Clinical analysis and identification of pediatric patients with colonic ulceration

Abstract

Background

A wide variety of diseases mimic inflammatory bowel disease (IBD). This study aimed to reduce the misdiagnosis among children with colonic ulcers.

Methods

Eighty-six pediatric patients with colonic ulcers detected by colonoscopy were enrolled in the retrospective study. Children were divided into different groups according to the final diagnosis. The clinical characteristics, laboratory examinations, endoscopic findings, and histopathological results were compared.

Results

IBD (n = 37) was just responsible for 43% of patients with colonic ulceration. Other diagnosis included autoimmune diseases (n = 9), infectious enteritis (n = 13), gastrointestinal allergy (n = 8), and other diseases (n = 19). Comparing IBD and non-IBD groups, children with IBD had a higher frequency of symptoms like weight loss/failure to thrive (P < 0.001), perianal lesions (P = 0.001), and oral ulcers (P = 0.022), and higher expression levels of platelet (P = 0.006), neutrophil-to-lymphocyte ratio (NLR) (P = 0.001), erythrocyte sedimentation rate (P < 0.001), C-reactive protein (P < 0.001), Immunoglobulin G (P = 0.012), Interleukin-1β (P = 0.003), Interleukin-6 (P = 0.024) and TNF-α (P = 0.026), and a wider ulcer distribution in the lower gastrointestinal tract (LGIT) (P < 0.001). Expression levels of hemoglobin (P < 0.001) and albumin (P = 0.001) were lower in IBD patients. Multivariate analysis showed hemoglobin, NLR, Score of ulceration in LGIT, and pseudopolyps contributing to the diagnosis of pediatric IBD with colonic ulcers.

Conclusions

We displayed potential indicators to help diagnose pediatric IBD differentiating from other disorders with colonic ulcers more prudently.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12887-024-05174-3.

Significance

What is known

• Many diseases can mimic IBD and reducing the misdiagnosis of IBD children is essential for the treatment and prognosis. Inflammatory factors like erythrocyte sedimentation rate, and fecal calprotectin have clinical value in diagnosing IBD.

What is new

• In our study, we investigated that IBD was just responsible for 43% of patients with colonic ulceration.

• Some potential indicators (like hemoglobin, neutrophil-to-lymphocyte ratio and ulcer distribution character) may contribute to the diagnosis of pediatric IBD with colonic ulcers.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12887-024-05174-3.

Introduction

Colonic ulcerations are a kind of localized defect of mucosal surface tissue found in colon and rectum [1]. The etiology of colonic ulceration is multiple, including inflammatory bowel disease (IBD), infectious colitis, allergic colitis, et al. [2–5]. A differential diagnosis is required for further treatment and prognosis evaluation [6]. It is challenging, and misdiagnosis is not rare in clinical practice. Many diseases (including infection, drug-induced disease, vascular disorders, diverticular disease, diversion proctocolitis, and immune disorders) can mimic IBD in clinical and pathological features [7]. Inflammatory factors like erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and fecal calprotectin (FCP) have been found for potential clinical diagnosis value with IBD but still need higher specificity [8, 9]. In most cases, acute infectious colitis is easily identified by clinical symptoms with no requirement for colonoscopy. But the patients with negative findings in pathogen tests, who present gastrointestinal bleeding, respond poorly to treatment, or have previous chronic gastrointestinal disorders, may undergo colonoscopic examination and confront the complex differential diagnosis [10]. Gastrointestinal-involved vasculitis, such as Immunoglobulin A (IgA) vasculitis and systemic lupus erythematosus (SLE), could be very similar to IBD in the symptoms, endoscopic findings, and histological assessment. European Crohn’s and Colitis Organisation (ECCO) recommended some features to help in the differential diagnosis between primary vasculitis and IBD [7]. The imaging, serological, fecal, and molecular tests could assist differentiation.

The manifestation of IBD in the pediatric population is more complicated and less specific than in adults for a higher ratio of genetic factors involved. IBD is diagnosed based on a comprehensive analysis of clinical manifestations, endoscopic examination, histopathological examination, and needed to distinguish from diseases like intestinal infectious diseases, and intestinal lesions of autoimmune diseases. Therefore, it is prone to be overdiagnosed in patients with colonic ulceration [1, 2, 11]. Pediatricians should also pay attention to gastrointestinal allergic diseases, which are more prevalent in children and increasing [12]. The data focused on the pediatric population is limited. To avoid diagnostic pitfalls, we observed the colonoscopy findings of ulcer features and combined them with the clinical manifestations and biochemical indexes of patients for the further recognition of related diseases, especially between IBD and non-IBD diseases. We summarized the differential manifestations of patients in the study and hoped to provide an alert for the misdiagnosis of children with colonic ulceration.

