Serum Leucine-Rich Alpha-2 Glycoprotein in Quiescent Crohn’s Disease as a Potential Surrogate Marker for Small-Bowel Ulceration detected by Capsule Endoscopy

Background: Small bowel (SB) lesions in quiescent Crohn’s disease (CD) are sometimes not identified by clinical activity or existing markers. We investigated the usefulness of a novel biomarker, leucine-rich α2-glycoprotein (LRG), for screening for the presence of SB ulcerative lesions detected by small-bowel capsule endoscopy (SBCE). Methods: We examined patients with a Crohn’s Disease Activity Index (CDAI) value < 150 and a C-reactive protein (CRP) value < 0.5 mg/dL with SB or SB colonic CD. The presence of small-bowel ulcerative lesions (≥0.5 cm) was grouped by SBCE results, and we then compared the groups’ LRG value to establish a cutoff value for screening for the presence of lesions. Results: In 40 patients with CD, the LRG values differed significantly between the patients with and without SB ulcerative lesions (Ul + 14.1 (2.1–16.5) μg/mL vs. Ul − 12.3 (9.3–13.5) μg/mL; p = 0.0105). The respective cutoff LRG values for the presence of SB ulcerative lesions was 14 μg/mL (areas under the ROC curve 0.77), with sensitivity 63.6%, specificity 82.8%, positive predictive values 58.3%, negative predictive values 85.7%, and accuracy 78%. Conclusion: These results indicate that LRG may be useful in predicting the presence of SB inflammation associated in patients with CD with CRP < 0.5 mg/dL and CDAI < 150, and in selecting patients for SBCE.


Introduction
Crohn's disease (CD) is a chronic, destructive, progressive inflammatory disease that causes inflammation and intestinal damage mainly in the small and large bowel [1,2]. The small bowel in particular is involved in approximately 70% of cases of individuals with CD. It has been pointed out that small bowel lesions in CD are difficult to assess accurately with the Crohn's Disease Activity Index (CDAI), which is a subjective assessment of disease activity in CD; it has also been posited that a CD patient's C-reactive protein (CRP) value does not reflect the disease activity of small intestinal lesions [3].
In patients with CD, the Crohn's Disease Endoscopic Index of Severity (CDEIS) and the simple endoscopic score for Crohn's Disease (SES-CD) are generally determined by ileocolonoscopy before the use of endoscopy [4]. Although the CDEIS and SES-CD scores are correlated, the standard use of ileocolonoscopy does not adequately assess the deep small bowel on the mouth-side from the terminal ileum, and it may underestimate the presence of small bowel lesions [5,6].
Leucine-rich α2-glycoprotein (LRG) has attracted attention as a new biomarker in ulcerative colitis and CD [7][8][9][10]. LRG is an interleukin (IL)-6-independent 50-kD protein produced in the localized inflammation of the intestinal tract [7]. In CD, the identification of an LRG value ≥ 16 µg/mL (according to the manufacturer's instructions; Sekisui Medical 2 of 10 Co., Tokyo, Japan) was reported to be useful for discriminating patients in remission (CDAI < 150 and SES-CD < 4) from patients with active CD (CDAI ≥ 150 and SES-CD ≥ 4). LRG was also reported to be useful for discriminating disease activity even in CRP-negative CD [10]. These findings did not involve evaluation for deep small-bowel lesions on the mouth-side from the terminal ileum. However, extensive small bowel involvement has been shown to be a poor prognostic factor in CD [11].
Balloon-assisted enteroscopy (BAE) and small bowel capsule endoscopy (SBCE) have been used to visualize inflammation in the deep small bowel [12]. Studies of BAE conducted to identify small bowel lesions of CD indicated that small bowel ulcerative lesions ≥ 0.5 cm are an independent risk factor for relapse and surgery [3]. A study of SBCE reported that in patient CD with a CDAI < 150 or a CDAI < 220 that did not require new therapeutic intervention in 3 months, an SBCE score (i.e., Lewis score (LS)) ≥ 350 was predictive of relapse at 24 months [13]. The monitoring of small bowel lesions in patients with CD in clinical remission is of high clinical relevance. However, the use of BAE or SBCE as a monitoring tool in all patients is not feasible due to the invasiveness and cost of these procedures. If the presence of small bowel lesions can be predicted by LRG, a noninvasive biomarker, the use of LRG will contribute to the selection of patients who should undergo endoscopy. We conducted the present study to determine the association between LRG values and SBCE-diagnosed small-bowel active inflammatory lesions in patients with CD in clinical remission with a CDAI value < 150 and a CRP level < 0.5 mg/dL.

