Faecal chitinase 3-like 1 is a reliable marker as accurate as
faecal calprotectin in detecting endoscopic activity in adult
patients with in
flammatory bowel diseases
A. Buisson*,†, E. Vazeille*,†, R. Minet-Quinard‡, M. Goutte*,†, D. Bouvier‡, F. Goutorbe*, B. Pereira§, N. Barnich†& G. Bommelaer*,†
*Gastroenterology Department, University Hospital Estaing, Clermont-Ferrand, France.
†Microbes, Intestine, Inflammation and Susceptibility of the Host, UMR 1071 Inserm/Universite d’Auvergne, USC-INRA 2018, Clermont-Ferrand, France. ‡Biochemistry Laboratory, University Hospital G. Montpied, Clermont-Ferrand, France.
§Biostatistics Unit- DRCI, GM– Clermont-Ferrand University and Medical Center, Clermont-Ferrand, France.
Correspondence to:
Dr A. Buisson, Gastroenterology Department, University Hospital Estaing, 1 Place Aubrac, 63100 Clermont-Ferrand, France. E-mail: a_buisson@hotmail.fr Dr E. Vazeille, Microbes, Intestine, Inflammation and Susceptibility of the Host, UMR 1071 Inserm/Universite d’Auvergne, USC-INRA 2018, CBRV, 28 Place Henri Dunant, 63 000 Clermont-Ferrand, France. E-mail: emilie.vazeille@udamail.fr
Publication data
Submitted 14 October 2015 First decision 22 November 2015 Resubmitted 20 January 2016 Resubmitted 17 February 2016 Resubmitted 18 February 2016 Accepted 18 February 2016 EV Pub Online 8 March 2016 AB and EV contributed equally to this work
SUMMARY Background
Faecal biomarkers are emerging tools in the assessment of mucosal healing in inflammatory bowel diseases (IBDs).
Aim
To evaluate the accuracy of faecal chitinase 3-like 1(CHI3L1) compared to calpro-tectin in decalpro-tecting endoscopic activity in IBD.
Methods
Overall, 86 IBD adults underwent colonoscopy consecutively and prospectively, with Crohn’s disease Endoscopic Index of Severity (CDEIS) or Mayo endoscopic sub-score calculation for ulcerative colitis, and stool collection. Faecal calprotectin was measured using quantitative immunochromatographic testing. Faecal CHI3L1 was quantified by ELISA. CHI3L1 cut-off value was determined using a receiver-operat-ing curve.
Results
In 54 Crohn’s disease patients, faecal CHI3L1 (q = 0.70, P < 0.001) and calprotectin (q = 0.74, P < 0.001) levels correlated with CDEIS and were significantly increased in patients with endoscopic ulceration. In patients with ileal Crohn’s disease, faecal CHI3L1 seemed to be better correlated with CDEIS than faecal calprotectin (q = 0.78 vs. q = 0.62, P < 0.001 for both). CHI3L1 > 15 ng/g detected endoscopic ulceration in Crohn’s disease with a sensitivity of 100% and a specificity of 63.6%, compared to faecal calprotectin> 250 lg/g showing a sensitivity of 90.5% and a specificity of 59.1%.
In 32 ulcerative colitis patients, faecal CHI3L1 and calprotectin levels correlated with Mayo endoscopic subscore (q = 0.44 and 0.61, respectively, P < 0.001 for both) and were significantly increased in ulcerative colitis patients with endoscopic activity. In ulcerative colitis patients, faecal CHI3L1 > 15 ng/g predicted endoscopic activity with a sensitivity of 81.8% and a specificity of 80.0%, compared to faecal calprotectin>250 lg/g showing a sensitivity of 86.4% and a specificity of 80.0%. Conclusion
Faecal CHI3L1 is a reliable biomarker in detecting endoscopic activity in IBD.
INTRODUCTION
Inflammatory bowel diseases (IBDs), including Crohn’s disease and ulcerative colitis, are chronic relapsing and remitting disorders and can lead to digestive damage and disability.1–4 To date, mucosal healing should be the therapeutic goal in IBD, as it is associated with sustained clinical remission, reduced rates of hospitali-sation and operations as well as a lower incidence of colorectal cancer.5–8 The definition of mucosal healing remains debated, but is mostly defined as the disap-pearance of ulceration.9, 10 Repeated colonoscopies are needed to monitor IBD and to check the achievement of this therapeutic goal. However, the burden experi-enced by patients and the potential risks11 have led IBD physicians to look for alternative non-invasive approaches.
One of the most attractive methods of evaluating IBD is the measurement of inflammatory proteins in the stools. Faecal calprotectin has been investigated for more than 10 years in IBD.12–26 Faecal calprotectin is able to differentiate IBD from irritable bowel syndrome patients, to predict clinical relapse and to avoid unnecessary colonoscopies in IBD patients.13, 16, 18, 21–24, 27 Recently, some authors have reported that faecal calprotectin cor-related with endoscopic scores and was able to detect endoscopic ulceration in IBD.17, 26, 28–30 As a conse-quence, faecal calprotectin has been increasingly used in the diagnosis and the monitoring of Crohn’s disease in daily practice.20 However, some studies suggest that fae-cal fae-calprotectin results are less relevant in some condi-tions, such as ileal involvement.17, 28, 29
Chitinase 3-like 1(CH3L1) also called YKL-40 is a gly-coprotein exhibiting a strong binding affinity to chitin, an abundant polysaccharide found in the cell walls (bac-teria, fungi and others), with no apparent glycohydrolase enzymatic activity.31 CH3L1 is expressed in a variety of cells (macrophages, neutrophils, fibroblasts, vascular smooth muscle cells, endothelial cells, epithelial cells, etc.) and is strongly induced at late stages of human macrophage differentiation. In addition, the dysregula-tion of CH3L1 was observed in several human diseases characterised by acute or chronic inflammation and tis-sue remodeling.32
Recently, Mizoguchi et al. reported that CHI3L1 is specifically upregulated in the colonic epithelial cells and lamina proprial macrophages in the inflamed mucosa from experimental murine colitis and IBD patients.33–35 The overexpression of CHI3L1 also increases the adher-ent-invasive Escherichia coli ability to colonise intestinal
epithelium.36 Aomatsu et al. showed a correlation between faecal CHI3L1 level and endoscopic scores in a paediatrics cohort.37 Whether this correlation remains in adults IBD patients remains unknown.