Materials and methods

Patients and study design

From December 2014 to June 2022, 86 pediatric patients (aged from 1 month to 15 years, and male: female = 42: 44) were analyzed in the study from Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine. Inpatients with colonic ulceration discovered by colonoscopy for the first time were eligible for the inclusion criteria. The patients with ulcers isolated in rectum were excluded. All children had relatively complete medical history records.

Demographic parameters, clinical characteristics, and laboratory examinations were collected from the medical record retrieval system. Demographic parameters included age and gender. Clinical characteristics harbored weight loss/failure to thrive, abdominal pain, diarrhea, constipation, hematochezia, fever, perianal lesions, rash, oral ulcer, and joint pain. The colonoscopy findings were analyzed retrospectively in combination with histopathological results. We evaluated all the results of colonoscopies based on a complete series of pictures (at least 3 figures in every part: the terminal ileum, ileocecal part, ascending colon, transverse colon, descending colon, sigmoid colon, and rectum) and the written results. Most patients have complete video recordings. Furthermore, we divided the lower gastrointestinal tract (LGIT) into terminal ileum, ileocecal part, ascending colon, transverse colon, descending colon, sigmoid colon, and rectum; every part with ulceration meant one score, and then we accumulated total score. The etiology was determined by clinical manifestation, examination, and long-term follow-up (one year at least). The IBD was diagnosed after excluding infectious, allergic, immunologic and other diseases. When genetic mutation associated with the primary immune deficiencies was detected, we diagnosed those patients as monogenic IBD-like disorders but not ulcerative colitis (UC) or Crohn’s disease (CD) in this study [7]. For children with unclear pathogens who have experienced acute gastroenteritis, were given empirical medication; infectious enteritis is considered if there is no symptom recurrence during follow-up (within 3–6 months). Nonspecific colonic ulcer was diagnosed in the patients who ruled out other diseases and recovered well without any treatment during the follow-up.

Statistical analysis

All data were analyzed using SPSS 26.0 (SPSS, Inc., IBM, Chicago, Illinois, USA). The measurement data were expressed using a median with an interquartile range (IQR). The enumeration data were presented by frequency or proportions. Mann–Whitney non-parametric test, Pearson’s chi-squared test or Fisher’s exact test were used for the comparison between different groups. A P value < 0.05 was defined as the statistical significance. Then, Backward: likelihood ratio test was used to select the most useful predictive features from the remaining features. Receiver operating characteristic (ROC) curve analysis was used to further identify these features with clinical significance.

Results

The diagnosis of children finding colonic ulcer

Eighty-six pediatric patients finding colonic ulcers were enrolled in the research. Etiological analysis of patients was shown in Table 1. Of note, the most common diagnosis was IBD (37/86), including UC (8/86, 9.3%), CD (22/86, 25.6%), and monogenic IBD-like disorders (7/86, 8.1%) with the mutation of IL10RA, EP300 mutation, PIK3CD mutation, Wiskott-Aldrich Syndrome, and immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome, respectively. Otherwise, there were autoimmune diseases (6 cases of IgA vasculitis (7.0%), 2 cases of other vasculitis (2.3%), and 1 case of SLE (1.2%), respectively), and gastrointestinal allergy (2 cases of food protein-induced enterocolitis syndrome (FPIES) (2.3%), 3 cases of eosinophilic gastroenteritis (3.5%), and 3 cases of other gastrointestinal allergy diseases (3.5%), respectively). A high ratio of children was also diagnosed with infectious enteritis (13/86, 15.1%). The etiology harbored Mycobacterium tuberculosis, Salmonellae, Campylobacter jejuni, and opportunistic pathogens (Clostridium difficile, Candida albicans, and Klebsiella pneumoniae), associated with antibiotic use. Other causes of colonic ulcers included nonspecific colonic ulcers (14/86, 16.3%), neonatal necrotizing enterocolitis (NEC) (1, 1.2%), Hirschsprung’s disease (1, 1.2%), Langerhans cell histiocytosis (LCH) (1, 1.2%), digestive perianastomotic ulcerations (1, 1.2%), and anal fistula (1, 1.2%).

