Leucine-rich alpha-2 glycoprotein as a superior biomarker to C-reactive protein for detecting small bowel lesions in Crohn’s disease

Abstract

BACKGROUND

Achievement of endoscopic healing (EH) is significant in the clinical practice of inflammatory bowel disease as it is correlated with improved prognosis. Existing biomarkers, including C-reactive protein (CRP), have relatively low accuracy for predicting EH, especially in small intestinal lesions in Crohn’s disease (CD); thus, noninvasive and more accurate biomarkers are required. Leucine-rich alpha-2 glycoprotein (LRG), a 50-kD protein, is produced under inflammatory conditions and has been reported to be useful in assessing disease activity in inflammatory bowel disease. However, the usefulness of LRG in small intestinal lesions in CD remains inconclusive.

AIM

To determine the usefulness of LRG for EH in small bowel lesions in CD and compare it with CRP.

METHODS

This study included 133 consecutive patients with CD who underwent balloon-assisted enteroscopy between June 2021 and March 2024 at Shiga University of Medical Science Hospital (Otsu, Japan). We retrospectively analyzed endoscopic scores in each of the ileum and colon and four markers including LRG, CRP, albumin, and Harvey-Bradshaw index (HBI). Spearman’s rank correlation coefficient and receiver operating characteristic analysis were performed.

RESULTS

Either active ileal or colonic lesions exhibited significant differences in LRG, CRP, albumin, and HBI compared with EH. CRP, albumin, and HBI showed a worse correlation with endoscopic activity in the ileum than that in the colon; however, LRG did not show a worse correlation (colon, r = 0.5218; ileum, r = 0.5602). Receiver operating characteristic analysis revealed that LRG for EH in the ileum and colon had the same cutoff values of 12.4 μg/mL. Comparing the areas under the curve of LRG and CRP for predicting EH in the ileum revealed a significantly higher areas under the curve of LRG (95% confidence interval, 0.017-0.194; P = 0.024), whereas the two showed no significant difference in the colon.

CONCLUSION

LRG is a useful biomarker in assessing the endoscopic activity of CD and is more useful than CRP in the small intestine.

Key Words: Crohn’s disease; Leucine-rich alpha-2 glycoprotein; Endoscopic activity; Ileum; Small intestine; Balloon-assisted enteroscopy

Core Tip: Targeting endoscopic healing (EH) is crucial, particularly in small bowel Crohn’s disease (CD), as EH is associated with better long-term outcomes. We have evaluated the usefulness of leucine-rich alpha-2 glycoprotein (LRG) as a biomarker in assessing EH and endoscopic activity in 133 consecutive patients with CD who underwent balloon-assisted enteroscopy. Our results indicate the usefulness of LRG in assessing EH and endoscopic activity in CD. Furthermore, the superiority of LRG over C-reactive protein has been demonstrated, especially in predicting EH in the small intestine.

INTRODUCTION

Crohn’s disease (CD), a type of inflammatory bowel disease (IBD), is a chronic inflammatory disorder that can affect any part of the gastrointestinal tract, with increasing incidence and prevalence worldwide[1,2]. Currently, it is suggested that the pathogenesis of IBD may involve in an abnormal mucosal immune response as a result of the interplay between genetic susceptibility, environmental factors, and intestinal microbiota[2-5]. As our understanding of the pathogenesis of IBD has increased, diagnostic techniques and treatment have also advanced. In this context, the treat-to-target (T2T) strategy for IBD has been validated and is now widely accepted[6]. In the T2T strategy, symptomatic remission and serum C-reactive protein (CRP) and fecal calprotectin (FC) level normalization are short- to intermediate-term targets, whereas endoscopic healing (EH) constitutes the long-term target. Targeting EH is critical, particularly in small bowel CD, because, although small bowel lesions can occur in up to 70% of cases with CD[7-9], symptoms and biomarkers, including CRP and FC, in small bowel CD are less sensitive and less reliable than those in colonic CD, limiting the ability of these indicators to predict EH in small bowel lesions. Furthermore, treating the target for EH is associated with better long-term outcomes, including prevention of clinical relapse, hospitalization, and surgery[6,10,11]. Although endoscopy is a crucial modality in evaluating EH in CD, detecting small bowel lesions using endoscopy remains challenging, especially in areas beyond the reach of conventional colonoscopy. The gold standard for deep small intestine imaging is balloon-assisted enteroscopy (BAE)[12]; however, owing to its invasive, time-consuming, and labor-intensive characteristics, frequent implementation is challenging. Therefore, a sensitive and noninvasive biomarker that can accurately predict EH or the activity of the small intestine would be useful.

