Fecal surrogate markers strongly correlate with endoscopic findings in pediatric onset inflammatory bowel disease: a retrospective study in Japan

Article information

Intest Res. 2025;.ir.2025.00103

Publication date (electronic) : 2025 December 12

doi : https://doi.org/10.5217/ir.2025.00103

Ryoko Yoshimura¹, Takahiro Kudo^,1, Masanori Toda¹, Kosuke Kashiwagi¹, Masumi Nagata¹, Kaori Aoki¹, Natsuki Ito¹, Kazuhide Tokita¹, Nobuyasu Arai¹, Reiko Kyodo¹, Masamichi Sato¹, Eri Miyata¹, Keisuke Jimbo¹, Yoshikazu Ohtsuka¹, Toshiaki Shimizu², Hiromichi Shoji²

¹Department of Pediatrics, Juntendo University Faculty of Medicine, Tokyo, Japan

²Department of Pediatrics and Adolescent Medicine, Juntendo University Graduate School of Medicine, Tokyo, Japan

Correspondence to Takahiro Kudo, Department of Pediatrics, Juntendo University Faculty of Medicine, 2-1-1 Hongo, Bunkyo-ku, Tokyo 113-8421, Japan. E-mail: t-kudo@juntendo.ac.jp

Received 2025 January 5; Revised 2025 March 25; Accepted 2025 April 2.

Abstract

INTRODUCTION

Inflammatory bowel disease (IBD) includes ulcerative colitis (UC), Crohn’s disease (CD), and IBD unclassified (IBDU). It causes chronic intestinal inflammation with periods of relapse and remission. As the cause remains unknown and no cure exists, treatment focuses on controlling inflammation and maintaining remission. The Selecting Therapeutic Targets in Inflammatory Bowel Disease study [1] found that reducing systemic inflammation improves outcomes in IBD management, emphasizing the role of endoscopic evaluation [2]. Endoscopy remains the gold standard for monitoring disease. However, its use in pediatric patients is limited due to its invasiveness, time demands, and limited facility availability. This has increased interest in less-invasive alternatives, with fecal biomarkers showing promise.

Fecal calprotectin (FC) is a calcium- and zinc-binding protein released from neutrophils in response to intestinal inflammation [3]. Similarly, fecal lactoferrin (FL), an iron-binding glycoprotein stored in neutrophil granules, is also released during inflammation and excreted in the stool [4]. Both biomarkers reflect neutrophil activity in the stool, providing a more direct measure of local intestinal inflammation than blood-based biomarkers [5]. Furthermore, FC and FL levels remain stable at room temperature for several days, enabling efficient measurement with small stool samples. While many studies have highlighted the utility of FC in pediatric IBD [3,6-10], the accuracy of these biomarkers has not been thoroughly assessed. This study focused on childhood-onset IBD. We simultaneously measured fecal immunochemical test (FIT) results, FC levels, and FL levels in stool samples using the colloidal gold agglutination method which gives rapid results and compared the correlation between these markers and their effectiveness in evaluating activity.

METHODS

RESULTS

1. Participant Characteristics

Tables 1 and 2 summarize the patient characteristics. During the study period, 88 patients underwent testing, and 1,105 stool samples were collected. Among the 88 patients, 65 had UC, 20 had CD, and 3 had IBDU. Patients were aged 3 to 27 years, including 56 males and 32 females. The median age at examination was 18 years, with a median age at onset of 11 years and a median disease duration of 7 years. In UC, the most common type was E4 (48 patients, 74%), followed by E1 (4 patients, 6%), E2 (11 patients, 17%), and E3 (2 patients, 3%). In CD, the most frequent type was the small intestine-colon type (16 patients, 80%), followed by the small intestine type (1 patient, 5%), and the colon type (3 patients, 15%). Stool samples (n=1,105) were analyzed, with 803 UC cases, 251 CD, and 51 IBDU cases. The median (interquartile range) FIT results, FC levels, and FL levels were 20 ng/mL (10–148), 266 μg/g (83–865), and 74 ng/mL (20–323), respectively (Table 1). Endoscopic examinations were performed in 68 patients (45 with UC, 18 with CD, and 5 with IBDU), and none developed severe conditions according to the endoscopic activity index (Table 2).

Table 1.

Patient Characteristics

Table 2.

Patient Characteristics

2. Correlation Analysis of Fecal Biomarkers

Fig. 1 shows the distribution of fecal biomarker values. A significant correlation was observed between FC and FL across all groups (Spearman’s rank correlation coefficient: all =0.92, UC=0.93, CD =0.90, and IBD-U =0.62). In contrast, the correlations between FIT and FC or FL were weaker, though still significantly positive.

Fig. 1.

Distribution map of stool biomarkers. (A) All patients, (B) patients with ulcerative colitis, (C) patients with Crohn’s disease, and (D) patients with inflammatory bowel disease unclassified. FL, fecal lactoferrin; FC, fecal calprotectin; FIT, fecal immunochemical test.