Table 1.

Etiological analysis of patients with colonic ulceration

Etiology	Number (n = 86) n (%)
IBD ( n  = 37)
*UC	8 (9.3%)
*CD	22 (25.6%)
Monogenic IBD-like disorders	7 (8.1%)
Non-IBD (n = 49)
Autoimmune diseases
Vasculitis
IgA vasculitis (Henoch-Schönlein purpura)	6 (7.0%)
Other vasculitis	2 (2.3%)
SLE	1 (1.2%)
Infectious enteritis	13 (15.1%)
Mycobacterium tuberculosis	2 (2.3%)
Salmonellae	3 (3.5%)
Campylobacter jejuni	1 (1.2%)
Clostridium difficile	1 (1.2%)
Candida albicans	1 (1.2%)
Klebsiella pneumoniae	1 (1.2%)
Unknown	4 (4.7%)
Gastrointestinal allergy
FPIES	2 (2.3%)
Eosinophilic gastroenteritis	3 (3.5%)
Other gastrointestinal allergy	3 (3.5%)
Others
Nonspecific colonic ulcer	14 (16.3%)
NEC	1 (1.2%)
Hirschsprung’s disease	1 (1.2%)
LCH	1 (1.2%)
Digestive perianastomotic ulcerations	1 (1.2%)
Anal Fistula	1 (1.2%)

IBD = Inflammatory bowel disease, *UC = Ulcerative colitis with no monogenic abnormality, *CD = Crohn’s disease with no monogenic abnormality, SLE = Systemic lupus erythematosus, FPIES = Food protein-induced enterocolitis syndrome, NEC = Necrotizing enterocolitis, LCH = Langerhans cell histiocytosis

Comparison of clinical characteristics in patients with colonic ulceration

Children were divided into IBD (n = 37) and non-IBD (n = 49) groups for further exploring the potential specific indicators of IBD. Statistical analysis displayed no significant difference between IBD and non-IBD groups in age and gender. The frequency of weight loss/failure to thrive (P < 0.001), perianal lesions (P = 0.001), and oral ulcer (P = 0.022) was significantly higher in IBD group than in non-IBD group (Table 2). The frequency of fever is higher in IBD group (18, 48.6%) than in non-IBD group (14, 28.6%), even though there was no significant difference between two groups (P = 0.057).

Table 2.

Comparison of characteristics of IBD and non-IBD patients

Events	IBD (n = 37), n (%)	Non-IBD (n = 49), n (%)	P value
Age, y,			0.115
Age < 2 y	8 (21.6%)	11 (22.4%)
2 y ≤ age < 6 y	3 (8.1%)	9 (18.4%)
6 y ≤ age < 10 y	8 (21.6%)	16 (32.7%)
Age ≥ 10 y	18 (48.6%)	13 (26.5%)
Gender (Male/Female)	16/21 (43.2%/56.8%)	26/23 (53.1%/46.9%)	0.367
Abnormal findings
Weight loss/failure to thrive	16 (43.2%)	6 (12.2%)	< 0.001***
Abdominal pain	17 (45.9%)	33 (67.3%)	0.046*
Diarrhea	19 (51.4%)	20 (40.8%)	0.331
Constipation	4 (10.8%)	6 (12.2%)	1.000
Hematochezia	15 (40.5%)	25 (51.0%)	0.335
Fever	18 (48.6%)	14 (28.6%)	0.057
†Perianal lesions	13 (35.1%)	3 (6.1%)	0.001**
Rash	7 (18.9%)	8(16.3%)	0.754
Oral ulcer	7 (18.9%)	1 (2.0%)	0.022*
Joint pain	4 (10.8%)	2 (4.1%)	0.432

IBD = Inflammatory bowel disease

†Perianal lesions = skin tag, perianal abscess, anal fistula, or anal fissure

*P < 0.05, **P < 0.01, ***P < 0.001

Then, we further divided children into IBD (UC or CD with no monogenic abnormality), monogenic IBD-like disorders and non-IBD groups (Table S1 – S2). Statistical analysis showed that the age of children with monogenic IBD-like disorders is significantly younger than children in IBD or non-IBD groups (P < 0.001).