Patients Selection
We retrospectively evaluated the cases of 59 patients with small bowel or small bowel colorectal-type CD attending who had undergone: (1) a patency evaluation with the use of a patency capsule, (2) small bowel capsule endoscopy (SBCE), and (3) LRG measurement during the period July 2020 to December 2021 at Tokyo Women's Medical University. We then excluded the patients who had a CDAI ≥ 150, a CRP value ≥ 0.5 mg/dL, with an interval between their SBCE and LRG measurements of >30 ± 7 days, colorectal CD, active colorectal ulcerative lesions ≥ 0.5 cm, active perianal lesions, stoma, age < 18 years, failure of patency capsule assessment, or failure of total small bowel observation. A final total of 40 patients was included in the analysis ( Figure 1). of an LRG value ≥ 16 μg/mL (according to the manufacturer's instructions; Sekisui Medical Co., Tokyo, Japan) was reported to be useful for discriminating patients in remission (CDAI < 150 and SES-CD < 4) from patients with active CD (CDAI ≥ 150 and SES-CD ≥ 4). LRG was also reported to be useful for discriminating disease activity even in CRP-negative CD [10]. These findings did not involve evaluation for deep small-bowel lesions on the mouth-side from the terminal ileum. However, extensive small bowel involvement has been shown to be a poor prognostic factor in CD [11].
Balloon-assisted enteroscopy (BAE) and small bowel capsule endoscopy (SBCE) have been used to visualize inflammation in the deep small bowel [12]. Studies of BAE conducted to identify small bowel lesions of CD indicated that small bowel ulcerative lesions ≥ 0.5 cm are an independent risk factor for relapse and surgery [3]. A study of SBCE reported that in patient CD with a CDAI < 150 or a CDAI < 220 that did not require new therapeutic intervention in 3 months, an SBCE score (i.e., Lewis score (LS)) ≥ 350 was predictive of relapse at 24 months [13]. The monitoring of small bowel lesions in patients with CD in clinical remission is of high clinical relevance. However, the use of BAE or SBCE as a monitoring tool in all patients is not feasible due to the invasiveness and cost of these procedures. If the presence of small bowel lesions can be predicted by LRG, a noninvasive biomarker, the use of LRG will contribute to the selection of patients who should undergo endoscopy. We conducted the present study to determine the association between LRG values and SBCE-diagnosed small-bowel active inflammatory lesions in patients with CD in clinical remission with a CDAI value < 150 and a CRP level < 0.5 mg/dL.

Patients Selection
We retrospectively evaluated the cases of 59 patients with small bowel or small bowel colorectal-type CD attending who had undergone: (1) a patency evaluation with the use of a patency capsule, (2) small bowel capsule endoscopy (SBCE), and (3) LRG measurement during the period July 2020 to December 2021 at Tokyo Women's Medical University. We then excluded the patients who had a CDAI ≥ 150, a CRP value ≥ 0.5 mg/dL, with an interval between their SBCE and LRG measurements of >30 ± 7 days, colorectal CD, active colorectal ulcerative lesions ≥ 0.5 cm, active perianal lesions, stoma, age < 18 years, failure of patency capsule assessment, or failure of total small bowel observation. A final total of 40 patients was included in the analysis ( Figure 1).