Leucine-rich alpha-2 glycoprotein (LRG) is a 50-kD protein that is produced by various cells, including neutrophils, macrophages, and intestinal epithelial cells[13]. As LRG, unlike CRP, is not directly dependent on interleukin-6 signaling, its serum level has recently been shown to be more reflective of the disease activity of IBD[14-17]. LRG is useful for small bowel lesions in CD evaluated using various modalities, such as colonoscopy[18], capsule endoscopy (CE)[19-22], magnetic resonance enterography (MRE)[23], intestinal ultrasonography[24], and BAE. As BAE is one of the most accurate modalities for small bowel lesion detection, LRG is regarded as useful in detecting small bowel lesions evaluated using BAE. However, studies on the utility of LRG for detecting small bowel lesions in CD evaluated using BAE alone remain limited[25-27], and the utility of LRG has not been fully validated. We here evaluated the usefulness of LRG as a biomarker in assessing EH and endoscopic activity in a relatively large number of patients with CD who underwent BAE and compared its usefulness to that of CRP, a widely used biomarker.

MATERIALS AND METHODS

Patients

This was an observational retrospective single-center study conducted at Shiga University of Medical Science Hospital. We evaluated 133 consecutive patients with CD who underwent BAE within 60 days before and after LRG measurement between June 2021 and March 2024. CD diagnosis was made according to the established diagnostic criteria[28]. The following were the exclusion criteria: The presence of (1) Colostomy/ileostomy; (2) Active perianal disease; (3) Obvious infection including abscess; (4) Malignancies; and (5) Changes in disease condition or treatment. Data with a shorter period between endoscopy and LRG measurement were adopted when one patient underwent more than one BAE.

Data collection

Data of the characteristics, symptoms, and laboratory tests (serum LRG, CRP, and albumin levels) of patients at the time of BAE were collected from the medical records. The Harvey-Bradshaw index (HBI) was employed for assessing clinical disease activity[29].

Endoscopic procedure

All cases underwent retrograde BAE using a single-balloon enteroscope (SIF-H190, Olympus Medical System, Tokyo, Japan). All BAEs were performed by three experts. The scope was inserted into the ileum as deep as possible[30]. To assess the endoscopic activity, we employed a modified simple endoscopic score for CD (mSES-CD)[31]. To reflect active inflammation, the mSES-CD excluded stenosis from the four items (size of ulcers, proportion of ulcerated surface, proportion of affected surface, and stenosis) in the original SES-CD. As in previous studies, most of the cases were observed using retrograde BAE, and not the entire jejunum was observed[25,31]. Therefore, in the present study, the small intestine was divided into two parts, the terminal ileum and proximal ileum, excluding the jejunum. The colon was divided into the right colon, transverse colon, left colon, and rectum. Each item was scored 0-3; subsequently, the sum of each part was determined as the mSES-CD. EH was defined as the absence of ulcerative lesions[31]. Therefore, mSES-CD < 3 points was considered EH since even one ulcer would result in a score of ≥ 3 points.

Ethical considerations

This study was conducted in accordance with the Declaration of Helsinki and approved by the Shiga University of Medical Science Research Ethics Committee (R2024-016). As this was a retrospective study using anonymized data, the need for informed consent was waived. The information disclosure document was presented on the hospital website.

