Original article 733
Serum antibodies to microbial antigens for Crohn’s disease progression: a meta-analysis Yao Xionga,b, Gou-Zhen Wanga,b, Jie-Qiong Zhoua,b, Bing-Qing Xiaa,b, Xin-Ying Wanga,b and Bo Jianga,b Objectives This meta-analysis evaluated the stratification powers of four well-studied serum antibodies to microbial antigens [ASCA (anti-Saccharomyces cerevisiae), anti-OmpC (anti-outer-membrane protein C), anti-I2 (anti-Pseudomonas fluorescens-associated sequence I2), and anti-CBir1 (anti-bacterial flagellin)] in characterizing progression of Crohn’s disease (CD). Methods Pooled sensitivity, specificity, and diagnostic odds ratio (DOR) with 95% confidence intervals (CI) for individual antibodies and antibody combination were used to evaluate and compare their stratification powers for CD-related complications and the need for surgery. Results Eleven studies were included in this metaanalysis. In terms of the outcomes for CD complication and surgery, ASCA had the highest sensitivities at 0.66 (CI 0.63–0.69) for complications and 0.66 (CI 0.63–0.68) for surgery, whereas anti-OmpC had the highest specificities at 0.83 (CI 0.80–0.85) for complications and 0.81 (CI 0.79–0.83) for surgery. Anti-OmpC had the highest DORs at 2.61 (CI 2.16–3.15) for complications and 2.93 (CI 2.48–3.47) for surgery, and a combination of at least
Introduction Crohn’s disease (CD) is a heterogeneous disorder manifesting with diverse clinical presentations. Although some patients will experience a benign clinical course, B80% can expect to develop complications including stricturing and penetrating disease behaviors and almost 50% will require surgery over a 10-year period [1,2]. The traditional treatment paradigm for CD includes a ‘stepup’ approach (e.g. topical steroids, ‘stepping up’ to systemic steroid use, followed by the use of immunosuppressants and biologicals if necessary). Although this approach may be effective in the near term, it may not prevent overall disease progression [3,4]. Recent evidence suggests that early effective therapies, such as the use of biologics and immunomodulators (referred to as the ‘top-down’ approach), might minimize the progression of disease complications [5,6]. Biologics, however, are expensive and are associated with rare but serious events including opportunistic infections and malignancies [7]. Therefore, it is important for physicians to identify factors that can stratify patients into at-risk populations for disease progression who would most benefit from an early and effective intervention. c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0954-691X
two antibodies presented pooled DORs at 2.93 (CI 2.42–3.56) for complications and 3.39 (CI 2.73–4.20) for surgery, superior to any single antibody. Conclusion Anti-OmpC had the highest stratification power among the four antibodies screened for the risk of both complications and surgery in CD patients, and the power became stronger when antibodies were assessed c in combination. Eur J Gastroenterol Hepatol 26:733–742 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. European Journal of Gastroenterology & Hepatology 2014, 26:733–742 Keywords: complication, Crohn’s disease, meta-analysis, microbial antigens, serum antibodies, surgery a Department of Gastroenterology, Nanfang Hospital, Southern Medical University and bGuangdong Provincial Key Laboratory of Gastroenterology, Guangzhou, China
Correspondence to Xin-Ying Wang and Bo Jiang, PhD, Department of Gastroenterology, Nanfang Hospital, Southern Medical University, Guangzhou 510515, China Tel: + 86 206 164 1540; fax: + 86 206 278 7385; e-mails: [email protected] and [email protected] Received 11 January 2014 Accepted 14 March 2014
Although current prognostic markers of CD rely on clinical risk factors such as the need for cortocosteroids, disease onset before the age of 40 years, smoking, the presence of perianal lesions at diagnosis, small bowel localization, and stricturing disease [8,9], these factors are neither sufficient nor accurate [10–13]. Identification of genetic markers has also been proposed for establishing the prognosis of CD [12]. However, the predictive power for the disease course of the NOD2/CARD15 genotype (the first and most promising genes associated with CD risk stratification) was weak, with poor area under the receiver operating characteristic curve, as shown recently by Adler et al. [14]. With emerging data showing that serologic antibodies to microbial antigens are associated with more aggressive CD phenotypes as well as a higher risk for surgery [15–20], these antibodies have become attractive prognostic markers. More than 10 kinds of antimicrobial serologic antibodies have been identified to be relevant to CD, including ASCA (antibodies directed against the oligomannan component of Saccharomyces cerevisiae, one of the members of anti-glycan antibodies), anti-OmpC (antibodies to the outer-membrane porin C of Escherichia DOI: 10.1097/MEG.0000000000000102
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
734
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
coli), anti-I2 (antibodies to the Pseudomonas fluorescensassociated sequence I2), anti-CBir1 (antibodies against bacterial flagellin), and other newly discovered antiglycan antibodies (AMCA, ALCA, ACCA, Anti-L, and Anti-C), and anti-flagellin antibodies (antibodies to A4Fla2 and Fla-X). Anti-glycan antibodies target cell wall carbohydrate epitopes found in microbiota such as yeasts and bacteria [21]. The most prominent member of this group of antibodies is ASCA, which targets the 200 kDa phosphopeptidomannan, a cell wall mannan of the common baker’s or brewer’s yeast S. cerveisiae [22]. Anti-OmpC is an antibody to an outer-membrane porin isolated from E. coli, and OmpC has been shown to help adherent-invasive E. coli thrive in the GI tract [23]. AntiI2 is an antibody to a bacterial DNA fragment that is a homolog of the bacterial transcription factor family derived from P. fluorescens [24]. Anti-CBir1 is an antibody to flagellin that is the primary structural component of bacterial flagella, and flagellin CBir1 has been shown to induce colitis in immunodeficient mice [25]. All of these antibodies have been shown to possess a certain degree of association with complicated disease behaviors or surgery in CD patients, but the results and conclusions of the existing studies [15–20] differ for each antibody. Therefore, before serological markers can aid in determining disease prognosis, their respective stratification powers and their specificities and sensitivities have to be defined clearly for their abilities to determine disease progression. This study describes a meta-analysis of serum antibodies to microbial antigens as stratification indicators of the progression and severity of CD. Here, we focused on four antibodies: ASCA, anti-OmpC, anti-I2, and anti-CBir1 because they have been the most well-characterized antibodies over the last 10 years [26] and represent antibody types specific for different microbial targets. As many studies [15,27,28] have found an association between the number of antimicrobial antibodies and CD phenotypes and disease progression (the so-called ‘serology dosage’ effect), we also assessed these antibodies in combination. To our knowledge, this is the first meta-analysis to evaluate the stratification powers of the four well-studied serum antibodies in determining complications and surgery in CD patients, although some comments and reviews [12,26,29] have been published on this subject and two meta-analyses have been carried out on the association between ASCA [30] and antiglycan antibodies [31] with progression of CD.