Evaluation Method
SBCE, which can image the entire small bowel, was used to identify small-bowel lesions. The serological markers CRP, LRG, and CDAI were measured within 30 ± 7 days of the SBCE. Inflammatory mucosal defects with white mucosa > 0.5 cm were defined as smallbowel ulcerative lesions [14], and their presence was confirmed by SBCE ( Supplementary  Figures S1 and S2). The diameter of a typical small intestine is 2.5 cm, and we calculated the circumference of a quarter of a small intestine to be 2 cm. Based on that, we estimated the size of the ulcer. This ulcerative lesion was considered to be classified as a deep ulcer, a finding primarily associated with Crohn's disease, according to the International Consensus Statement [15,16]. Each patient's capsule endoscopic score, Lewis score (LS), Capsule Endoscopy Crohn's Disease Activity Index (CECDAI) [4], and Crohn's Disease Activity in Capsule Endoscopy (CDACE) were also calculated [17].
The primary endpoint of the study was to compare LRG values according to the presence or absence of small intestinal ulcerative lesions. We also compared whether LRG values differed in LS ≥ 350. As a result, we determined the appropriate LRG screening value by setting the cutoff value using the Youden index from the receiver operating characteristic (ROC) curve of the LRG values measured in the presence of small intestinal ulcerative lesions and an LS ≥ 350. We then calculated the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of the cutoff value. As a secondary endpoint, we calculated the correlations between the CRP values, LRG values, and each SBCE score, and we compared the background factors and SBCE scores of the patients with quiescent CD grouped by the previously reported LRG cutoff value (16 µg/mL) and our newly calculated cutoff value. Finally, LRG values and SBCE scores in the use of biological agents were evaluated.

SBCE Procedure
Intestinal patency was confirmed in all patients before the SBCE procedure by the use of patency capsules (PC; Medtronic, Minneapolis, MN, USA). SBCE was performed by practitioners with experience conducting > 1200 SBCE procedures. The capsule endoscopy device used for all patients was the PillCam™ SB3 (Medtronic, Minneapolis, MN, USA). The patient was instructed to fast from 21:00 on the day before the SBCE examination. At 8:00 the following morning, oral mosapride (15 mg) was administered as a pretreatment drug. An intestinal cleansing procedure was considered unnecessary for the SBCE. The timepoint at which the patient swallowed the SBCE with water marked the start of the procedure. Drinking water was provided 2 h after the SBCE ingestion, and a meal was provided 4 h after the patient swallowed the SBCE endoscope. The excretion of the capsule was confirmed visually after the completion of the examination.

Ethical Considerations
The study protocol was approved by the Institutional Ethics Review Committee of Tokyo Women's Medical University, and each patient provided written informed consent (IRB no. 2021-0114).

Statistical Analysis
All the data are expressed as the median (interquartile range (IQR)). Wilcoxon's test was used in the univariate analysis of background factors. Spearman's rank correlation coefficient was used to analyze the correlations of the LRG score with the CDACE, LS, CECDAI, and CDAI scores. Probability (p)-values < 0.05 were considered significant. With regard to the presence of small intestinal ulcers (≥0.5 cm) and LS ≥ 350, we calculated the cutoff value of each score using the Youden index from the ROC curve and computed the sensitivity, specificity, PPV, NPV, and accuracy. The JMP statistical analysis software (ver. 12; SAS, Cary, NC, USA) was used in all analyses.

Evaluation Using LRG for Small-Bowel Ulcerative Lesions
Based  Table S1).
The LRG value with the highest area under the curve (AUC) was calculated from the ROC curves obtained in the presence and absence of small intestinal ulcerative lesions: LRG 14 µg/mL (AUC 0.77, Figure 2). The use of LRG 14 µg/mL to detect the presence of small intestinal ulcerative lesions showed 63.6% sensitivity, 82.8% specificity, 58.3% PPV, 85.7% NPV, and 78% accuracy.
The LRG value with the highest area under the curve (AUC) was calculated from the ROC curves obtained in the presence and absence of small intestinal ulcerative lesions: LRG 14 μg/mL (AUC 0.77, Figure 2). The use of LRG 14 μg/mL to detect the presence of small intestinal ulcerative lesions showed 63.6% sensitivity, 82.8% specificity, 58.3% PPV, 85.7% NPV, and 78% accuracy.