Statistical analysis

Continuous variables were expressed as medians with interquartile ranges. To compare changes in each parameter between baseline and follow-up, the pairwise t test was used. The Kruskal-Wallis test was used to compare the four groups. Categorical values were presented as numbers, and differences were analyzed using the χ² test or Fisher’s exact test. To analyze the correlation between the mSES-CD and LRG, CRP, albumin, and HBI, Spearman’s rank correlation coefficient was employed. To determine the optimal cut-off value of the biomarkers in predicting EH, receiver operating characteristic (ROC) analysis was performed. Additionally, the area under the curve (AUC) was calculated. All statistical analyses were performed using GraphPad Prism (GraphPad Software, version 9.4.0; San Diego, CA, United States) except for the comparative analysis of ROC curves, which was performed using R package pROC[32]. P < 0.05 was considered statistically significant.

RESULTS

Patient characteristics

The patient inclusion flow chart is shown in Figure 1. We identified 139 patients in whom BAE was performed within 60 days before and after LRG measurement. Six patients were excluded, and a total of 133 patients were analyzed. The characteristics of these patients are presented in Table 1. The median age was 38 years, and 103 patients (77.4%) were males. Steroids were administered in only one case, whereas biologics were used in 96 patients (72.2%).

Figure 1 Flow chart of the study inclusion and exclusion process. CD: Crohn’s disease; BAE: Balloon-assisted enteroscopy; LRG: Leucine-rich alpha-2 glycoprotein.

Table 1 Overall characteristics of the study participants, n (%).

Variables	n = 133
Sex, male	103 (77.4)
Age, median years (IQR)	38 (28-48.5)
BMI, median years (IQR)	20.9 (19.25-24.25)
Period between endoscopy and LRG measurement, median days (IQR)	18 (3-40)
Location
L1: Ileal	43 (32.3)
L2: Colonic	10 (7.5)
L3: Ileocolonic	80 (60.2)
Behavior
B1: Inflammatory	51 (38.3)
B2: Stricturing	54 (40.6)
B3: Penetrating	28 (21.0)
Previous intestinal resection	42 (31.6)
Concomitant treatment
5-Aminosalicylic acid	70 (52.6)
Steroids	1 (0.8)
Immunomodulator	34 (25.5)
Biologics	96 (72.2)
Anti-TNF alpha	44 (33.1)
Anti-integrin	4 (3.0)
Anti-interleukin 12/23 p40	40 (30.1)
Anti-interleukin 23 p19	6 (4.5)
JAK inhibitor	1 (0.8)
LRG, median μg/mL (IQR)	12.5 (9.4-18.0)
CRP, median mg/dL (IQR)	0.1 (0.03-0.26)
Albumin, g/dL (IQR)	4.2 (3.9-4.5)
HBI (IQR)	0 (0-2)
mSES-CD in the ileum (IQR)	3 (0-4)
mSES-CD in the colon (IQR)	0 (0-3)

IQR: Interquartile range; BMI: Body mass index; TNF: Tumor necrosis factor; JAK: Janus kinase; LRG: Leucine-rich alpha-2 glycoprotein; CRP: C-reactive protein; HBI: Harvey-Bradshaw index; mSES-CD: Modified simple endoscopic score for Crohn’s disease.

LRG was elevated and correlated with endoscopic activity during colitis and ileitis in patients with CD

Overall, 61 patients achieved EH in both the ileum and colon (EH group), whereas 43 patients, 16 patients, and 13 patients had inflammation in the ileum only (ileal group), colon only (colonic group), and both the ileum and colon (ileocolonic group), respectively. As shown in Table 2, patients’ characteristics differ in each group, especially regarding the use of biologics and disease behavior in the Montreal classification[33]. In addition to LRG, we used the following three parameters: CRP, albumin, and HBI as these parameters are frequently used for assessing CD disease activity. The ileal, colonic, and ileocolonic groups had significantly higher serum LRG concentrations than the EH group (Figure 2A). Furthermore, the EH group exhibited significantly different CRP, albumin, and HBI compared with the ileal, colonic, and ileocolonic groups (Figure 2B-D).