Comparison of biochemical indexes in patients with colonic ulceration

The biochemical indexes were then analyzed in different groups (Table 3, Table S3 – S4). The levels of hemoglobin (P < 0.001) and albumin (P = 0.001) were lower in IBD group than in non-IBD group, while the levels of platelet (P = 0.006), neutrophil-to-lymphocyte ratio (NLR) (P = 0.001), ESR (P < 0.001), CRP (P < 0.001), Immunoglobulin G (IgG) (P = 0.012), and Interleukin-1β (IL-1β) (P = 0.003), IL-6 (P = 0.024) and TNF-α (P = 0.026) were significantly higher in IBD group than in non-IBD group. In addition, the level of FCP was > 1500 ug/g in 22 cases (78.6%) of IBD children and in 14 cases (58.3%) of non-IBD children (P = 0.065) (Table 3). NLR, ESR, CRP and TNF-α showed higher level in CD group than in gastrointestinal allergy group (Table S4).

Table 3.

Comparison of biochemical indexes of IBD and non-IBD patients

Events	IBD (n = 37)	Non-IBD (n = 49)	P value
Hemoglobin, g/dL, median (IQR)	109.00, (87.50–121.00) n = 37	127.00, (111.50–132.50) n = 49	< 0.001***
Albumin, mg/dL median, (IQR)	36.60, (31.45–41.25) n = 37	42.60, (37.20–45.55) n = 49	0.001**
Platelet, 10^9/L median, (IQR)	397, (310–521) n = 37	302, (258–391) n = 49	0.006**
D-Dimer median, (IQR)	0.55, (0.32–0.86) n = 29	0.315, (0.23–0.827) n = 42	0.124
NLR median, (IQR)	2.49, (1.80–4.42) n = 37	1.41, (0.96–2.70) n = 49	0.001**
AST, U/L median, (IQR)	29.00, (19.85–40.50) n = 37	34.00, (25.35–42.15) n = 49	0.081
ALT, U/L median, (IQR)	16.00, (10.00–21.00) n = 37	19.00, (11.00–26.00) n = 49	0.230
ESR, mm/h, median, (IQR)	36.50, (20.50–64.25) n = 36	16.00, (2.50–26.00) n = 33	< 0.001***
CRP, mg/L			< 0.001***
< 8, n (%)	10 (27.0%)	33 (67.3%)
≥ 8, n (%)	27 (73.0%)	16 (32.7%)
Positive fecal transferrin, n (%)	16 (84.2%)	27 (90.0%)	0.877
Positive fecal occult blood, n (%)	21 (60.0%)	28 (62.2%)	0.840
FCP, ug/g			0.065
≤ 50, n (%)	1 (3.6%)	2 (8.3%)
50 < FCP ≤ 500, n (%)	1 (3.6%)	6 (25%)
500 < FCP ≤ 1000, n (%)	2 (7.1%)	2 (8.3%)
1000 < FCP ≤ 1500, n (%)	2 (7.1%)	0 (0.0%)
> 1500, n (%)	22 (78.6%)	14 (58.3%)
IgA, g/L			0.279
< 0.7, n (%)	5 (17.9%)	7 (26.9%)
0.7 ≤ IgA < 4, n (%)	20 (71.4%)	18 (69.2%)
≥ 4, n (%)	3 (10.7%)	1 (3.8%)
IgG, g/L			0.012*
< 7, n (%)	2 (7.1%)	7 (26.9%)
7 ≤ IgG < 16, n (%)	20 (71.4%)	18 (69.2%)
≥ 16, n (%)	6 (21.4%)	1 (3.8%)
IL-8, pg/mL			0.921
< 62, n (%)	10 (43.5%)	9 (45.0%)
≥ 62, n (%)	13 (56.5%)	11 (55.0%)
IL-1β, pg/mL			0.003**
< 5, n (%)	2 (8.7%)	10 (50.0%)
≥ 5, n (%)	21 (91.3%)	10 (50.0%
IL-6, pg/mL			0.024*
< 5.9, n (%)	4 (17.4%)	10 (50.0%)
≥ 5.9, n (%)	19 (82.6%)	10 (50.0%)
IL-10, pg/mL			0.988
< 9.1, n (%)	15 (65.2%)	13 (65.0%)
≥ 9.1, n (%)	8 (34.8%)	7 (35.0%)
TNF-α, pg/mL			0.026*
< 8.1, n (%)	0 (0.0%)	4 (20.0%)
≥ 8.1, n (%)	23 (100.0%)	16 (80.0%)
IL-2R, pg/mL			0.080
< 710, n (%)	3 (12.5%)	7 (35.0%)
≥ 710, n (%)	21 (87.5%)	13 (65.0%)
25-(OH)-VitD, nmol/L			0.142
≤ 50, n (%)	16 (80.0%)	6 (54.5%)
> 50, n (%)	4 (20.0%)	5 (45.5%)