In this study, we aimed to evaluate the correlation between faecal CHI3L1 level and endoscopic scores in adult IBD patients; to determine the best threshold of faecal CHI3L1 to detect endoscopic activity in IBD and to assess the performance of faecal CHI3L1 cut-off in detecting endoscopic activity in IBD compared to faecal calprotectin.
METHODS
Ethical considerations
This study was performed in accordance with the Decla-ration of Helsinki, Good Clinical Practice and applicable regulatory requirements. The study was approved by local Ethics Committee (Ref 2015/CE76).
Patients
We led an observational study of a single-centre cohort in which standardised evaluation was performed by experienced clinicians, in all patients. Patients from the Clermont–Ferrand IBD Unit with an established diagno-sis of Crohn’s disease or ulcerative colitis38 undergoing ileocolonoscopy, regardless of the indication, were prospectively and consecutively included between December 2014 and June 2015. Clinical parameters including the Crohn’s Disease Activity Index (CDAI) and the Simple Clinical Colitis Activity Index (SCCAI) for ulcerative colitis were collected (Table 1).39 Blood samples were taken prior to the endoscopy and were used to measure high sensitive serum C-reactive protein (CRP) level by immunonephelemetric method (Vista; Siemens, Berlin, Germany).
Patients who took NSAIDs or aspirin within the 4 weeks prior to the measurement of calprotectin were not included.
Endoscopy
Patients followed a bowel cleansing protocol via oral ingestion of 2L of PEG (Fortrans, Ipsen Pharma, Paris, France) the evening prior to and 2L the morning of the examination. Endoscopies were performed under anaes-thesia with propofol (Propofol Dakota Pharm; Sano fi-Aventis, Paris, France), by two experienced endoscopists (AB, GB), using column video colonoscopy (QFC L 140; Olympus, Tokyo, Japan). The endoscopists were blinded
from the faecal calprotectin or CHI3L1 values. The severity and extent of endoscopic lesions were assessed by Crohn’s disease Endoscopic Index of Severity (CDEIS)40 for Crohn’s disease patients, and the Mayo endoscopic subscore in ulcerative colitis, as routinely used in the department.
Faecal calprotectin and CHI3L1 measurement
Stools were collected in the morning the day before the endoscopy to reduce intraindividual variation, and imme-diately stored at 4 °C. The bowel cleansing was started in all patients after stool collection. Patients were instructed to transport the stool samples in a dedicated container at
Table 1 | Characteristics of the inflammatory bowel disease patients included in the study
Crohn’s disease patients (n = 54) Ulcerative colitis patients (n = 32) Age, mean (s.d.) 36.3 16.4 42.4 14.5
Disease duration at inclusion, mean (s.d.) 7.9 10.8 7.5 9.2
Male gender,n (%) 22 (40.7) 16 (50.0)
Tobacco use,n (%)
Nonsmokers 21 (38.9) 16 (50.0)
Former smokers 13 (24.1) 11 (34.4)
Active smokers 20 (37.0) 5 (15.6)
Previous intestinal resection,n (%) 18 (33.3) – Montreal classification Disease location,n (%) – L1 15 (27.7) – L2 11 (20.4) – L3 28 (51.9) – L4 0 (0.0) – Behaviour,n (%) B1 25 (46.3) – B2 16 (29.6) – B3 13 (24.1) – Perianal lesions,n (%) 16 (29.6) – Extension,n (%) – E1 – 4 (12.5) E2 – 15 (46.9) E3 – 13 (40.6) Concomitant therapies,n (%) 5-ASA 8 (14.8) 16 (50.0) Corticosteroids 7 (13.0) 11 (34.4) Budesonide 3 (5.6) 0 (0.0) Thiopurines 17 (31.5) 7 (21.8) Methotrexate 2 (3.7) 0 (0.0) Infliximab 17 (31.5) 8 (25.0) Adalimumab 13 (24.1) 3 (9.4) Ustekinumab 1 (1.9) 0 (0.0) Vedolizumab 0 (0.0) 1 (3.1)
CDAI, median [IQR] 162.5 [64.5–231.0] – Total CDEIS, median [IQR] 3.25 [0.0–5.0] –
SCCAI, median [IQR] – 5 [2–8]
Mayo endoscopic subscore, median [IQR] – 2 [1–3] CRP, median [IQR] 8.6 [3.0–24.5] 4 [2.9–18.4] Faecal calprotectin (lg/g) median [IQR] 499.0 [145.0–1800.0] 1762.5 [129.5–1800.0] Faecal chitinase 3-like 1 (ng/g) median [IQR] 46.5 [6.0–208.0] 89.5 [6–919] Indications for colonoscopy,n (%)
Monitoring of endoscopic activity after therapy 32 (59.3) 14 (43.7) Evaluation of endoscopic
severity before switching therapy
22 (40.7) 16 (50.0)
Screening for colorectal dysplasia 0 (0.0) 2 (6.3)
s.d., standard deviation; IQR, interquartile;n, number; CDAI, Crohn’s Disease Activity Index; CDEIS, Crohn’s disease endoscopic index of severity; SCCAI, Simple Clinical Colitis Activity Index; CRP, C-reactive protein.
4 °C. Faecal samples were immediately transferred, upon patient arrival, to the Clermont–Ferrand hospital Bio-chemistry Laboratory. Stool cultures were performed on all IBD samples to exclude gastrointestinal infection. Cal-protectin was measured, as routinely performed in our IBD centre, using quantitative immunochromatographic test Quantum Blue High Range (B€uhlmann Laboratories AG, Sch€onenbuch, Switzerland), according to the manu-facturer’s instructions. Laboratory personnel, who were blinded from the current clinical and endoscopic disease activity of the patients, performed the analyses. The lower and the upper limits of detection for calprotectin were 100 and 1800 lg/g, respectively. Consequently, all calprotectin levels <100 and >1800 lg/g were considered as equal to 100 and 1800lg/g, respectively.