Discussion
While it has been reported that the presence of small-bowel lesions in CD cannot be determined by clinical activity or existing markers, small bowel ulcerative lesions ≥ 0.5 cm are an independent risk factor for relapse and surgery [3], and the monitoring of such lesions is of high clinical significance even in patients with CD who are judged to be in clinical remission. We therefore investigated whether the use of a patient's LRG value could predict the presence or absence of small bowel ulcerative lesions in quiescent patients with CD in clinical remission with a CRP level < 0.5 mg/dL. The results showed that a lower LRG value of 14 µg/mL rather than the previously reported LRG cutoff value of 16 µg/mL significantly predicted the presence of small intestinal ulcerative lesions ≥ 0.5 cm in diameter.
The prediction of ≥0.5 cm small bowel ulcerative lesions by LRG values was reported in a study of BAE [14]. In that report, the detection sensitivity was 79%, with 82% specificity, 93% PPV, and 58% NPV at the cutoff value of 13.4 µg/mL LRG. Another report indicated that the LRG value predicted a modified SES-CD score of 0, which represents only the inflammatory findings of SES-CD [9]. In that report, AUC 0.82 and 77% accuracy were obtained at a cutoff LRG value of 13 µg/mL, indicating that the prediction ability of LRG was equivalent to that of calprotectin. The discrepancy between the findings obtained in those studies and our present investigation may be explained by the fact that our patient series was limited to those with quiescent CD with clinical remission and negative CRP values, and the monitoring modalities used were different. The use of SBCE, which can image the entire small bowel, may have increased the visualized bowel length and contributed to the identification of lesions, resulting in different cutoff values.
Although the relationship between SBCE scores and relapse risk is still unclear, it was reported that an LS ≥ 350 is useful for predicting relapse after 24 months post baseline [13]. In the present study, the use of the LRG value < 14 µg/mL was able to significantly discriminate patients with an LS < 350 with 80% specificity and the NPV of 100%. This suggests that the use of LRG may be able to select patients with inflammatory findingseven those in clinical remission-who should be prioritized for a small bowel search. Our analyses also revealed that LRG was correlated with each SBCE score (LS, CECDAI, and CDACE) in patients with quiescent CD. Thus, LRG may reflect the extent of small bowel lesions detectable by SBCE that are not reflected by the CDAI or the CRP level.
In addition, biologic use status was associated with significantly lower LRG values and significantly lower capsule endoscopy scores in cases with CDAI < 150 and negative CRP. On the other hand, LS ≥ 350 and the presence of small intestinal ulcerative lesions larger than 0.5 cm did not differ significantly. It is necessary to be careful in interpreting these results, as the fact of using biologics may be influenced by the high potential disease activity, among other factors. However, it may be suggested that LRG measurement may be one of the indicators to evaluate therapeutic intervention in small intestinal lesions as well.
The strength of this study is that it is the first to compare LRG values of the entire small bowel in patients with CD with the use of SBCE, a modality that enables the easy observation of the entire small bowel; however, SBCE is expensive, and its indications must be carefully considered. It may be possible to predict patients with active small-bowel lesion CD who need to be evaluated by SBCE by measuring the LRG value beforehand.
There are some study limitations to consider. The cutoff value for LRG may be different in patients with small-bowel stenosis that cannot be passed by SBCE. The study's design was retrospective and included a small number of patients (n = 40) at a single institution. Calprotectin was not measured; this is because the measurement of calprotectin in patients with CD is not covered by Japan's health insurance system. However, a comparison of the efficacy of LRG and calprotectin has been reported [9], and further investigation of this parameter is warranted. No bowel cleansing agents were used in preparation for SBCE in this study. Therefore, there could potentially be lesions that were not visualized by capsule endoscopy. In addition, since we included patients with small-bowel colonic CD and since only 30% of the patients underwent colonoscopy at the same time that their SBCE was conducted and their LRG was measured, the activity of the colonic lesions may have affected the LRG values. However, the CDAI and the CRP level have been shown to be poor reflectors of the activity of small-bowel lesions [18][19][20], and thus, active lesions of the colon may be more reflective of clinical activity. All of the patients included in this study had CDAI values < 150 and CRP levels < 0.5 mg/dL, which we believe reduced this effect.

Conclusions
We suggest that the determination of the LRG value may be useful for (1) predicting the presence of small bowel inflammation associated with relapse in patients with CD with CRP values < 0.5 mg/dL and a CDAI < 150, and (2) selecting patients for SBCE. However, a multicenter validation in a larger number of patients is needed to strengthen this conclusion.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/jcm11092494/s1, Figure S1: Small bowel ulcerative lesions < 0.5 cm; Figure S2: Small bowel ulcerative lesions ≥ 0.5 cm; Table S1: Comparison of the patients with and without small bowel ulcer (>0.5 cm); Table S2: Comparison of the patients with and without Lewis score ≥ 350; Table S3: Comparison with LRG ≥ 16 µg/mL as the cutoff; Table S4: Comparison by use of biologics.