Methods
combination: Crohn’s, Crohn, inflammatory bowel disease, ASCA, anti-Saccharomyces cerevisiae antibodies, I2, anti-I2, antibodies to the Crohn’s disease-related bacterial sequences, Pseudomonas fluorescens-related protein, antiEscherichia coli outer membrane porin C, anti-OmpC, anti-outer-membrane protein C, Omp, anti-flagellin, antiCBir1, and antibodies to flagellin. The reference lists of eligible studies and review articles were also checked to identify other publications relevant to this topic. Inclusion/exclusion criteria
Studies were included in the present analysis when the following criteria were fulfilled: (a) comparison in humans of at least two of the four antibodies [ASCA (or gASCA), anti-OmpC, anti-I2, anti-CBir1]; (b) analysis and comparison of associations between these markers with complications or surgery (we defined complications as stricturing and/or penetrating or non-inflammatory disease and surgery as CD-related surgery or small-bowel resections or abdominal surgery); (c) if raw numbers of each genotype–phenotype combination could be found or clearly calculated from the published data; (d) in cases of potential or suspected overlapping patient populations (on the basis of either location or authorship), the latest publication or the study with the largest numbers of patients was included in the meta-analysis; and (e) written in English. We excluded reviews, editorials, and meeting abstracts. Review processes and data extraction
Titles of all candidate articles were read, and for those that were not excluded solely on the basis of titles, the abstracts were reviewed. We then retrieved the full texts of published papers that were not excluded on the basis of the titles and abstracts alone. Each article was read to confirm whether it fulfilled the inclusion and exclusion criteria. Two reviewers assessed the studies selected on the basis of the authors, journal, year of publication, country, population source, total sample size, patient age, and serum antibodies measured. Study quality was assessed using the QUADAS (Quality Assessment of Studies of Diagnostic Accuracy Included in Systematic Reviews) checklist (maximum score, 14) [32]. In cases of discordance, a consensus was reached through discussion with the senior author. True-positive (TP), false-positive (FP), falsenegative (FN), and true-negative (TN) numbers were calculated for each study. Some included data were recomputed from original article descriptions that potentially resulted in slightly different values. Corresponding authors were contacted for missing data.
Database searches
Electronic databases including PubMed, Embase, Ovid, and Web of Science were searched to identify studies on the association between serum antimicrobial antibodies and CD complications or the need for surgery up to December 2013. This search strategy used a combination of the following MeSH headings and keywords alone or in
Statistical analysis
The primary outcome of interest for this analysis was the stratification powers of serum antibodies for CD progression, including complications and the need for surgery. Standard methods recommended for meta-analyses of diagnostic tests were used. Statistical analyses were
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Antimicrobial antibodies for CD progression Xiong et al. 735
carried out using Stata 12.0 (Stata Corporation, College Station, Texas, USA) and Meta-DiSc 1.4 [33]. P-values less than 0.05 were considered statistically significant. Pooled sensitivities, specificities, and diagnostic odds ratios (DORs) with 95% confidence interval (CI) were calculated. Here, the sensitivity means the odds of a positive antibody state in patients with an aggressive disease course (complications or surgery); the specificity means the odds of a negative antibody state in patients without an aggressive disease course (no complications or surgery); and the DOR is defined as the odds of a positive antibody state in patients with an aggressive disease course (complications or surgery) compared with the odds of a positive antibody state in patients without an aggressive disease course (no complications or surgery). A higher DOR indicated a higher association of serum antibodies and disease progression [34]. I2 was used to assess statistical heterogeneity, with values at least 50% indicating significant heterogeneity. The P-value of the Cochran’s Q-test was calculated to detect statistically significant heterogeneity across studies. We used the fixed-effects model by Mantel–Haenszel to calculate the summary DOR and its 95% CI among studies with homogeneous results, whereas the random-effects model by DerSimonian and Laird [35] was used when significant heterogeneity was found. Forest plots were derived to show the results of individual studies and pooled analyses. Deeks’ funnel plots [36] were used to detect any publication bias, for which a P-value less than 0.1 was suggestive of significant bias.