From the same stool samples, faecal CHI3L1 was mea-sured (duplicates) using Human chitinase 3-like 1 quan-tikine ELISA kit (R&D systems; Minneapolis, MN, USA) in the laboratory ‘Microbes, Intestine, Inflammation and Susceptibility of the Host, UMR 1071 Inserm/Universite d’Auvergne; USC-INRA 2018’ Clermont–Ferrand, France. Laboratory personnel, who were blinded from the current clinical and endoscopic disease activity of the patients and calprotectin results, performed the analyses. Results were given in ng/g of stool.
Data managing and statistical analysis
Study data were collected and managed using REDCap electronic data capture tools hosted at Clermont-Ferrand University Hospital.41 REDCap (Research Electronic Data Capture) is a secure, web-based application designed to support data capture for research studies, providing (i) an intuitive interface for validated data entry; (ii) audit trails for tracking data manipulation and export procedures; (iii) automated export procedures for seamless data downloads to common statistical packages and (iv) procedures for importing data from external sources.
Statistical analysis was performed usingSTATAsoftware (version 13, StataCorp, College Station, USA). The tests were two sided, with a Type I error set ata = 0.05. Base-line characteristics were presented for each pathology [Crohn’s disease patients (n = 54) and ulcerative colitis patients (n= 32)] as mean (s.d.) or median [interquar-tile range] according to statistical distribution (assump-tion of normality assessed using the Shapiro–Wilk test) for continuous data and as the number of patients and associated percentages for categorical parameters. Corre-lation coefficients (Pearson or Spearman according to statistical distributions) were calculated to study relations
between quantitative parameters. Comparisons of patients’ characteristics between the independent groups (e.g. no endoscopic ulceration and endoscopic ulceration) were performed using chi-squared or Fisher’s exact tests for categorical variables, and Student’s t-test or Mann– Whitney test if assumptions of t-test were not met (i) normality and (ii) assumption of homoscedasticity studied using Fisher–Snedecor test for quantitative parameters (faecal calprotectin and faecal CHI3L1). A ROC curve was used to determine the best CHI3L1 threshold to detect endoscopic ulceration in Crohn’s dis-ease and ulcerative colitis, taking into account the clini-cal relevance and usual recommended indexes (Youden, Liu and efficiency). Sensitivity, specificity, predictive val-ues (negative and positive) and likelihood ratios (nega-tive and posi(nega-tive) were presented with 95% confidence intervals for each estimated threshold. As biomarker specificity and sensitivity depend on disease prevalence, we performed a calculation of a summary measure like net benefit42 that takes into account the relative clinical costs of false positive and false-negative tests in the setting of varying disease prevalence. We calculated the weighted comparison (WC)42, 43 WC= D sensitiv-ity+[(1 p/p) 9 relative cost(false positives/true posi-tives)9 Dspecificity], where p = prevalence.
The relative misclassification cost (false positives/true positives) is a parameter that seeks to define how many false positives a true positive is worth. Clearly, such a ‘cost’ is very difficult to estimate. In the context of detec-tion of ulceradetec-tion in Crohn’s disease and endoscopic activity in ulcerative colitis, it may be argued that high test sensitivity – in order to identify most of true posi-tives – with the accompanying risk of false positives (e.g. unnecessary colonoscopies or cross-sectional imaging) is more acceptable than tests with low sensitivity but high specificity which risk false-negative diagnoses (i.e. missing active IBD leading to an increased risk of dis-ease complications). Previous publications considering pancreatic cancer screening42 or dementia diagnosis44 used an arbitrary value of relative cost equal to 0.1 con-sidering a 10-fold higher relative weighting for true-posi-tive diagnoses compared with false-positrue-posi-tive diagnoses. In addition, a work about colorectal or polyps screening determined 1/9 as the value of this relative cost.45 According to clinical relevance, we considered that miss-ing active IBD is probably less detrimental than missmiss-ing a pancreatic cancer. For this study, we decided to give the results using a ratio false positives/true posi-tives = 0.5. We also calculated the WC using a ratio of 0.1 and 1 to confirm that regardless the choice of this
value we obtained the same result either a benefit or a loss. To aid interpretation, another parameter has been calculated using WC, namely the equivalent increase in true-positive patients per 1000, using the equation42: weighted comparison 9 prevalence 9 1000. Whereas positive values were taken to indicate a net test benefit, negative values indicated a net loss.
RESULTS
Baseline characteristics of the patients
Overall, 86 IBD patients were included in the study. Among them, 54 Crohn’s disease patients were included (59.3% female), with a mean age of 36.3 16.4 years and a mean Crohn’s disease duration of 7.9 10.8 years at the inclusion time. Fifteen patients (27.5%) presented with pure ileal disease (L1 according to Montreal classifi-cation), 11 (20.4%) with colonic disease (L2) and 28 (51.9%) with ileocolonic Crohn’s disease (L3). The med-ian CDAI and CRP were 162.5 [64.5–231.0] and 8.6 [3.0–24.5] mg/L, respectively. Patients’ characteristics are shown in Table 1.
Endoscopic ulcerations were found in 23 Crohn’s dis-ease patients (42.6%), whereas 12 patients (22.2%) had severe endoscopic activity defined as CDEIS >6.
Among the 32 ulcerative colitis patients (50.0% female), the mean age at the time of inclusion was 42.4 14.5 years and the mean ulcerative colitis dura-tion was 7.5 9.2 years. While ulcerative colitis extent was limited to the rectum in 4 patients (12.5%), 15 patients (46.9%) had sigmoid/left colon involvement and 13 patients (40.6%) had pancolitis. The median SCCAI and CRP were 5 [2–8] and 4 [2.9–18.4] mg/L, respec-tively. Patients’ characteristics are given in Table 1. Overall, mucosal healing, defined as Mayo endoscopic subscore of 0 or 1, was observed in 9 ulcerative colitis patients (28.1%).
