used a clear definition of “healthy children,” although most studies did report an absence of gastrointestinal symptoms.
To generate reference values, future studies should be strict on inclusion and exclusion criteria and use a clear definition of “healthy,” use protocolled bowel preparation and uniformly use high-frequency probes. Bowel wall thickness should be measured separately in each segment and in different age categories, whereby it would be worth a consideration splitting the youngest age groups, taking into account the results of the studies in infants presented in this systematic review. In addition, it would be of value to investigate the presence of other ultrasonographic markers of inflammation in healthy children, such as increased vascularity of the bowel wall, and presence of lymph nodes.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.