Figure 2 Comparison of each parameter among the endoscopic healing (n = 61), ileal (n = 43), colonic (n = 16), and ileocolonic groups (n = 13). ^aP < 0.05; ^bP < 0.01; ^cP < 0.001. The endoscopic healing group exhibited significantly different leucine-rich alpha-2 glycoprotein, C-reactive protein, albumin, and Harvey-Bradshaw index compared with the ileal, colonic, and ileocolonic groups. A: Leucine-rich alpha-2 glycoprotein; B: C-reactive protein; C: Albumin; D: Harvey-Bradshaw index. LRG: Leucine-rich alpha-2 glycoprotein; EH: Endoscopic healing; CRP: C-reactive protein; HBI: Harvey-Bradshaw index.

Table 2 Characteristics of the patients in each group, n (%).

Variables	EH group (n = 61)	Ileal group (n = 43)	Colonic group (n = 16)	Ileocolonic group (n = 13)	P value
Sex, male	46 (75.4)	34 (79.1)	12 (75.0)	11 (84.6)	0.886
Age, median years (IQR)	35 (25.5-47)	38 (29-50.5)	42.5 (25.5-48.3)	36 (21.5-51)	0.870
BMI, median years (IQR)	20.9 (19.4-24.4)	21.2 (19.3-24.6)	20.8 (18.8-23.8)	20.2 (18.3-23.1)	0.760
Location					N/A
L1: Ileal	20 (35.0)	23 (56.1)	0 (0)	0	-
L2: Colonic	6 (9.8)	0 (0)	4 (25.0)	0	-
L3: Ileocolonic	35 (57.4)	20 (32.8)	12 (75.0)	13 (100)	-
Behavior					0.0002
B1: Inflammatory	35 (57.4)	4 (6.6)	7 (43.8)	5 (38.5)	-
B2: Stricturing	15 (24.6)	28 (45.9)	6 (37.5)	5 (38.5)	-
B3: Penetrating	11 (18.0)	11 (18.0)	3 (18.8)	3 (23.1)	-
Previous intestinal resection	9 (14.8)	20 (32.8)	7 (43.8)	6 (46.2)	0.002
Concomitant treatment
5-Aminosalicylic acid	35 (57.4)	24 (39.3)	6 (37.5)	5 (38.5)	0.356
Steroids	0 (0)	1 (16.4)	0 (0)	0 (0)	0.541
Immunomodulator	17 (27.9)	13 (21.3)	3 (18.8)	1 (7.7)	0.357
Biologics	51 (83.6)	26 (42.6)	11 (68.8)	6 (46.2)	0.012
Anti-TNF alpha	29 (47.5)	6 (9.8)	5 (31.3)	4 (30.8)	0.005
Anti-integrin	1 (16.4)	2 (3.3)	1 (6.3)	0 (0)	0.622
Anti-interleukin; 12/23 p40	18 (29.5)	16 (26.2)	5 (31.3)	1 (7.7))	0.245
Anti-interleukin; 23 p19	3 (4.9)	2 (3.3)	0 (0)	1 (7.7)	0.780
JAK inhibitor	0 (0)	1 (1.6)	0 (0)	0 (0)	0.655

IQR: Interquartile range; BMI: Body mass index; TNF: Tumor necrosis factor; JAK: Janus kinase; N/A: Not applicable.

Subsequently, we examined whether endoscopic activity was correlated with these four parameters. The ileocolonic group was excluded from this analysis to evaluate inflammation in the colon and ileum. As shown in Figure 3A-D, all parameters showed relatively good correlations with endoscopic activity in the colon. Notably, LRG was markedly correlated with endoscopic activity in the ileum (Figure 4A), suggesting its superiority in detecting small bowel lesions in patients with CD. Conversely, CRP, albumin, and HBI did not show strong correlations with endoscopic activity in the ileum (Figure 4B-D).

Figure 3 Correlation of each parameter with the endoscopic activity in the colon (n = 77). Endoscopic activity was assessed using modified simple endoscopic score for Crohn’s disease. All parameters showed relatively good correlations with endoscopic activity in the colon. A: Leucine-rich alpha-2 glycoprotein; B: C-reactive protein; C: Albumin; D: Harvey-Bradshaw index. mSES-CD: Modified simple endoscopic score for Crohn’s disease; LRG: Leucine-rich alpha-2 glycoprotein; CRP: C-reactive protein; HBI: Harvey-Bradshaw index.