Crohn’s disease patients
Correlation studies. The mean faecal CHI3L1 level was 620.1 1455.3 ng/g. Considering the overall 54 Crohn’s disease patients, the faecal CHI3L1 value correlated with CDAI (q = 0.42, P = 0.001), CRP (q = 0.61, P < 0.001), faecal calprotectin value (q = 0.68, P < 0.001) and CDEIS (q = 0.70, P < 0.001) (Figure 1a). CDAI, CRP and faecal calprotectin value correlated also with CDEIS (q = 0.43, q = 0.64 and 0.74, respectively, P < 0.001 for each of them) (Figure 1b).
In Crohn’s disease patients with pure ileal involve-ment (L1 according to the Montreal classification), faecal
CHI3L1 and faecal calprotectin correlated with CDEIS (q = 0.78, P < 0.001 and q = 0.62, P < 0.001, respec-tively). Faecal CHI3L1 (q = 0.86, P = 0.002) and faecal calprotectin (q = 0.92, P < 0.001) correlated also with CDEIS in patients with pure colonic Crohn’s disease (L2 according to Montreal classification). Considering the Crohn’s disease patients with ileocolonic involvement (L3 according to the Montreal classification), we found a significant correlation between faecal CHI3L1 or faecal calprotectin levels and CDEIS with coefficient of, respec-tively, 0.65 (P< 0.001) and 0.75 (P < 0.001).
Detection of endoscopic ulceration. The mean value of faecal CHI3L1 was 15-fold higher in the Crohn’s disease patients presenting with endoscopic ulceration than in those with no ulcer (P= 0.001) (Table 2). We observed also a twofold increase in the mean level of faecal calpro-tectin in patients with ulcerated lesions compared to patients with no ulceration (P= 0.03) (Table 2). In
0 0 1000 6000 4000 2000 0 0 10 20 30 40 2000 3000 10 20 30 40
Faecal chitinase3-like 1 level (ng/g)
Faecal calprotectin level (µg/g)
CDEIS
CDEIS (a)
(b)
Figure 1 | Correlation between faecal calprotectin level (a) or faecal chitinase 3-like 1 (b) and Crohn’s disease endoscopic index of severity (CDEIS) in 54 Crohn’s disease patients.
addition, the levels of these two faecal biomarkers were significantly higher in patients experiencing severe endoscopic activity defined as CDEIS >6, with 1890 1953 ng/g vs. 399.6 1278 ng/g, P = 0.0008 for faecal CHI3L1 and 1771 89 lg/g vs. 681 694 lg/g, P = 0.0004 for faecal calprotectin. Using a ROC curve, we attempted to determine the best CHI3L1 threshold to detect the presence of endoscopic ulceration in Crohn’s disease. The area under the curve was 0.82 CI95%(0.71– 0.92) for faecal CHI3L1, whereas it was only 0.74 CI95% (0.62–0.87) for faecal calprotectin (Figure 2). We found that a CHI3L1 cut-off value of 15 ng/g detected Crohn’s disease endoscopic ulceration with a sensitivity and a specificity of 100% and 63.6%, respectively (Table 2). The most used threshold of faecal calprotectin, i.e.
250lg/g,28, 46showed a sensitivity of 90.5% and a speci-ficity of 59.1% (Table 2). The addition of the two dosages, i.e. faecal CHI3L1 and faecal calprotectin, did not improve the accuracy of the two markers in detect-ing endoscopic ulceration in Crohn’s disease.
The calculation of WC comparing faecal CHI3L1 to faecal calprotectin showed a value of 0.12 (ranging from 0.10 to 0.14 according to the value of the misclassi fica-tion cost). As WC is positive, there is an increased net benefit favouring CHI3L1. Accordingly, we calculated a benefit equivalent to 58 [49–70] more true-positive patient results per 1000 patients.
Ulcerative colitis patients
Correlation studies. The mean faecal CHI3L1 level was 1132.6 2658.9 ng/g. In the 32 included ulcerative colitis patients, the faecal CHI3L1 value correlated with SCCAI (q = 0.52, P = 0.001), plasma CRP concentration (q = 0.35, P < 0.001), faecal calprotectin value (q = 0.62, P < 0.001) and Mayo endoscopic subscore (q = 0.47, P < 0.001). SCCAI, CRP and faecal calprotectin values correlated also with Mayo endoscopic subscore (q = 0.55, 0.51, 0.61, respectively, P < 0.001 for each of them).
Detection of endoscopic activity. The mean values of faecal CHI3L1 or faecal calprotectin were significantly higher in the ulcerative colitis patients with Mayo endoscopic subscore of 2 or 3 than in those with mucosal healing defined as Mayo endoscopic subscore <2 (1541.0 3075.6 ng/g vs. 134.2 328.6 ng/g, P < 0.001 for faecal CHI3L1 and 1421.8 667.9 lg/g vs. 324.7 544.3 lg/g, P < 0.001 for faecal calpro-tectin) (Figure 3a and b). We found an 11-fold
1-Specificity 0.00 0.00 0.25 0.25 0.50 0.50 0.75 0.75 1.00 1.00 Sensitivity
Faecal chitinase 3-like 1: 0.87 Faecal calprotectin: 0.75
Reference
Figure 2 | Receiver-operating curve illustrating the performances of faecal chitinase 3-like 1 and faecal calprotectin in detecting endoscopic ulceration in 54 Crohn’s disease patients.
Table 2 | Performances of faecal chitinase 3-like 1 and calprotectin in detecting endoscopic ulceration, in 54 Crohn’s disease patients
Mean value of faecal marker (s.d.)