Results Study characteristics
The initial search of the databases identified 469 articles and among these, 11 articles [15–18,27,28,37–41] fulfilled the criteria established in this meta-analysis. The detailed steps of our literature search and reasons for exclusions are shown in Fig. 1. One study by Papp et al. [19] was excluded from the meta-analysis because it included anti-Omp, which is against multiple bacterial proteins derived from two species of intestinal bacteria according to the subsequent publication by the same author [42] and thus different from anti-OmpC in this study. All the 11 studies were retrospective and had been published between 2004 and 2013. Four studies were carried out in Europe, four in the USA, two in Canada, and one in Japan. Two studies included pediatric patients [16, 27] and the others focused primarily on adult patients. The clinical characteristics and QUADAS scores of these studies are summarized in Table 1. Complicated behavior
Three to six studies with adequate quantitative details were used for each analysis of the four individual antibodies and their combination in terms of the outcomes of the complicated behavior (Table 2 and Fig. 2). For individual antibodies, ASCA and antiCBir1 had higher sensitivities [0.66 (95% CI 0.63–0.69) and 0.60 (95% CI 0.56–0.65), respectively], but lower specificities [0.57 (95% CI 0.54–0.60) and 0.55 (95% CI 0.52–0.58), respectively], whereas anti-OmpC and anti-I2 had higher specificities [0.83 (95% CI 0.80–0.85) and 0.81
Fig. 1
469 articles identified in database search after duplicates removed
220 full-text articles assessed for eligibility
Records excluded (249): Case report (6) Reviews, commentary, or editorial (45) In vitro and animal studies (9) Not about CD (189)
Records excluded (209): Review, commentary, or case report (65) In vitro and animal studies (10) Not about IBD or the four antibodies (50) Different outcome (65) Did not compare two antibodies (11) No individual values of markers given (4)
11 articles included in this meta-analysis
Overlap of patients (1) Not in English (3)
Flow chart describing the selection criteria of the studies included in the meta-analysis. CD, Crohn’s disease; IBD, inflammatory bowel disease.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
736
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
Table 1
Patient characteristics of the 11 studies included in the meta-analyses
References, country
Mean or median age/mean or median age at CD patients diagnosis (years) no.
Source, population
Arnott et al., UK [15] Mow et al., USA [18] Dubinsky et al., USA [16]
Western General Hospital, Edinburgh Cedars-Sinai Medical Center Western Regional Pediatric IBD Research Alliance Ferrante et al., Belgium [17] The University Hospital in Leuven, Belgium Suzuki et al., Japan [40] The hospital of Hyogo College of Medicine Papadakis et al., USA [38] Cedars-Sinai Medical Center Dubinsky et al., USA [27]
Western Regional Pediatric IBD Research Alliance, the Pediatric IBD Collaborative Research Group, the Wisconsin Pediatric IBD Alliance O’Donnell et al., Ireland [37] A specialist IBD clinic at a university hospital Elkadri et al., Canada [28] Mount Sinai Hospital and The Hospital for Sick Children Ryan et al., Canada [39] Manitoba IBD Cohort Michielan et al., Italy [41]
The gastroenterology outpatient unit at the University Hospital of Padua
39/27 –/–(23)a 13/12
142 303 196
35/22 34.5/24.7 –/–
738 104 428
–/12
796
38/26 32.8/22.1 38.5/–
179 391 127
36.5/–
60
Serum antibodies measured
QUADAS quality score
ASCA, anti-OmpC, anti-I2 ASCA, anti-OmpC, anti-I2 ASCA, anti-OmpC, anti-Cbir1, anti-I2 ASCA, anti-OmpC ASCA, anti-I2 ASCA, anti-OmpC, anti-Cbir1, anti-I2 ASCA, anti-OmpC, anti-Cbir1
11 10 11
ASCA, anti-OmpC, anti-Cbir1 ASCA, anti-OmpC, anti-Cbir1 ASCA, anti-OmpC, anti-Cbir1, anti-I2 ASCA, anti-OmpC
11 11 10
9 10 11 9
11
anti-CBir1, antibodies against bacterial flagellin; anti-I2, antibodies to the Pseudomonas fluorescens-associated sequence I2; anti-OmpC, antibodies to the outer-membrane porin C of Escherichia coli; ASCA, antibodies directed against the oligomannan component of Saccharomyces cerevisiae; CD, Crohn’s disease; IBD, inflammatory bowel disease; QUADAS, Quality Assessment of Studies of Diagnostic Accuracy Included in Systematic Reviews. a Median age of onset.
(95% CI 0.75–0.86), respectively], but lower sensitivities [0.38 (95% CI 0.36–0.41) and 0.39 (95% CI 0.32–0.46), respectively]. When looking at the DORs, anti-OmpC had the highest value of 2.61 (95% CI 2.16–3.15) and anti-CBir1 had the lowest value of 1.94 (95% CI 1.58–2.38); ASCA and anti-I2 had intermediate values of 2.39 (95% CI 2.01–2.85) and 2.27 (95% CI 1.43–3.60), respectively. Analysis of antibody combination identified that more than one antibody resulted in a pooled DOR of 2.93 (95% CI 2.42–3.56) that was higher than any individual antibody (2.93 vs. 1.94–2.61), indicating a numerical but not statistically significant difference of having a greater chance of developing a complication.