3. Correlation Analysis of Fecal Biomarkers and Endoscopic Activity Index in UC

Fig. 2 presents distribution maps for the MES and UCEIS scores, as well as stool biomarker measurements. All stool biomarkers exhibited significant positive correlations with the endoscopic activity index. The strongest correlation was observed for FIT, followed by FC and FL.

Fig. 2.

Comparisons of fecal biomarker levels in patients with ulcerative colitis by endoscopic activity score. (A) Mayo Endoscopic Subscore (MES). (B) Ulcerative Colitis Endoscopic Index of Severity (UCEIS). FC, fecal calprotectin; FL, fecal lactoferrin; FIT, fecal immunochemical test.

4. FC, FL, and FIT levels Were Compared According to UC Severity Based on Endoscopic Findings

Patients were classified according to the MES into 2 groups: those with endoscopic activity (MES ≥2) and those in endoscopic remission (MES ≤1). Stool biomarker measurements were compared between these groups (Fig. 3). In the MES ≤ 1 group, median fecal biomarker values were 303 μg/g for FC, 93 ng/mL for FL, and 20 ng/mL for FIT. In the MES ≥ 2 group, median values were 1,124 μg/g for FC, 375 ng/mL for FL, and 536 ng/mL for FIT. The MES ≥2 group showed significantly higher levels of all stool biomarkers compared to the MES ≤ 1 group (P<0.05). Based on UCEIS scores, the patients were further categorized into the endoscopic remission group (UCEIS ≤1), mild disease group (UCEIS of 2–4), and moderate disease group (UCEIS of 5–6), with comparisons made among these groups (Fig. 4). In the UCEIS ≤ 1 group, median FC was 229 μg/g, FL level was 28 ng/mL, and FIT value was 15 ng/mL. In the UCEIS 2–4 group, median FC was 927 μg/g, FL was 212 ng/mL, and FIT was 54 ng/mL. The UCEIS 5–6 group showed median values of 2,988 μg/g for FC, 988 ng/mL for FL, and 8,793 ng/mL for FIT. Significant differences were observed among the 3 groups, with increasing stool biomarker levels correlating with higher UCEIS scores (P<0.05).

Fig. 3.

Median fecal biomarker levels across different grades of endoscopic activity according to the Mayo Endoscopic Subscore (MES). FC, fecal calprotectin; FL, fecal lactoferrin; FIT, fecal immunochemical test.

Fig. 4.

Median fecal biomarker endoscopic subscore across different grades of endoscopic activity according to the Ulcerative Colitis Endoscopic Index of Severity (UCEIS). FC, fecal calprotectin; FL, fecal lactoferrin; FIT, fecal immunochemical test.

5. Prediction of Endoscopic Remission Based on FC and FL Levels in UC

A MES of ≤1 was classified as endoscopic remission (negative), and a MES of ≥2 as endoscopically active (positive), both analyzed using ROC (Fig. 5A). For FC, the AUC was 0.76 with a cutoff of 983 μg/g, resulting in 69% sensitivity and 83% specificity. The AUC for FL was 0.73, with a cutoff value of 368 ng/mL, yielding 60% sensitivity and 83% specificity. Similarly, ROC analysis defined endoscopic remission (negative) as UCEIS ≤1 and endoscopically active (positive) as UCEIS ≥2 (Fig. 5B). For FC, the AUC was 0.75, with a cutoff of 927 μg/g, yielding 57% sensitivity and 100% specificity. The AUC for FL was 0.74, with a cutoff of 204 ng/mL, yielding 60% sensitivity and 87% specificity.

Fig. 5.

AUC for predicting endoscopic disease activity by FC and FL levels. (A) Mayo Endoscopic Subscore of <2. (B) Ulcerative Colitis Endoscopic Index of Severity of <2. FC, fecal calprotectin; FL, fecal lactoferrin; AUC, area under the curve; CI, confidence interval.

6. Correlation Analysis for Fecal Biomarker and Endoscopic Activity Index in CD

The distribution of SES-CD and stool biomarkers is shown in Fig. 6. The Spearman’s correlation coefficients were 0.76 for FC, 0.83 for FL, and 0.68 for FIT, with all markers showing positive correlations.

Fig. 6.

Comparisons of fecal biomarker levels by endoscopic activity score according to the Simple Endoscopic Score for Crohn’s Disease (SES-CD) in patients with Crohn’s disease. FIT, fecal immunochemical test; FC, fecal calprotectin; FL, fecal lactoferrin.

7. Comparison of Stool Biomarkers by CD Severity Based on Endoscopic Findings

Fig. 7 compares stool biomarker levels across the SES-CD groups: the endoscopically inactive group (SES-CD ≤2), the mild group (SES-CD ≤ 6), the moderate group (SES-CD ≤ 15), and the severe group (SES-CD ≥16). Significant differences were observed in all stool biomarkers (P<0.05).

Fig. 7.