Figure 4 Correlation of each parameter with the endoscopic activity in the ileum (n = 104). Endoscopic activity was assessed using mSES-CD. C-reactive protein, albumin, and Harvey-Bradshaw index did not show strong correlations with endoscopic activity in the ileum, whereas leucine-rich alpha-2 glycoprotein was markedly correlated with endoscopic activity. A: Leucine-rich alpha-2 glycoprotein; B: C-reactive protein; C: Albumin; D: Harvey-Bradshaw index. mSES-CD: Modified simple endoscopic score for Crohn’s disease; LRG: Leucine-rich alpha-2 glycoprotein; CRP: C-reactive protein; HBI: Harvey-Bradshaw index.

Accuracy of LRG in detecting active colonic and ileal lesions in patients with CD

Next, by performing ROC analysis, we examined the accuracy of the four parameters. The ROC curves of LRG, CRP, albumin, and HBI in predicting EH in the colon and ileum are depicted in Figure 5. The AUCs of LRG, CRP, albumin, and HBI in the colon were relatively high, with values of 0.863, 0.805, 0.831, and 0.756, respectively. Conversely, in the ileum, all of these values were lower than those in the colon; however, only LRG remained above 0.8 (LRG: 0.809; CRP: 0.705; albumin: 0.703; and HBI: 0.639). This finding indicates that LRG can be a good biomarker for small bowel lesions that are difficult to detect using other parameters. The cut-off values, sensitivity, and specificity for EH determination in the colon and ileum are presented in Table 3. We further investigated whether a combination of LRG and other parameters could improve small bowel lesion detection and noted that LRG alone had greater accuracy than LRG combined with other parameters (Supplementary Table 1). As CRP level normalization is a significant and common goal in the T2T strategy, we compared the AUC of LRG to that of CRP[6]. Notably, the AUC of LRG was significantly superior to that of CRP in the ileum [95% confidence interval (CI): 0.017-0.194; P = 0.024)] (Table 3). The AUCs of LRG and CRP in the colon showed no significant differences (95%CI: -0.050 to 0.166; P = 0.295 (Table 3).

Figure 5 Receiver operating characteristic analysis of each parameter in predicting the endoscopic activity in the colon and ileum. The area under curves of leucine-rich alpha-2 glycoprotein, C-reactive protein, albumin, and Harvey-Bradshaw index in the colon were relatively high. Conversely, in the ileum, all of these values were lower than those in the colon; however, only leucine-rich alpha-2 glycoprotein remained relatively high. A-D: Receiver operating characteristic analysis of each parameter in predicting the endoscopic activity in the colon; E-H: Receiver operating characteristic analysis of each parameter in predicting the endoscopic activity in the ileum. LRG: Leucine-rich alpha-2 glycoprotein; CRP: C-reactive protein; HBI: Harvey-Bradshaw index; AUC: Area under curves.

Table 3 Area under curves, cut-off value, sensitivity, specificity, and accuracy of leucine-rich alpha-2 glycoprotein and C-reactive protein in predicting endoscopic healing in the ileum and colon.

Ileum/colon	Variables	AUC (95%CI)	Cutoff value	Sensitivity (95%CI)	Specificity (95%CI)	Accuracy	P value (95%CI)
Ileum	LRG	0.809 (0.724-0.894)	12.4 μg/mL	0.767 (0.623-0.869)	0.787 (0.669-0.871)	0.779	0.024 (0.017-0.194)
	CRP	0.705 (0.602-0.808)	0.20 mg/dL	0.465 (0.325-0.611)	0.869 (0.762-0.932)	0.702
Colon	LRG	0.863 (0.747-0.978)	12.4 μg/mL	0.875 (0.640-0.978)	0.787 (0.669-0.871)	0.805	0.295 (-0.050 to 0.166)
	CRP	0.805 (0.689-0.921)	0.13 mg/dL	0.750 (0.505-0.898)	0.738 (0.616-0.832)	0.740

AUC: Area under the curve; CI: Confidence interval; LRG: Leucine-rich alpha-2 glycoprotein; CRP: C-reactive protein; EH: Endoscopic healing.