Surgery
Three to eight studies that fulfilled the inclusion criteria were used for the analyses of individual antibodies and their combination in terms of the outcomes of surgery (Table 2 and Fig. 3). Similar to the results of the complicated behavior, the analysis of the need for surgery for individual antibodies showed that ASCA and antiCBir1 had higher sensitivities [0.66 (95% CI 0.63–0.68) and 0.61 (95% CI 0.57–0.65), respectively], but lower specificities [0.58 (95% CI 0.55–0.60) and 0.52 (95% CI 0.49–0.55), respectively], whereas anti-OmpC and anti-I2 had higher specificities [0.81 (95% CI 0.79–0.83) and 0.63 (95% CI 0.57–0.69), respectively], but lower sensitivities [0.43 (95% CI 0.41–0.46) and 0.59 (95% CI 0.53–0.65), respectively]. The DOR for surgery was the highest for anti-OmpC, with a value of 2.93 (95% CI 2.48–3.47), and the lowest for anti-CBir1, with a value of 1.83 (95% CI 1.52–2.20). The pooled DOR for surgery when a combination of at least two antibodies was used
was 3.39 (95% CI 2.73–4.20), which was also superior to any single antibody examined (3.39 vs. 1.83–2.93). Publication bias
Deeks’ funnel plots were used to evaluate publication bias of the articles, with the results of P-values shown in Table 2. Except for the analysis of ASCA, the tests showed a statistically nonsignificant value (P > 0.1), indicating that there was generally no potential publication bias.
Discussion Reliable prognostic markers are needed to identify CD patients who might benefit from the timely introduction of immunomodulators or biologicals. Although antibody responses toward microbial antigens have questionable clinical benefit as diagnostic tools because of their low sensitivities, they can be used to characterize specific and adverse outcomes associated with CD, making them promising prognostic markers. Before recommending the routine use of serological markers for predicting disease course, diagnostic studies reporting sensitivities and specificities of serological markers in terms of their stratification powers for CD progression are warranted. If found to be useful and their predictive powers are further proved in prospective studies, these markers could potentially be added to clinical indicators used to predict the risk for developing adverse outcomes associated with CD. Recently, two meta-analyses investigated the association between several serum antibodies against microbial antigens and CD progression. One meta-analysis carried out by Zhang et al. [30] focused on ASCA and suggested that a positive ASCA status was a risk factor for CD-related complications and surgery. The other
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
meta-analysis carried out by Kaul et al. [31] investigated a group of anti-glycan antibody biomarkers (ASCA, AMCA, ALCA, ACCA, anti-L, and anti-C) and showed that these anti-glycan antibodies had pooled DORs of around 2, with no prominent marker associated with disease progression. The results of the latter meta-analysis also showed that ACCA had the highest association with complications, whereas both ASCA and ACCA associated equally with the need for surgery and that the combination of more than one antibody had a higher association with disease course. These two meta-analyses only focused on one antibody or antibodies belonging to the group of antiglycan antibodies. Different from them, this metaanalysis examined four antibodies (ASCA, anti-OmpC, anti-I2, and anti-CBir1) that have recently been well studied in terms of their relationship with CD progression and are specific for different microbial targets. Our analyses for individual antibodies (Table 2 and Figs 2 and 3) indicated that ASCA and anti-CBir1 had higher sensitivities, but relatively lower specificities compared with anti-OmpC, which had the highest specificity, but the lowest sensitivity for complications and surgery. These results may reflect some difference in the prevalences of these antibodies in CD patients. The meta-analysis by Kamat et al. [43] showed that the prevalences of ASCA IgA/IgG (45%) and anti-CBir1 (55.2%) in CD patients were relatively high, whereas the prevalence of antiOmpC (29.4%) was relatively low. In terms of the DORs of individual antibodies in this study, anti-OmpC was shown to have the highest stratification power, which was superior to ASCA (the traditional and best-studied serum antibody). This suggests that future studies on serum antibodies should pay more attention toward the detection of anti-OmpC.
Sensitivity (95% CI) I2%, P-value of Q-test Specificity (95% CI) I2%, P-value of Q-test DOR (95% CI) I2%, P-value of Q-test Publication bias P-value
CI, confidence interval; DOR, diagnostic odds ratio. Other abbreviations as in Table 1. a Defined as either IgA or IgG ASCA positive.