P-value
Cut-off
value AUC Se Spe NPV PPV PLR NLR No endoscopic ulceration Endoscopic ulceration Faecal calprotectin 643 730 lg/g 1236 684 lg/g 0.03 250lg/g 0.74 (0.62–0.87) 90.5% (69.6–98.8) 59.1% (36.4–79.3) 86.7% (59.5–98.3) 67.9% (47.6–84.1) 2.2 (1.3–3.7) 0.2 (0.1–0.6) Faecal chitinase 3-like 1 94.4 215 ng/g 1466 2079 ng/g 0.001 15 ng/g 0.82 (0.71–0.92) 100% (83.9–100) 63.6% (40.7–80.2) 100% (76.8–100) 72.4% (52.8–87.3) 2.8 (1.6–4.8) 0.0 (0.0–0.0)
s.d., standard deviation; AUC, area under the curve; Se, sensitivity; Spe, specificity; NPV, negative predictive value; PPV, positive predic-tive value; PLR, posipredic-tive likelihood ratio; NLR, negapredic-tive likelihood ratio. Values within parentheses represent 95% confident interval.
increase in mean CHI3L1 level in patients with endo-scopic activity compared to those with mucosal healing (P= 0.007), whereas mean calprotectin level was mul-tiplied by four (P= 0.0003). Using a ROC curve, we attempted to determine the best CHI3L1 threshold to detect the presence of endoscopic activity defined as Mayo ≥2.
The area under the curve was 0.81 CI95% (0.65– 0.97) and/or 0.83 CI95% (0.68–0.98) for faecal CHI3L1 and faecal calprotectin, respectively. In ulcerative colitis patients, 15 ng/g was the best CHI3L1 threshold to detect ulcerative colitis endoscopic activity with sensi-tivity and specificity of 81.6% and 80.0%, respectively (Table 3). The faecal calprotectin cut-off value of
1800 1600 1400 1200 1000 800 600 400 200 400 600 800 1000 1200 1400 1600 1800 4000 6000 8000 10 000 Mean Median 0 200 0
Mayo* < 2 Mayo* ≥2 Mayo* < 2
*Mayo = Mayo endoscopic sub-score 6 ng/g Mayo* ≥2 250 µg/g F aecal calprotectin le v el (µg/g) F
aecal chitinase 3-lik
e 1 le
v
el (ng/g)
(a)
(b)
Figure 3 | Faecal chitinase 3-like 1 (a) and faecal calprotectin (b) values according to endoscopic activity, defined as Mayo endoscopic subscore ≥2, in 32 ulcerative colitis patients.
Table 3 | Performances of faecal chitinase 3-like 1 and calprotectin in detecting endoscopic activity defined as Mayo endoscopic subscore≥2 in 32 ulcerative colitis patients
Mean value of faecal marker (s.d.)
P-value
Cut-off
value AUC Se Spe NPV PPV PLR NLR Mayo endoscopic subscore<2 Mayo endoscopic subscore≥2 Faecal calprotectin 324.7 544.3 lg/g 1421.8 667.9 lg/g 0.0002 250lg/g 0.83 (0.68–0.98) 86.4% (65.1–97.1) 80.0% (44.4–97.5) 72.7% (39.0–94.0) 90.5% (69.6–98.8) 4.3 (1.2–15.1) 0.2 (0.1–0.5) Faecal chitinase 3-like 1 134.2 328.6 ng/g 1541.0 3075.2 ng/g 0.007 15 ng/g 0.81 (0.65–0.97) 81.8% (59.7–94.8) 80.0% (44.4–97.5) 66.7% (34.9–90.1) 90.0% (68.3–98.8) 4.1 (1.2–14.4) 0.2 (0.1–0.6)
s.d., standard deviation; AUC, area under the curve; Se, sensitivity; Spe, specificity; NPV, negative predictive value; PPV, positive predic-tive value; PLR, posipredic-tive likelihood ratio; NLR, negapredic-tive likelihood ratio. Values within parentheses represent 95% confident interval.
250lg/g28, 46 showed a sensitivity of 86.4% and a specificity of 80.0% (Table 3). As faecal CHI3L1 was more sensitive and faecal calprotectin more specific in our cohort, we investigated the performances of mea-suring successively faecal CHI3L1 and faecal calpro-tectin. Among the ulcerative colitis patients with CHI3L1 level <15 ng/g, we found four patients with Mayo endoscopic subscore ≥2 (false negative). A faecal calprotectin cut-off value >250 lg/g allowed to accu-rately classify three of these four patients.
The calculation of WC comparing faecal CHI3L1 to faecal calprotectin showed a value of 0.04 (regardless the value of the misclassification cost as Dspeci-ficity = 0). As WC is negative, there is net loss for CHI3L1. Accordingly, we calculated a benefit equivalent to 27 more true-positive patient results per 1000 patients for faecal calprotectin.
DISCUSSION
This study is the first that reports the correlation between faecal CHI3L1 value and endoscopic scores in adults suffering from IBD and shows the performances of faecal CHI3L1 measurement in detecting endoscopic ulceration in Crohn’s disease and endoscopic activity in ulcerative colitis adults.