IBD = Inflammatory bowel disease, NLR = neutrophil-to-lymphocyte ratio, AST = Alanine aminotransferase, ALT = Alanine aminotransferase, ESR = Erythrocyte sedimentation rate, CRP = C-reactive protein, FCP = Fecal calprotectin, IgA = Immunoglobulin A, IgG = Immunoglobulin G, IL-8 = Interleukin-8, IL-1β = Interleukin-1β, IL-6 = Interleukin-6, IL-10 = Interleukin-10, TNF-α = tumor necrosis factor-α, IL-2R = Interleukin-2 receptor, 25-(OH)-VitD = 25-hydroxyvitamin D. *P < 0.05, **P < 0.01, ***P < 0.001

The comparison of endoscopic characteristics in patients with colonic ulceration

Next, we collected the endoscopic figures of patients with different diagnoses (UC, CD, monogenic IBD-like disorders and non-IBD, respectively) in Fig. 1, and sorted out the endoscopic information of patients (Table 4, Table S5 – S6). The existence of upper gastrointestinal ulceration presented no significant difference in IBD and non-IBD groups. Colonoscopy revealed that ulcers could be found in any segment of the colon, and a wider ulcer distribution range in LGIT was found in IBD group than in non-IBD group (P < 0.001), especially between CD and infectious enteritis groups. Besides, Table 4 displayed that cobblestone appearance, deep ulcer (mainly in CD group as Table S6), and pseudopolyps were mainly showed in IBD group.

Fig. 1.

Endoscopic figures of patients with different diagnosis in colon. (A) UC: multiple continuous shallow ulcers with white coating on the surface. (B) CD: Cobblestone appearance with ulceration. (C) monogenic IBD-like disorder: Ulcerative lesions covered with pus coating, and accompanied by scattered inflammatory polypoid hyperplasia around the ulcer. (D) Allergy: clear multiple ulcers

Table 4.

Comparison of endoscopic characteristics in IBD and non-IBD patients

Events	IBD (n = 37)	Non-IBD (n = 49)	P value
Ulcer distribution
Upper gastrointestinal ulceration, n (%)	9 (24.3%)	4(10.3%)	0.104
*Score of ulceration in LGIT, median (IQR)	5 (2–6)	2 (1–2)	< 0.001***
Abnormal findings
Deep ulcer, n (%)	16 (43.2%)	5 (10.2%)	< 0.001***
Pseudopolyps, n (%)	10 (27.0%)	2 (4.1%)	0.004**
Bleeding changes, n (%)	7 (18.9%)	10 (20.4%)	1.000
Erythema, n (%)	1 (2.7%)	3 (6.1%)	0.631
Cobblestone appearance, n (%)	4 (10.2%)	0 (0.0%)	0.031*
Narrow, n (%)	3 (8.1%)	2 (4.1%)	0.648
Distortion of ileocecal valve, n (%)	3 (8.1%)	0 (0.0%)	0.076

IBD = Inflammatory bowel disease, * Score of ulceration in LGIT: We divided the lower gastrointestinal tract (LGIT) into terminal ileum, ileocecal part, ascending colon, transverse colon, descending colon, sigmoid colon and rectum; every part with ulceration meant one score and we accumulated total score. *P < 0.05, **P < 0.01, ***P < 0.001

The comparison of histopathological results in patients with colonic ulceration

The histopathological presentations were including inflammatory cell infiltration, cryptitis, crypt abscess, granulation tissues hyperplasia, and granuloma (Fig. 2). We found IBD patients presented more pathological changes, including crypt abscess (P = 0.013), granulation tissues hyperplasia (P < 0.001), and granuloma (P = 0.016), compared to non-IBD patients (Table 5). When comparing IBD, monogenic IBD-like disorders and non-IBD groups (Table S7), only the characteristic of granulation tissues hyperplasia had statistical differences between children with monogenic IBD-like disorders (3, 42.9%) and non-IBD groups (1, 2.1%). Table S8 further displayed that CD group had significant ratio of granulation tissues hyperplasia. Meanwhile, there was no significant difference about the number of Eosinophilic granulocytes between patients with IBD and non-IBD groups (Table S9 - S10).