0.59 (0.53–0.65) 95.7, 0.000 0.63 (0.57–0.69) 91.2, 0.000 2.71 (1.95–93.77) 0.0, 0.533 0.181
Arnott and colleagues [15,17,18,28,41] 0.47 (0.43–0.50) 96.4, 0.000 0.77 (0.74–0.80) 95.0, 0.000 3.39 (2.73–4.20) 0.0, 0.794 0.101 Dubinsky and colleagues [27,28,37,38,39] 0.61 (0.57–0.65) 74.9, 0.003 0.52 (0.49–0.55) 47.5, 0.107 1.83 (1.52–2.20) 12.0, 0.337 0.084 Arnott and colleagues [15,18,39]
Arnott and colleagues [15,17,18,27,28,37,39,41] 0.43 (0.41–0.46) 91.7, 0.000 0.81 (0.79–0.83) 91.1, 0.000 2.93 (2.48–3.47) 47.3, 0.065 0.669
Sensitivity (95% CI) I2%, P-value of Q-test Specificity (95% CI) I2%, P-value of Q-test DOR (95% CI) I2%, P-value of Q-test Publication bias P-value Surgery Studies included
Arnott and colleagues [15,17,18,27,28,41] 0.66 (0.63–0.68) 19.0, 0.290 0.58 (0.55–0.60) 86.0, 0.000 2.82 (2.03–3.91) 63.9, 0.017 0.028
0.39 (0.32–0.46) 85.7, 0.001 0.81 (0.75–0.86) 0.0, 0.765 2.27 (1.43–3.60) 4.5, 0.351 0.762
Arnott and colleagues [15,17,27,28,41] 0.42 (0.39–0.45) 96.6, 0.000 0.75 (0.72–0.78) 92.9, 0.000 2.93 (2.42–3.56) 0.0, 0.475 0.646 Dubinsky and colleagues [27,28,37,39] 0.60 (0.56–0.65) 71.9, 0.014 0.55 (0.52–0.58) 44.2, 0.146 1.94 (1.58–2.38) 5.9, 0.364 0.683 Ferrante and colleagues [17,27,28,37,39,41] 0.38 (0.35–0.41) 93.8, 0.000 0.83 (0.80–0.85) 88.1, 0.000 2.61 (2.16–3.15) 0.0, 0.546 0.779 Ferrante and colleagues [17,27,28,41] 0.66 (0.63–0.69) 63.6, 0.041 0.57 (0.54–0.60) 74.8, 0.008 2.39 (2.01–2.85) 7.4, 0.3563 0.028
Anti-OmpC ASCAa
Dubinsky and colleagues [27,39,40]
Combination (Z 2 antibodies)
Complication Studies included
Table 2
Pooled results of serum antibodies to microbial antigens
Anti-I2
Anti-CBir1
Antimicrobial antibodies for CD progression Xiong et al. 737
The present results also suggested that individual antibodies were useful stratification markers but their respective values indicated by sensitivities, specificities, and DORs were not as high as expected. The highest sensitivity, specificity, and DOR of individual antibodies for complications were 0.66, 0.83, and 2.61 (Table 2), respectively, with similar outcomes in terms of surgery. However, these findings of individual antibodies did not maximize the role that serum antibodies played in indicating the complications and surgery. As shown in Table 2 and Figs 2 and 3, the combination of at least two antibodies had an elevated DOR of 3.05 for complications and 3.16 for surgery, suggesting that analysis of the presence of antibodies in combination had a better stratification power. It is regrettable that we cannot find a way to meta-analyze the association between the number of antibodies with the disease course. However, it is important to note that an increasing number of positive antibodies [15,27,28] appeared to be associated with a higher risk for CD-related complications and surgery. Furthermore, elevated immune responses toward individual [17–19,44] or multiple [15,18,27,37] microbial
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
738
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
Fig. 2
Diagnostic OR (95% CI) (a)
2.34 (1.73 − 3.18) Ferrante et al. [17] Dubinsky et al. [27] 2.23 (1.62 − 3.08) Elkadri et al. [28] 2.43 (1.59 − 3.69) Michielan et al. [41] 6.57 (2.11 − 20.48) Fixed-effects model Pooled diagnostic odds ratio = 2.39 (2.01 − 2.85) Cochran-Q = 3.24; d.f. = 3 (P = 0.3563) Inconsistency (I 2) = 7.4%
ASCA
0.01
1
100.0
Anti-OmpC
1
0.01 (c)
Diagnostic OR (95% CI)
2.11 (1.49 − 2.97) Ferrante et al. [17] 2.73 (1.86 − 4.01) Dubinsky et al. [27] 1.16 (0.07 − 18.88) Ryan et al. [39] 3.15 (2.06 − 4.80) Elkadri et al. [28] 2.11 (0.70 − 6.35) Michielan et al. [41] 3.90 (1.94 − 7.83) O’Donnell et al. [37] Fixed-effects model Pooled diagnostic odds ratio = 2.61 (2.16 − 3.15) Cochran-Q = 4.03; d.f. = 5 (P = 0.5456) Inconsistency (I 2) = 0.0 %
(b)
100.0 Dubinsky et al. [16] Suzuki et al. [40] Ryan et al. [39]
Anti-I2
Diagnostic OR (95% CI) 2.68 (1.37 − 5.25) 3.67 (0.96 − 14.03) 1.32 (0.54 − 3.27)
Fixed-effects model Pooled diagnostic odds ratio = 2.27 (1.43 − 3.60) Cochran-Q = 2.09; d.f. = 2 (P = 0.3509) Inconsistency (I 2) = 4.5 % 1
0.1 (d)
10.0 Dubinsky et al. [27] Ryan et al. [39] Elkadri et al. [28] O’Donnell et al. [37]
Anti-CBir1
0.1
1
(e)
Combination (≥ 2 antibodies)
Diagnostic OR (95% CI) 1.77 (1.28 − 2.47) 1.26 (0.61 − 2.58) 2.50 (1.66 − 3.76) 2.05 (1.12 − 3.76)
Fixed-effects model Pooled diagnostic odds ratio = 1.94 (1.58 − 2.38) Cochran-Q = 3.19; d.f. = 3 (P = 0.3637) Inconsistency (I 2) = 5.9 % 10.0 Diagnostic OR (95% CI) Arnott et al. [15] 2.38 (1.14 − 4.99) Ferrante et al. [17] 3.23 (2.06 − 5.05) Dubinsky et al. [27] 2.43 (1.74 − 3.39) 3.89 (2.56 − 5.93) Elkadri et al. [28] 2.51 (0.76 − 8.37) Michielan et al. [41] Fixed-effects model Pooled diagnostic odds ratio = 2.93 (2.42 − 3.56) Cochran-Q = 3.52; d.f. = 4 (P = 0.4748) Inconsistency (I 2) = 0.0 %
0.01
1
100.0
Forest plots of pooled diagnostic odds ratios (DORs) for complication. (a) ASCA, (b) anti-OmpC, (c) anti-I2, (d) anti-CBir1, and (e) combination (Z 2 antibodies). Plots show pooled DOR probabilities of included studies with corresponding 95% confidence interval (CI), with circles indicating the area proportional to the study weight in the meta-analysis.