As clinical indices are not reliable in assessing mor-phological disease activity and in predicting the disease course,47, 48 endoscopy remains the reference to monitor IBD patients. IBD physicians are currently developing non-invasive biomarkers which would be more accept-able than endoscopy to manage IBD in daily practice. Although serum markers could be helpful, we consider that faecal markers, reflecting digestive inflammation, are more specific in IBD. Faecal biomarkers can be divided schematically into three categories by their sources: the faecal excretion of leucocytes, serum proteins or leuco-cyte products and the faecal excretion of epithelial cells.37 In this study, we chose to investigate faecal CHI3L1, which is almost undetectable in healthy subject, and is derived mainly from epithelial cells and macro-phages, contrary to faecal calprotectin which is mainly excreted by neutrophils. In addition, it has been shown that CHI3L1 could have a role in the inflammatory pro-cess as it is induced by pro-inflammatory cytokines (Il1-b, Il-6 and TNF-a) and favoured the bacterial adhesion and invasion of the colonic epithelial cells, in particular adherent-invasive E. coli.33, 35, 49 Preliminary reports demonstrated that faecal CHI3L1 level is almost unde-tectable in healthy subjects.35, 37 We confirmed, in our work, that faecal CHI3L1 level was increased in IBD
patients with active disease as previously reported, and was able to differentiate moderate from severe endo-scopic activity in adults.35, 37 Aomatsu et al.37 described a correlation between faecal CHI3L1 level and simplified endoscopic score for Crohn’s disease (SES-CD)50 in a cohort of 25 paediatrics Crohn’s disease with a correla-tion coefficient of 0.61 (P = 0.001).50 Our data validated these preliminary results in a cohort of 54 adults with Crohn’s disease, reporting a correlation coefficient of 0.70 (P< 0.001). In addition, we compared this correla-tion to faecal calprotectin which has become the gold standard of the faecal markers in IBD. The correlation intensity was quite similar to faecal calprotectin (q = 0.74, P < 0.001). Faecal calprotectin measurement is often considered as less reliable in pure ileal Crohn’s disease,17, 28 although we suggested recently that the main endoscopic factors influencing faecal calprotectin value were the affected area and the presence of ulcera-tion, in a depth-related manner.29 In analysing the sub-group of Crohn’s disease with pure ileal involvement, we showed that the correlation seemed to be higher for CHI3L1 compared to faecal calprotectin (0.78 vs. 0.62), suggesting that CHI3L1 could be particularly interesting in ileal Crohn’s disease. In contrast, faecal calprotectin could be slightly more useful than CHI3L1 in pure colo-nic Crohn’s disease (q = 0.92 and 0.86, respectively). However, this tendency could be taken with caution and should not lead to draw any definitive conclusion because of the limited sample size of this subgroup. As mucosal healing is commonly considered as the disap-pearance of ulceration in Crohn’s disease,9 we investi-gated the performances of faecal markers to detect endoscopic ulceration. In this regard, we highlighted that a cut-off value of CHI3L1> 15 ng/g was close to the cut-off value previously used in paediatrics (13.6 ng/g). In addition, CHI3L1 threshold of 15 ng/g was superior to faecal calprotectin cut-off value of 250 lg/g in detect-ing Crohn’s disease endoscopic ulceration, suggestdetect-ing that faecal CHI3L1 was a promising inflammatory mar-ker in Crohn’s disease. In our cohort, faecal CHI3L1 showed sensitivity and negative predictive value both equal to 100%, illustrating the potential role in eliminat-ing the presence of endoscopic ulceration in Crohn’s dis-ease. In ulcerative colitis patients, we confirmed that faecal CHI3L1 level was increased in patients with active ulcerative colitis and we found, for the first time, a cor-relation between faecal CHI3L1 and the Mayo endo-scopic subscore. In ulcerative colitis, mucosal healing is admitted as Mayo endoscopic subscore <2.51 Accord-ingly, we investigated the performances of CHI3L1 and
calprotectin in detecting mucosal healing. We found sim-ilar performances between these two faecal markers (a sensitivity of 81.6% and a specificity of 80.0% for CHI3L1 > 15 ng/g vs. a sensitivity of 86.4% and a speci-ficity of 80.0% for calprotectin >250 lg/g). In addition, we suggested a complementarity between these two fae-cal markers in ulcerative colitis. However, these data should be confirmed in an independent larger cohort. Regarding the plot distribution of CHI3L1 values, we can observe an important gap between the faecal CHI3L1 levels of IBD patients with endoscopic ulcera-tion and those with no ulcer. We could hypothesise that CHI3L1 secretion was highly increased in presence of mucosal damage.
Although the role of CHI3L1 remains poorly investi-gated, it could exacerbate intestinal inflammation by enhancing bacterial adhesion and invasion in colonic epithelial cells.35 In this line, previous data have shown that CHI3L1 could promote adherent-invasive E. coli.49 As accumulating evidence suggested the implication of adherent-invasive E. coli in the Crohn’s disease gene-sis,52–54 faecal CHI3L1 dosage could be also tested as marker of microbiota-mediated Crohn’s disease which could benefit from therapeutics targeting the microbiota especially adherent-invasive E. coli. As CHI3L1 also acti-vates specific cellular signalling pathway in intestinal epithelial cells, including PI3K/AKT,34 further investiga-tions are warranted to extend our knowledge of this sur-rogate marker.
Our sample size could be considered as a limitation, although it was large enough to provide widely signi fi-cant results and remains the largest reported so far. In addition, we chose to not include healthy controls because several studies previously demonstrated that fae-cal CHI3L1 levels are almost undetectable in healthy subjects.35, 37 Our study has also several strengths, such
as the prospective design and the choice of a novel bio-marker deriving from sources (macrophages and epithe-lial cells) different from those of faecal calprotectin. Interestingly, we compared our novel marker to the ref-erence, i.e. faecal calprotectin, and could suggest a possi-ble complementarity between these two markers.
In conclusion, faecal CHI3L1 measurement could be a reliable assay to assess endoscopic activity both in Crohn’s disease and ulcerative colitis. These data should be confirmed in independent cohorts. The ability of fae-cal CHI3L1 to predict clinifae-cal relapse or favourable out-come under therapy remains to be addressed.
AUTHORSHIP
Guarantor of the article: Anthony Buisson.
Author contributions: A Buisson: study concept and design; acquisi-tion of data; analysis and interpretaacquisi-tion of data; drafting of the manuscript. E Vazeille: study concept and design; acquisition of data; analysis and interpretation of data; critical revision of the manuscript for important intellectual content. R Minet-Quinard: acquisition of data (biochemistry); analysis and interpretation of data; critical revision of the manuscript for important intellectual content. M Goutte: study concept and design; acquisition of data; analysis and interpretation of data; critical revision of the manu-script for important intellectual content. D Bouvier: acquisition of data (biochemistry); critical revision of the manuscript for important intellectual content. F Goutorbe: critical revision of the manuscript for important intellectual content. B Pereira: statistical analysis. N Barnich: study concept and design; critical revision of the manu-script for important intellectual content. G Bommelaer: study con-cept and design; critical revision of the manuscript for important intellectual content.
All authors approved the final version of the article, including the authorship list.
ACKNOWLEDGEMENTS
Pr Doug Altman and Dr Susan Mallett for the statistical section; Juliette Barrau and Etienne Rambaud; Kelly Shields.