Fig. 2.

Representative images of H&E staining for colon tissues of patients with different diagnosis. (A) UC: formation of crypt abscess with partial gland deformation (200×). (B) CD: acute and chronic inflammatory cell aggregation (200×). (C) monogenic IBD-like disorder: infiltration of plasma cells and neutrophils, occasional crypt abscess (200×). (D) The figure of the child with allergy: numerous eosinophils (200×)

Table 5.

Comparison of histopathological results in IBD and non-IBD patients

Events	IBD (n = 37)	Non-IBD (n = 48)	P value
Cryptitis, n (%)	26 (70.3%)	31 (64.6%)	0.580
Crypt abscess, n (%)	13 (35.1%)	6 (12.5%)	0.013*
Granulation tissues hyperplasia, n (%)	13 (35.1%)	1 (2.1%)	< 0.001***
Granuloma, n (%)	9 (24.3%)	2 (4.2%)	0.016*

IBD = Inflammatory bowel disease. *P < 0.05, ***P < 0.001

Multivariate logistic regression analysis of potential indicators of IBD with colonic ulceration

When we finished the analysis of the clinical characteristics, laboratory examinations, histopathological results, and colonoscopy results compared between the IBD and non-IBD groups, multivariate analysis on the data was made for a prediction rule for colonic ulcers in pediatric patients (Table 6). We found the hemoglobin (OR = 0.955, P = 0.002), NLR (OR = 1.368, P = 0.014), Score of ulceration in LGIT (OR = 1.866, P < 0.001), and pseudopolyps (OR = 12.605, P = 0.014) contributing to the diagnosis of pediatric IBD with colonic ulcers. Then ROC curve analysis was used in these variables (Fig. 3). Our findings showed the area under the curve (AUC) > 0.5 in NLR (0.712), Score of ulceration in LGIT (0.759), and pseudopolyps (0.615), and AUC < 0.5 in hemoglobin (0.266). Besides, we found combined diagnosis of these variables had larger AUC (0.883).

Table 6.

Multivariate logistic regression analysis of potential indicators of IBD

Events	Regression coefficient	OR	95%CI	P value
Hemoglobin	-0.046	0.955	0.928–0.983	0.002**
NLR	0.314	1.368	1.066–1.757	0.014*
*Score of ulceration in LGIT	0.624	1.866	1.338–2.601	< 0.001***
Pseudopolyps	2.534	12.605	1.685–94.294	0.014*

OR = Odds ratio, CI = Confidence interval, NLR = neutrophil-to-lymphocyte ratio

*Score of ulceration in LGIT: We divided the lower gastrointestinal tract (LGIT) into terminal ileum, ileocecal part, ascending colon, transverse colon, descending colon, sigmoid colon and rectum; every part with ulceration meant one score and we accumulated total score

*P < 0.05, **P < 0.01, ***P < 0.001

Fig. 3.

Receiver operating characteristic (ROC) curve analysis. The figure showed the ROC of hemoglobin (P < 0.001, the cutoff value = 118 with sensitivity of 73.0% and specificity of 71.4%), NLR (P = 0.001, the cutoff value = 1.69 with sensitivity of 78.4% and specificity of 59.2%), Score of ulceration in LGIT (P < 0.001, the cutoff value = 2.50 with sensitivity of 70.3% and specificity of 77.6%), and pseudopolyps (P = 0.070)

Discussion

With the development of endoscopic technology, colonic ulcers were found in children with different diagnosis. Due to the particular stage of growth and development in children and different disease spectrum from adults, it is essential to do early differential diagnosis and provide precise treatment. Our study displayed some potential indicators for the help of the diagnosis of pediatric IBD differentiating from other disorders with colonic ulcers, though we need larger sample size to confirm the findings.