LRG is useful in assessing improvement of small bowel lesions

We finally examined whether each parameter could predict improvement in endoscopic ileal activity. 35 patients underwent follow-up BAE after more than 6 months. Patients with CD who had at least one point of improvement in the mSES-CD in the ileum had significantly lower LRG, CRP, and albumin levels (Figure 6). In contrast, these parameters did not show significant differences in patients whose SES-CD remained unchanged. The HBI consistently showed no significant differences, likely because several patients had an HBI score of 0. These results suggest that LRG, CRP, and albumin are useful in assessing improvement of small bowel lesions; however, the HBI, a parameter of clinical symptom, is not.

Figure 6 Changes in each parameter in patients who underwent follow-up balloon-assisted enteroscopy. ^aP < 0.05, ^bP < 0.01. Patients who improved the Modified simple endoscopic score for Crohn’s disease (mSES-CD) in the ileum had significantly lower leucine-rich alpha-2 glycoprotein, C-reactive protein, and albumin levels, whereas these parameters did not show significant differences in patients whose simple endoscopic score for Crohn’s disease remained unchanged. The Harvey-Bradshaw index consistently showed no significant differences. A: Patients with at least one point of improvement in the mSES-CD in the ileum (n = 16); B: Patients with unchanged mSES-CD in the ileum (n = 19). LRG: Leucine-rich alpha-2 glycoprotein; CRP: C-reactive protein; HBI: Harvey-Bradshaw index.

DISCUSSION

We here demonstrated the superiority of LRG over CRP as a biomarker for ileal inflammation detection as assessed using BAE in patients with CD. CRP is frequently used in the clinical practice of IBD for assessing clinical activity, and CRP level normalization is an intermediate-term goal in the Selecting Therapeutic Targets in IBD-II statement[6]. CRP elevation has been associated with clinical and endoscopic activities[34]. Moreover, patients with high CRP levels have been shown to be associated with higher hospitalization rates, even among those without symptoms[35]. However, CRP has a relatively low sensitivity for endoscopic activity detection in CD, especially in small bowel lesions[7-9,36]. Furthermore, our findings showed that the AUC of CRP in EH detection in the small intestine was lower than that in the colon. EH is associated with fewer hospitalizations, decreased future steroid treatment, and fewer complications including fistula and abscess[37,38]; thus, CRP level normalization is inadequate as a long-term target, and EH should be achieved. In this regard, a sensitive biomarker for small bowel lesion detection is clearly needed.

The usefulness of LRG as an inflammation biomarker was initially reported in rheumatoid arthritis and CD, and subsequently in ulcerative colitis[39,40]. The first study reported that LRG was markedly correlated with clinical activity in CD. However, the weak correlation between the clinical activity index represented by CD activity index and endoscopic activity has long been noted[41]. Regarding the correlation between LRG and endoscopic activity in CD, a previous study observed that LRG was not significantly correlated with SES-CD[42]; however, Shinzaki et al[14] evaluated patients with IBD receiving adalimumab using colonoscopy and reported that LRG was significantly correlated with SES-CD, and the subsequent study reported similar results[14,17]. This discrepancy is presumably due to the inability of routine colonoscopy to sufficiently observe small bowel lesions.