Declaration of personal interests: AB has served as a speaker or a consultant for MSD, Abbvie, Ferring, Takeda and Vifor pharma. The other authors declare no conflict of interest.
Declaration of funding interests: None.
REFERENCES
1. Peyrin-Biroulet L, Loftus EV Jr, Colombel J-F, et al. The natural history of adult Crohn’s disease in population-based cohorts. Am J Gastroenterol 2010; 105: 289–97.
2. Pariente B, Cosnes J, Danese S, et al. Development of the Crohn’s disease digestive damage score, the Lemann score. Inflamm Bowel Dis 2011;17: 1415–22.
3. Pariente B, Mary J-Y, Danese S, et al. Development of the Lemann index to
assess digestive tract damage in patients with Crohn’s disease. Gastroenterology 2015;148: 52–63 e3.
4. Peyrin-Biroulet L, Cieza A, Sandborn WJ, et al. Development of thefirst disability index for inflammatory bowel disease based on the international classification of functioning, disability and health. Gut 2012;61: 241–7. 5. Baert F, Moortgat L, Van Assche G,
et al. Mucosal healing predicts
sustained clinical remission in patients with early-stage Crohn’s disease. Gastroenterology 2010;138: 463–8. 6. Schnitzler F, Fidder H, Ferrante M,
et al. Mucosal healing predicts long-term outcome of maintenance therapy with infliximab in Crohn’s disease. Inflamm Bowel 2009; 91: 295–301. 7. Frøslie KF, Jahnsen J, Moum BA, et al.
Mucosal healing in inflammatory bowel disease: results from a Norwegian
population-based cohort.
Gastroenterology 2007;133: 412–22. 8. Pineton de Chambrun G,
Peyrin-Biroulet L, Lemann M, et al. Clinical implications of mucosal healing for the management of IBD. Nat Rev Gastroenterol Hepatol 2010;7: 15–29. 9. Ferrante M, Colombel J-F, Sandborn
WJ, et al. Validation of endoscopic activity scores in patients with Crohn’s disease based on a post hoc analysis of data from SONIC. Gastroenterology 2013;145: 978–86 e5.
10. Colombel JF, Sandborn WJ, Reinisch W, et al. Infliximab, azathioprine, or combination therapy for Crohn’s disease. N Engl J Med 2010;362: 1383– 95.
11. Buisson A, Chevaux J-B, Hudziak H, et al. Colonoscopic perforations in inflammatory bowel disease: a retrospective study in a French referral centre. Dig Liver Dis 2013;45: 569–72. 12. Tibble JA, Sigthorsson G, Bridger S,
et al. Surrogate markers of intestinal inflammation are predictive of relapse in patients with inflammatory bowel disease. Gastroenterology 2000; 119: 15–22.
13. Costa F, Mumolo MG, Ceccarelli L, et al. Calprotectin is a stronger predictive marker of relapse in ulcerative colitis than in Crohn’s disease. Gut 2005;54: 364–8.
14. Sipponen T, K€arkk€ainen P, Savilahti E, et al. Correlation of faecal calprotectin and lactoferrin with an endoscopic score for Crohn’s disease and
histologicalfindings. Aliment Pharmacol Ther 2008;28: 1221–9.
15. Molander P, F€arkkil€a M, Ristim€aki A, et al. Does fecal calprotectin predict short-term relapse after stopping TNFa-blocking agents in inflammatory bowel disease patients in deep remission? J Crohns Colitis 2015; 9: 33–40.
16. van Rheenen PF, van de Vijver E, Fidler V. Faecal calprotectin for screening of patients with suspected inflammatory bowel disease: diagnostic meta-analysis. BMJ 2010;341: c3369. 17. Schoepfer AM, Beglinger C, Straumann
A, et al. Fecal calprotectin correlates more closely with the Simple Endoscopic Score for Crohn’s disease (SES-CD) than CRP, blood leukocytes, and the CDAI. Am J Gastroenterol 2010;105: 162–9.
18. Jensen MD, Kjeldsen J, Nathan T. Fecal calprotectin is equally sensitive in Crohn’s disease affecting the small bowel and colon. Scand J Gastroenterol 2011;46: 694–700.
19. Sipponen T, Savilahti E, K€arkk€ainen P, et al. Fecal calprotectin, lactoferrin, and endoscopic disease activity in
monitoring anti-TNF-alpha therapy for Crohn’s disease. Inflamm Bowel Dis 2008;14: 1392–8.
20. Louis E. Fecal calprotectin: towards a standardized use for inflammatory bowel disease management in routine practice. J Crohns Colitis 2015;9: 1–3. 21. Vieira A, Fang CB, Rolim EG, et al.
Inflammatory bowel disease activity assessed by fecal calprotectin and lactoferrin: correlation with laboratory parameters, clinical, endoscopic and histological indexes. BMC Res Notes 2009;2: 221.
22. Mooiweer E, Severs M, Schipper MEI, et al. Low fecal calprotectin predicts sustained clinical remission in inflammatory bowel disease patients: a plea for deep remission. J Crohns Colitis 2015;9: 50–5.
23. Langhorst J, Elsenbruch S, Koelzer J, et al. Noninvasive markers in the assessment of intestinal inflammation in inflammatory bowel diseases: performance of fecal lactoferrin, calprotectin, and PMN-elastase, CRP, and clinical indices. Am J Gastroenterol 2008;103: 162–9.
24. Røseth AG, Aadland E, Grzyb K. Normalization of faecal calprotectin: a predictor of mucosal healing in patients with inflammatory bowel disease. Scand J Gastroenterol 2004;39: 1017–20. 25. Lasson A, Stotzer P-O, €Ohman L, et al.
The intra-individual variability of faecal calprotectin: a prospective study in patients with active ulcerative colitis. J Crohns Colitis 2015;9: 26–32. 26. Mosli MH, Zou G, Garg SK, et al.
C-reactive protein, fecal calprotectin, and stool lactoferrin for detection of endoscopic activity in symptomatic inflammatory bowel disease patients: a systematic review and meta-analysis. Am J Gastroenterol 2015;110: 802–19; quiz 820.