antigens were also shown to be associated with higher frequencies of CD complications and surgery. For example, a study by Mow et al. [18] showed that only 18% of patients with the lowest-level response toward three tested microbial antigens had small bowel surgery, whereas 90% of patients with the highest-level response toward these microbial antigens had small bowel surgery. On the basis of these discussions, the combined presence and levels of a
number of serum antibodies could be the strongest stratification markers. It should be noted that the studies included in this analysis were retrospective and were mostly cross-sectional, with samples taken at various random points during the disease course. Thus, the detection of the antibodies described in this report was more an indicator for the association of, rather
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Antimicrobial antibodies for CD progression Xiong et al. 739
Fig. 3
Diagnostic OR (95% CI) (a)
Mow et al. [18] Arnott et al. [15] Ferrante et al. [17] Dubinsky et al. [27] Elkadri et al. [28] Michielan et al. [41]
ASCA
1
0.01
100.0
Anti-OmpC
(1.97 – 5.04) (3.45 – 17.48) (1.59 – 2.89) (1.54 – 3.15) (1.33 – 3.06) (2.16 – 22.10)
Random effects model Pooled diagnostic odds ratio = 2.82 (2.03 – 3.91) Cochran-Q = 13.87; d.f. = 5 (P = 0.0165) Inconsistency (I 2) = 63.9 % τ2 = 0.0968 Mow et al. [18] Arnott et al. [15] Ferrante et al. [17] Dubinsky et al. [27] Ryan et al. [39] Elkadri et al. [28] Michielan et al. [41] O’Donnell et al. [37]
(b)
3.15 7.76 2.15 2.20 2.01 6.91
Diagnostic OR (95% CI) 2.01 (1.27 – 3.19) 1.35 (0.68 – 2.70) 2.82 (2.01 – 3.95) 4.08 (2.74 – 6.06) 5.24 (0.25 – 111.44) 4.36 (2.79 – 6.83) 2.67 (0.90 – 7.94) 2.95 (1.47 – 5.90)
Fixed effects model Pooled diagnostic odds ratio = 2.93 (2.48 – 3.47) Cochran-Q = 13.29; d.f. = 7 (P = 0.0654) Inconsistency (I 2) = 47.3 % 0.01
1
100.0 Mow et al. [18] Arnott et al. [15] Ryan et al. [39]
(c) Anti-I2
Diagnostic OR (95% CI) 2.76 (1.72 – 4.43) 3.37 (1.69 – 6.72) 1.75 (0.69 – 4.39)
Fixed effects model Pooled diagnostic odds ratio = 2.71 (1.95– 3.77) Cochran-Q = 1.26; d.f. = 2 (P = 0.5332) Inconsistency (I 2) = 0.0 % 1
0.1
10.0 Papadakis et al. [38] Dubinsky et al. [27] Ryan et al. [39] Elkadri et al. [28] O’Donnell et al. [37]
(d) Anti-CBir1
Diagnostic OR (95% CI) 1.73 (1.17 – 2.55)) 2.03 (1.41 – 2.93) 1.03 (0.50 – 2.10) 2.28 (1.51 – 3.44) 1.43 (0.75 – 2.74)
Fixed effects model Pooled diagnostic odds ratio = 1.83 (1.52 – 2.20) Cochran-Q = 4.55; d.f. = 4 (P = 0.3372) Inconsistency (I 2) = 12.0 % 0.1
1
10.0 Mow et al. [18] Arnott et al. [15] Ferrante et al. [17] Elkadri et al. [28] Michielan et al. [41]
(e) Combination (≥ 2 antibodies)
Diagnostic OR (95% CI) 2.89 (1.81 – 4.61) 2.85 (1.39 – 5.85) 4.16 (2.69 – 6.44) 3.39 (2.23 – 5.16) 2.67 (0.84 – 8.44)
Fixed effects model Pooled diagnostic odds ratio = 3.39 (2.73 – 4.20) Cochran-Q = 1.68; d.f. = 4 (P = 0.7939) Inconsistency (I 2) = 0.0 % 0.01
1
100.0
Forest plots of pooled diagnostic odds ratios (DORs) for surgery. (a) ASCA, (b) anti-OmpC, (c) anti-I2 (d), anti-CBir1, and (e) combination (Z 2 markers). Plots show pooled DOR probabilities of included studies with corresponding 95% confidence interval (CI), with circles indicating the area proportional to the study weight in the meta-analysis.
than a predictor of, a severe disease course. Despite this limitation, this study paved the way for opportunities to explore and validate the predictive value of these markers.