Median fecal biomarker levels across different grades of endoscopic activity according to the Simple Endoscopic Score for Crohn’s Disease (SES-CD). FC, fecal calprotectin; FL, fecal lactoferrin; FIT, fecal immunochemical test.

8. Prediction of Endoscopic Remission by FC and FL Levels in CD

Endoscopic remission (negative) was defined as SES-CD ≤ 2, and endoscopically active (positive) as SES-CD ≥3. ROC analysis (Fig. 8) showed an AUC of 0.96 for FC with a cutoff value of 630 μg/g, yielding 86% sensitivity and 100% specificity. For FL, the AUC was 0.95, with a cutoff value of 79 ng/mL, also yielding 86% sensitivity and 100% specificity.

Fig. 8.

AUC for predicting endoscopic disease activity (Simple Endoscopic Score for Crohn’s Disease <3) based on FC and FL levels. FC, fecal calprotectin; FL, fecal lactoferrin; AUC, area under the curve; CI, confidence interval.

DISCUSSION

Data from 1,105 stool samples were analyzed, including only patients with childhood-onset IBD. This is a large sample size for a pediatric study. In our study, stool biomarkers were measured using the gold colloid method. In Japan, the ELISA (enzyme-linked immunosorbent assay) and FEIA (fluorescent enzyme immunoassay) methods have been widely used to measure fecal biomarkers, but it took about 1 week to obtain results. The gold colloid method used in this study could measure FIT, FC, and FL in about 10 minutes from a small amount of stool specimen, so the test results could be used in clinical practice to make decisions about medical treatment without a time lag.

The present study found a modest correlation between FIT and FC, consistent with previous reports in children, but a significant correlation between FC and FL [16]. FC, FL and FIT positively correlated with endoscopic activity indices in UC and CD, as shown in earlier children [6] and all-age studies [17]. In UC, the correlation was strongest for FIT, followed by FC and FL. This result was similar to previous report on adults [18]. In CD, the strongest correlation was with FL, followed by FC and FIT, differing from UC.

Previous studies in adults have shown no correlation between FIT results and SES-CD in adult patients with small intestine- type CD [19]. However, this study found a significant correlation, as the endoscopic activity evaluation included both small and large intestine types. Each fecal biomarker has distinct characteristics, and factors like stool collection methods and bowel movement frequency can influence results. FIT values are also elevated in patients with hemorrhoidal bleeding and menstruation, even without inflammation. FC levels are influenced by intestinal inflammation unrelated to IBD, nonsteroidal anti-inflammatory drugs and proton pump inhibitor use, dietary factors, obesity, and bowel movement frequency. Notably, FC levels are higher in infancy than in adulthood [20-23]. FL levels may increase in patients with rectal cancer, polyps, and breastfed children [24,25]. Additionally, some studies in children [26] and adults [27-29] have shown that FC and FL levels often rise several months before a relapse. In this study, patients in remission, based on the endoscopic activity index, who had high FC and FL levels did not experience recurrence or require therapy changes during follow-up. These findings emphasize the need to understand each test’s unique properties and use them as complementary diagnostic tools alongside clinical evaluations.

The optimal cutoff values for FC in assessing endoscopic remission of UC were 983 μg/g for MES and 927 μg/g for UCEIS, with AUC values of 0.781 and 0.728, respectively. For FL, the optimal cutoff values were 217 ng/mL and 214 ng/mL for MES and UCEIS, with AUC values of 0.737 and 0.726, respectively. Specificity and sensitivity were 87% versus 77% for MES and 58% versus 67% for UCEIS, indicating strong diagnostic performance. These findings highlight FC and FL as reliable markers for assessing endoscopic activity in UC.

The optimal cutoff value for FL in evaluating endoscopic activity in IBD remains inconsistent, with significant variability across studies in children [26] and adults [30-34]. Further research is needed to determine a definitive FL cutoff value for assessing endoscopic remission in pediatric IBD patients. In conclusion, FC and FL were equally useful in predicting endoscopic remission in UC and CD. No difference in usefulness of FC and FL was found among the severity and classification of each disease.

This study has several limitations. First, the retrospective design may have introduced errors in test values due to a time lag of up to 3 months between stool sample collection and endoscopic examination. Second, only 70 patients’ endoscopic findings were analyzed, highlighting the need for a larger sample size and additional studies for more robust conclusions. Third, although the present study showed a good correlation between SES-CD and FIT, stratified analysis was not possible due to the small number of patients with small and large intestinal types of CD, and no conclusions could be drawn regarding the utility of FIT in small intestinal CD. Lastly, patient selection bias may have occurred due to the study’s single-center design. Future research should involve a large, multicenter, prospective study to confirm the correlation between endoscopic findings and stool biomarkers.

In conclusion, a strong correlation was observed between FL and FC measured using the gold colloid method. In patients with UC and CD, significant correlations between the endoscopic activity index and both FC and FL levels were noted. This method serves as an effective and rapid surrogate marker for disease activity, potentially reducing the need for frequent endoscopic examinations in patients with IBD.

Article information Continued