In the total of 86 pediatric patients found with ulcers by colonoscopy, the most common diagnosis was IBD. Early-onset IBD patients (< 10 years), especially very early-onset IBD (VEO-IBD) patients (< 6 years), present a closer relationship with genetic mutations [13]. We defined VEO-IBD patients as monogenic IBD-like disorders when they have a monogenic mutation according to ECCO [7]. In our study, 3 cases were with the mutation of IL10RA, which was of high frequency in monogenic IBD-like disorders, and the conventional therapy of IBD has poor efficacy, even with haematopoietic stem cell transplantation [14, 15]. Four cases were also with genetic abnormality, including EP300 mutation, PIK3CD mutation, Wiskott-Aldrich Syndrome, and IPEX syndrome, respectively. These children need early gene detection to confirm the diagnosis for the severe disease progression [16]. Autoimmune diseases had 9 cases, mainly including IgA vasculitis (6/86). IgA vasculitis is one of the most frequent vasculitides in childhood and had similar gastrointestinal manifestations (abdominal pain, hematochezia) with IBD, which needs care for the prognosis [17]. A total of 8 cases of gastrointestinal allergy were found in our study, and the disease occurs frequently in children. Nambu et al. have found the most common diagnoses except for IBD in children who performed colonoscopies, were eosinophilic gastrointestinal disorders, which need food avoidance [18]. Our findings showed that the levels of NLR, ESR, CRP, and TNF-α may distinguish CD and gastrointestinal allergy. Furthermore, a high proportion of children had intestinal ulcers for infectious enteritis (13/86), which may alert us to strictly control the indications of colonoscopy examination. The level of ESR was also higher in CD group than infectious enteritis group.

We mainly focused on comparing IBD with other diseases and the differences may help to increase the differential diagnosis of disease profile. The results displayed that the age of IBD group was mainly ≥ 10 years (60.0%), while the age of children with monogenic IBD-like disorders was mainly < 2 years (85.7%), which was consistent with previous research [19]. Clinical symptoms of children with colonic ulcers, weight loss/failure to thrive, perianal lesions, and oral ulcers had a higher frequency in IBD children than in non-IBD children. These abnormal findings fulfilled the criteria of clinical features in IBD [20, 21]. Our findings showed less abdominal pain in IBD children. The cause indicates that the mild pain was easily neglected due to other serious complaints in this retrospective study. And abdominal pain is a subjective feeling while children may lack correct expression.

Laboratory examinations presented lower hemoglobin and albumin expression levels in the IBD group than in the non-IBD group. These were typically elevated in active disease while mild IBD could show normal results in routine blood tests [21]. Sabery et al. found anemia as a predictor for IBD was 83% sensitivity, and albumin is also a sensitive maker when combined with inflammatory indexes like CRP or ESR [8, 9]. The intestinal ulcer lesions may contribute to the condition that IBD children are more susceptible to anemia than other diseases. Our multivariate analysis also identified lower hemoglobin with sensitivity of 73.0% and specificity of 71.4%. Platelet aggregation at the colonic mucosae was common in colitis and IBD [22], and our findings of the high level of platelet and D-Dimer in IBD children (age ≥ 6 y) indicated the IBD children had an increased risk for thrombus development than children with other colitis (Table S3). As a marker of reflecting systemic inflammation, NLR has been found to be significantly increased at diagnosis and active disease of IBD children compared with healthy subjects [23–25]. Our results further found that NLR was also considerably higher in IBD group than in non-IBD group, even though all patients had colonic ulceration. When NLR reaching the value of 1.69 (sensitivity of 78.4% and specificity of 59.2%) in a child with colonic ulcerations, we may need to consider IBD.

The expression levels of ESR and CRP were still significantly higher in IBD group. These markers were relatively sensitive and specific to distinguish IBD (especially CD) from other diseases in children with ulcers in colon. Besides, FCP is a kind of protein from neutrophils, considered as a noninvasive and valuable marker of intestinal inflammation [26, 27]. We found that the expression levels of FCP between two groups did not show statistical differences (P = 0.065). FCP in both IBD (78.6%) and non-IBD patients (58.3%) was mainly > 1500 ug/g. The reason indicates that young children have a higher baseline value of FCP compared with normal adults [27]. In addition, the research object in our study were all children with colonic ulceration, whose intestinal inflammation was relatively severe. Herein, FCP may not be sensitive to distinguishing IBD among children with colonic ulcerations.

Otherwise, IgG was also at a higher level in IBD group. IBD is a disorder closely related to immune reaction. IgG was predicted to be associated with IBD severity and response to gut microbiota [28]. Pro-inflammatory factors including IL-1β, and TNF-α, were increasing the expression level in IBD group. The finding indicates that IL-1β may be a potential specific marker of IBD in pediatric patients. The upstream mechanisms and signal pathways of IL-1β may be interesting perspectives of pathogeny which identify IBD from colonic ulcerative diseases, like mediating NLRP3 activation which regulates the generation and maturation of IL-1β [29–31]. In addition, anti-TNF-α drugs have been widely used in treating children with IBD [32]. Anti- IL-1β may also be an effective therapeutic target in IBD patients in the future.