CE and MRE are less invasive modalities for detecting small bowel lesions in CD than BAE. LRG has been shown to correlate with endoscopic activity in the small intestine as assessed using CE[20-22]. Furthermore, compared with CRP, LRG showed a high correlation with transmural inflammation as assessed using MRE[23]. Therefore, studies demonstrating the positive correlation between LRG and disease activity in the small intestine are accumulating. Nevertheless, BAE can more readily detect small mucosal lesions, including aphthae, erosions, or small ulcers, than other modalities and remains to be the gold standard[12,43]. However, frequent implementation of BAE is labor-intensive, time-consuming, invasive, and challenging. Moreover, only a few facilities have the equipment and experts necessary for performing BAE. In this regard, the present study, which included a relatively large number of BAEs, is considered valuable. Our search extracted only three previous studies on the correlation between LRG and small bowel lesions detected using BAE[25-27]. These three previous studies along with the present study are summarized in Table 4[25-27]. The first study conducted by Kawamoto et al[25] reported that LRG was markedly correlated with endoscopic activity in the small intestine, with AUC, cutoff value, sensitivity, and specificity of 0.850 (95%CI: 0.766-0.933), 13.4 mg/mL, 0.79 (95%CI: 0.74-0.82), and 0.82 (95%CI: 0.68-0.92), respectively (Table 4). The reason for the slightly better results compared with our results may be that the study included a small number of patients with ileal and colonic ulcers. The same group recently reported that LRG combined with FC is appropriate for assessing endoscopic activity in the small intestine[26]. Furthermore, Kawamura et al[27] showed that LRG can be a reliable marker of the endoscopic activity of CD. Of the four studies, the study conducted by Kawamura et al[27] has the best sensitivity and specificity, possibly due to the relatively small number of patients with small bowel CD.

Table 4 Area under curves, cut-off value, sensitivity, and specificity of leucine-rich alpha-2 glycoprotein in predicting endoscopic healing in the small intestine determined using balloon-assisted enteroscopy according to the previous and present studies.

Ref.	Number of patients	AUC (95%CI)	Cut-off value	Sensitivity	Specificity
Kawamoto et al[25]	108	0.850 (0.766-0.933)	13.4 μg/mL	0.79	0.82
Kawamoto et al[26]	75	0.805 (0.703-0.907)	13 μg/mL	0.71	0.80
Kawamura et al[27]	36	0.885 (0.744-1.000)	8.9 μg/mL	0.926	0.778
Present study	104	0.809 (0.724-0.894)	12.4 μg/mL	0.767	0.787

AUC: Area under the curve; CI: Confidence interval.

Our study had several limitations. First, this was a retrospective single-center study. Although we analyzed consecutive patients, a selection bias may still exist. Second, in most patients, BAE was performed on a different day rather than the day of LRG measurement. Although this may affect the correlation between endoscopic findings and parameter levels, a previous study has also used a period of ≤ 8 weeks[25], which is comparable to the present study. Furthermore, in the present study, the median period between endoscopy and LRG measurement was 18 days, which does not appear to have strongly affected the results. Third, differences existed in the characteristics of patients in the EH, ileal, colonic, and ileocolonic groups compared in this study. The differences, including the use of biologics and disease behavior, may have influenced the results. Fourth, FC, another essential biomarker, was not analyzed. The usefulness of FC and LRG should be compared; however, LRG has the advantage of being simpler than FC because it is a blood test. Fifth, we did not perform antegrade BAE in most cases and did not evaluate the entire jejunum. Although there are fewer lesions in the upper small intestine than in the lower small intestine, the accuracy of each parameter may be underestimated. Sixth, the definition of EH can vary across studies[44]. In the present study, EH was defined as the absence of ulcer lesions. Therefore, mSES-CD < 3 points was considered EH; however, when 0 point was defined as EH, the cut-off values would be changed. However, the fact remains that LRG has an excellent correlation with endoscopic activity. Despite these limitations, LRG appears to be a useful biomarker for detecting lesions in CD, especially small bowel lesions, and further multicenter prospective studies are warranted.

CONCLUSION

Our results show that LRG is useful for EH and endoscopic activity assessment in CD and is superior to CRP, especially in predicting EH in the small intestine. An LRG of < 12.4 μg/mL may be considered an EH indicator. The use of LRG offers a potential alternative to endoscopy and may appropriately predict the timing of therapeutic intervention. Further validation studies are warranted.

ACKNOWLEDGEMENTS

The authors wish to acknowledge Dr. Takeshi Yamamura of the Department of Gastroenterology and Hepatology, Nagoya University Graduate School of Medicine for providing the data upon our request.