Some studies reported stable antibody responses over time [39,45,46] or after treatment or surgery [18,47,48]; hence, if a patient was seropositive early during disease, there
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
740
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
should be a concern that this patient may at some time develop a complicated disease or may require surgery. Recently, a few studies [15,27,49] have assessed the ability of antibodies as predictors of disease course with serum usually drawn at diagnosis or before the presentation of complications or surgery. Their results strongly support the use of antibodies to microbial antigens as prediction markers. However, large prospective studies are still required to evaluate their predictive values. In addition, to make these markers further useful in choosing patients appropriate for top-down therapy, prospective studies are also needed to verify that antibody-driven risk stratification alters the progression of CD. If the presence of antimicrobial antibodies indicates treatment-refractory disease with resistance to immunomodulators/biologicals, aggressive therapy would be ineffective. However, this possibility might be small as many studies [50–52] failed to show a relationship between serotype and response to infliximab. The mechanism behind the relationship between antibody responses toward microbial antigens and severe disease phenotypes has not yet been fully understood. There are now basically two hypotheses on the mechanism. One possibility is that increased intestinal permeability or chronic bacterial translocation caused by aggressive transmural inflammation leads to immune responses and antimicrobial antibody formation. Evidences supporting this hypothesis are that significant antimicrobial antibody expression was also reported in other clinical conditions with impaired mucosa such as celiac disease [53], liver cirrhosis [42], and AIDS [43], and that antibody response could be lost with mucosal healing in celiac disease patients [53]. Some researchers, however, argue that this hypothesis is less likely because research data also suggested that antibodies could be found in preclinical settings [54], and that their presence was not associated with intestinal permeability alterations [55,56] or disease activity [45]. Another possibility is that the distinct antibody response patterns may indicate unique pathophysiological mechanisms in CD progression and are thus important determinants of disease phenotype and disease progression. This is a corollary to the current theory supported by animal models that chronic intestinal inflammation is the result of an aberrant immunologic response to gut commensal bacteria, with genetic defects as predisposing factors [57–59]. Whether antibody responses are the consequences or the triggers of severe disease phenotypes is still in dispute and their role in the pathophysiology of CD should be examined further. This meta-analysis had certain limitations. First, inevitable biases were introduced by pooling different observational studies, reflected by the statistical heterogeneity present throughout the analysis. The heterogeneity observed could be because of various sources including pharmacological treatment, disease duration, age at diagnosis, ethnic origin, and different definitions of surgery. However, we could not obtain sufficient detailed information from the original studies to further stratify the analysis. Future studies should
pay more attention to these potential confounding factors. Second, the studies included in this analysis used different panels of antibodies in determining the combined prognostic power of these markers that could lead to heterogeneity in the combination analysis. As the respective studies did not specifically name the antibody combinations used, it was impossible to know which combination worked best. Third, most of the studies included in this meta-analysis [15,16,18,27,28,37–39] have carried out serum assays in Cedars-Sinai Medical Center and Prometheus Laboratories, and may thus simply reflect the experience of these few expert laboratories. The operation and availability of such assays should be made easier for local laboratories; thus, the widespread use of serum antibody markers in clinical practice could be feasible. In conclusion, ASCA had the highest sensitivity and antiOmpC had the highest specificity for both complications and the need for surgery, and overall, the presence of antiOmpC had the highest stratification values among the four serum antibodies screened. However, analyses of at least two antibodies were more effective than any individual marker alone in the context of CD progression. To be clinically useful in identifying patients who might be ideal for ‘top-down’ therapy, detection of serum antibodies to microbial antigens and their respective levels require further evaluation in prospective studies to better define their predictive powers.
Acknowledgements The authors thank Dr J.D. Ryan (University of Manitoba, Winnipeg, Canada), Dr Abdul A. Elkadri, and Dr Mark S. Silverberg (Mount Sinai Hospital, Toronto, Canada), Dr Dubinsky (Cedars-Sinai Medical Center, Los Angeles, USA), Dr Andrea Michielan (University of Padua, Padua, Italy), and Dr Severine Vermeire and Dr Marc Ferrante (University Hospital Gasthuisberg, Leuven, Belgium) for providing additional information about their studies. Comments of four anonymous reviewers have considerably helped to improve the manuscript. This study was funded by 2012B050600022 Science and Technology Planning Project of Guangdong Province. Conflicts of interest
There are no conflicts of interest.