27. D’Inca R, Dal Pont E, Di Leo V, et al. Calprotectin and lactoferrin in the assessment of intestinal inflammation and organic disease. Int J Colorectal Dis 2007;22: 429–37.
28. D’Haens G, Ferrante M, Vermeire S, et al. Fecal calprotectin is a surrogate marker for endoscopic lesions in inflammatory bowel disease. Inflamm Bowel Dis 2012;18: 2218–24.
29. Goutorbe F, Goutte M, Minet-Quinard R, et al. Endoscopic factors influencing calprotectin value in Crohn’s disease. J Crohns Colitis 2015;9: 1113–9. 30. Yamamoto T, Shimoyama T,
Matsumoto K. Consecutive monitoring of faecal calprotectin during mesalazine suppository therapy for active rectal inflammation in ulcerative colitis. Aliment Pharmacol Ther 2015;42: 549– 58.
31. Zhu Z, Zheng T, Homer RJ, et al. Acidic mammalian chitinase in asthmatic Th2 inflammation and IL-13 pathway activation. Science 2004;304: 1678–82.
32. Lee CG, Da Silva CA, Dela Cruz CS, et al. Role of chitin and chitinase/ chitinase-like proteins in inflammation, tissue remodeling, and injury. Annu Rev Physiol 2011;73: 479–501. 33. Kawada M, Chen C-C, Arihiro A, et al.
Chitinase 3-like-1 enhances bacterial adhesion to colonic epithelial cells through the interaction with bacterial chitin-binding protein. Lab Invest J Tech Methods Pathol 2008;88: 883–95. 34. Chen C-C, Llado V, Eurich K, et al.
Carbohydrate-binding motif in chitinase 3-like 1 (CHI3L1/YKL-40) specifically activates Akt signaling pathway in colonic epithelial cells. Clin Immunol 2011;140: 268–75.
35. Mizoguchi E. Chitinase 3-like-1 exacerbates intestinal inflammation by enhancing bacterial adhesion and invasion in colonic epithelial cells. Gastroenterology 2006;130: 398–411. 36. Low RN, Francis IR, Politoske D, et al.
Crohn’s disease evaluation: comparison of contrast-enhanced MR imaging and single-phase helical CT scanning. J Magn Reson Imaging 2000;11: 127–35. 37. Aomatsu T, Imaeda H, Matsumoto K, et al. Faecal chitinase 3-like-1: a novel biomarker of disease activity in paediatric inflammatory bowel disease. Aliment Pharmacol Ther 2011; 34: 941–8.
38. Lennard-Jones JE. Classification of inflammatory bowel disease. Scand J Gastroenterol Suppl 1989;170: 2–6; discussion 16–19.
39. Walmsley RS, Ayres RC, Pounder RE, et al. A simple clinical colitis activity index. Gut 1998;43: 29–32. 40. Mary JY, Modigliani R. Development
and validation of an endoscopic index of the severity for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Therapeutiques des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989;30: 983–9. 41. Harris PA, Taylor R, Thielke R, et al.
Research electronic data capture (REDCap)– a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42: 377–81.
42. Mallett S, Halligan S, Thompson M, et al. Interpreting diagnostic accuracy studies for patient care. BMJ 2012;345: e3999.
43. Moons KG, Stijnen T, Michel BC, et al. Application of treatment thresholds to diagnostic-test evaluation: an alternative to the comparison of areas under
receiver operating characteristic curves. Med Decis Mak Int J Soc Med Decis Mak 1997;17: 447–54.
44. Larner AJ. Comparing diagnostic accuracy of cognitive screening instruments: a weighted comparison approach. Dement Geriatr Cogn Dis Extra 2013;3: 60–5.
45. Halligan S, Mallett S, Altman DG, et al. Incremental benefit of computer-aided detection when used as a second and concurrent reader of CT colonographic data: multiobserver study. Radiology 2011;258: 469–76.
46. Nancey S, Boschetti G, Moussata D, et al. Neopterin is a novel reliable fecal marker as accurate as calprotectin for predicting endoscopic disease activity in patients with inflammatory bowel diseases. Inflamm Bowel Dis 2013; 19: 1043–52.
47. Peyrin-Biroulet L, Reinisch W, Colombel J-F, et al. Clinical disease
activity, C-reactive protein
normalisation and mucosal healing in Crohn’s disease in the SONIC trial. Gut 2014;63: 88–95.
48. Lahiff C, Safaie P, Awais A, et al. The Crohn’s disease activity index (CDAI) is similarly elevated in patients with Crohn’s disease and in patients with irritable bowel syndrome. Aliment Pharmacol Ther 2013;37: 786–94. 49. Low D, Tran HT, Lee I-A, et al.
Chitin-binding domains of Escherichia coli ChiA mediate interactions with intestinal epithelial cells in mice with colitis. Gastroenterology 2013;145: 602– 12, e9.
50. Daperno M, D’Haens G, van Assche G, et al. Development and validation of a new, simplified endoscopic activity score for Crohn’s disease: the SES-CD. Gastrointest Endosc 2004;60: 505–12. 51. Dignass A, Eliakim R, Magro F, et al.
Second European evidence-based
consensus on the diagnosis and management of ulcerative colitis part 1: definitions and diagnosis. J Crohns Colitis 2012;6: 965–90.
52. Darfeuille-Michaud A, Neut C, Barnich N, et al. Presence of adherent Escherichia coli strains in ileal mucosa of patients with Crohn’s disease. Gastroenterology 1998; 115: 1405–13.
53. Darfeuille-Michaud A, Boudeau J, Bulois P, et al. High prevalence of adherent-invasive Escherichia coli associated with ileal mucosa in Crohn’s disease. Gastroenterology 2004;127: 412–21.
54. Barnich N, Carvalho FA, Glasser AL, et al. CEACAM6 acts as a receptor for adherent-invasive E. coli, supporting ileal mucosa colonization in Crohn disease. J Clin Invest 2007; 76: 1566–74.