Colonoscopy revealed statistical differences in ulcer distribution between IBD and non-IBD groups. The discrepancy in ulcer distribution range may help to distinguish IBD and non-IBD with colonic ulceration, especially between CD and infectious enteritis. ROC curve analysis indicated that when the ulcer affects more than 2.5 parts (≥ 3 parts in fact) of the intestinal segments, there is a possibility of diagnosis as IBD with sensitivity of 70.3% and specificity of 77.6%. This still needs to be confirmed by increasing the sample size. Deep ulcer (mainly in CD), and pseudopolyps were also more found in IBD group. Otherwise, cobblestone appearance was only found in IBD group, which fulfilled the criteria of endoscopic characteristics in CD [20]. Histopathological results play essential roles in the different diagnoses of intestinal disorders [21]. The characteristics of mucosal tissue with hematoxylin & eosin staining included cryptitis, crypt abscess, granulation tissues hyperplasia, granuloma, lymphocyte proliferation, and eosinophil infiltration. The histological phenotype is hardly distinguished for the overlapping in morphological aspects between IBD and non-IBD patients [10]. In children with colonic ulcers, we found the significant differences in the change of crypt abscess, granulation tissues hyperplasia, and granuloma between IBD and non-IBD groups. Granuloma is a kind of local change of chronic inflammation, which could be observed typically in CD [21]. We did not see significant eosinophil infiltration in IBD (even monogenic IBD-like disorders) or Infectious enteritis groups. Characteristic pathological changes still need further exploration in other intestinal disorders.

We still have some limitations in the research. In our IBD group, there were mainly CD children (n = 22). The sample size of UC is relatively small and still needs to be expanded. A larger sample size is necessary to further verify our findings and reduce bias in data analysis.

Conclusion

In conclusion, the etiologies caused by colonic ulcers are complex and diverse, and primary diseases are IBD, infectious enteritis, autoimmune diseases, and gastrointestinal allergy. Hemoglobin, albumin, NLR, ESR, and CRP could be considered essential reference indicators for CD in children with colonic ulcers. Platelet, IgG, IL-1β, IL-6, and TNF-α may be potential indicators. The sensitivity and specificity still need a large sample size to verify these indexes. Joint monitoring of hemoglobin, NLR, ulcer distribution range, and pseudopolyps may have clinical significance in identifying IBD from other disease when children with colonic ulcers. The differences found in clinical characteristics, laboratory examinations, histopathological results and endoscopic findings of IBD and non-IBD groups in the study are hoped to help further understand intestinal diseases and improve the diagnosis in children with colonic ulcers.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ALT

Alanine aminotransferase

AST

Alanine aminotransferase

AUC

Area under the curve

CD

Crohn’s disease

CI

Confidence interval

CRP

C-reactive protein

ECCO

European Crohn’s and Colitis Organisation

ESR

Erythrocyte sedimentation rate

FCP

Fecal calprotectin

FPIES

Food protein-induced enterocolitis

LGIT

Lower gastrointestinal tract

IBD

Inflammatory bowel disease

Ig

Immunoglobulin

IL

Interleukin

IPEX syndrome

Immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome

IQR

Interquartile range

LCH

Langerhans cell histiocytosis

NEC

Neonatal necrotizing enterocolitis

NLR

Neutrophil-to-lymphocyte ratio

NLRP3

NOD-like Receptor Family Pyrin Domain Containing 3

OR

Odds ratio

ROC

Receiver operating characteristic

SLE

Systemic lupus erythematosus

TNF-α

Tumor necrosis factor-α

UC

Ulcerative colitis

VEO-IBD

Very early-onset IBD

25-(OH)-VitD

25-hydroxyvitamin D

Author contributions

W. Y. and Y. W. developed the study design. Y. Y. and Y. T. collected the data and wrote the manuscript. Y. C., H. F., and Q. W. analyzed the data. W. Y. and Y. W. also reviewed and modified the manuscript. All of the authors approved the manuscript to be published.

Funding

This study was supported by National Natural Science Foundation of China (81701486), and Foundation of Science and Technology Commission of Shanghai Municipality (16ZR1428400).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

In accordance with the Declaration of Helsinki, the study was approved by the Ethics Committee of Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (XHEC-D-2022-250). Written informed consent was obtained from the parents of all individual participants included in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.