References 1
2
3
4
Cosnes J, Cattan S, Blain A, Beaugerie L, Carbonnel F, Parc R, et al. Long-term evolution of disease behavior of Crohn’s disease. Inflamm Bowel Dis 2002; 8:244–250. Cosnes J, Nion-Larmurier I, Beaugerie L, Afchain P, Tiret E, Gendre JP. Impact of the increasing use of immunosuppressants in Crohn’s disease on the need for intestinal surgery. Gut 2005; 54:237–241. Faubion WA Jr, Loftus EV Jr, Harmsen WS, Zinsmeister AR, Sandborn WJ. The natural history of corticosteroid therapy for inflammatory bowel disease: a population-based study. Gastroenterology 2001; 121:255–260. Vermeire S, van Assche G, Rutgeerts P. Review article: altering the natural history of Crohn’s disease – evidence for and against current therapies. Aliment Pharmacol Ther 2007; 25:3–12.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Antimicrobial antibodies for CD progression Xiong et al. 741
5
6
7 8 9
10
11
12
13
14
15
16
17
18
19
20
21 22
23 24
25
26
27
28
Feagan BG, Panaccione R, Sandborn WJ, D’Haens GR, Schreiber S, Rutgeerts PJ, et al. Effects of adalimumab therapy on incidence of hospitalization and surgery in Crohn’s disease: results from the CHARM study. Gastroenterology 2008; 135:1493–1499. Peyrin-Biroulet L, Oussalah A, Williet N, Pillot C, Bresler L, Bigard MA. Impact of azathioprine and tumour necrosis factor antagonists on the need for surgery in newly diagnosed Crohn’s disease. Gut 2011; 60:930–936. Hanauer SB. Positioning biologic agents in the treatment of Crohn’s disease. Inflamm Bowel Dis 2009; 15:1570–1582. Beaugerie L, Seksik P, Nion-Larmurier I, Gendre JP, Cosnes J. Predictors of Crohn’s disease. Gastroenterology 2006; 130:650–656. Romberg-Camps MJ, Dagnelie PC, Kester AD, Hesselink-van de Kruijs MA, Cilissen M, Engels LG, et al. Influence of phenotype at diagnosis and of other potential prognostic factors on the course of inflammatory bowel disease. Am J Gastroenterol 2009; 104:371–383. Dias CC, Rodrigues PP, da Costa-Pereira A, Magro F. Clinical prognostic factors for disabling Crohn’s disease: a systematic review and meta-analysis. World J Gastroenterol 2013; 19:3866–3871. Cerqueira RM, Lago PM. Clinical factors predictive of Crohn’s disease complications and surgery. Eur J Gastroenterol Hepatol 2013; 25: 129–134. Lichtenstein GR. Emerging prognostic markers to determine Crohn’s disease natural history and improve management strategies: a review of recent literature. Gastroenterol Hepatol 2010; 6:99–107. Tamboli CP, Doman DB, Patel A. Current and future role of biomarkers in Crohn’s disease risk assessment and treatment. Clinic Exp Gastroenterol 2011; 4:127–140. Adler J, Rangwalla SC, Dwamena BA, Higgins PD. The prognostic power of the NOD2 genotype for complicated Crohn’s disease: a meta-analysis. Am J Gastroenterol 2011; 106:699–712. Arnott IDRMD, Landers CJPD, Nimmo EJPD, Drummond HEBS, Smith BKRBS, Targan SRMD, et al. Sero-reactivity to microbial components in Crohn’s disease is associated with disease severity and progression, but not NOD2/CARD15 genotype. Am J Gastroenterol 2004; 99:2376–2384. Dubinsky MC, Lin YC, Dutridge D, Picornell Y, Landers CJ, Farrior S, et al. Serum immune responses predict rapid disease progression among children with Crohn’s disease: immune responses predict disease progression. Am J Gastroenterol 2006; 101:360–367. Ferrante M, Henckaerts L, Joossens M, Pierik M, Joossens S, Dotan N, et al. New serological markers in inflammatory bowel disease are associated with complicated disease behaviour. Gut 2007; 56:1394–1403. Mow WS, Vasiliauskas EA, Lin YC, Fleshner PR, Papadakis KA, Taylor KD, et al. Association of antibody responses to microbial antigens and complications of small bowel Crohn’s disease. Gastroenterology 2004; 126:414–424. Papp M, Altorjay I, Norman GL, Shums Z, Palatka K, Vitalis Z, et al. Seroreactivity to microbial components in Crohn’s disease is associated with ileal involvement, noninflammatory disease behavior and NOD2/ CARD15 genotype, but not with risk for surgery in a Hungarian cohort of IBD patients. Inflamm Bowel Dis 2007; 13:984–992. Zholudev A, Young WB, Leichtner AM, Bousvaros A. Sensitivity and specificity of p-ANCA, ASCA, and anti-OmpC serologic testing for inflammatory bowel disease in children and young adults. Gastroenterology 2002; 122:A12–A12. Dotan N, Altstock RT, Schwarz M, Dukler A. Anti-glycan antibodies as biomarkers for diagnosis and prognosis. Lupus 2006; 15:442–450. Main J, McKenzie H, Yeaman GR, Kerr MA, Robson D, Pennington CR, et al. Antibody to Saccharomyces cerevisiae (bakers’ yeast) in Crohn’s disease. BMJ 1988; 297:1105–1106. Barnich N, Darfeuille-Michaud A. Adherent-invasive Escherichia coli and Crohn’s disease. Curr Opin Gastroenterol 2007; 23:16–20. Wei B, Huang T, Dalwadi H, Sutton CL, Bruckner D, Braun J. Pseudomonas fluorescens encodes the Crohn’s disease-associated I2 sequence and T-cell superantigen. Infect Immun 2002; 70:6567–6575. Lodes MJ, Cong Y, Elson CO, Mohamath R, Landers CJ, Targan SR, et al. Bacterial flagellin is a dominant antigen in Crohn disease. J Clin Invest 2004; 113:1296–1306. Prideaux L, De Cruz P, Ng SC, Kamm MA. Serological antibodies in inflammatory bowel disease: a systematic review. Inflamm Bowel Dis 2012; 18:1340–1355. Dubinsky MC, Kugathasan S, Mei L, Picornell Y, Nebel J, Wrobel I, et al. Increased immune reactivity predicts aggressive complicating Crohn’s disease in children. Clin Gastroenterol Hepatol 2008; 6:1105–1111. Elkadri AA, Stempak JM, Walters TD, Lal S, Griffiths AM, Steinhart AH, et al. Serum antibodies associated with complex inflammatory bowel disease. Inflamm Bowel Dis 2013; 19:1499–1505.
29 30
31
32
33
34
35 36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
Lewis JD. The utility of biomarkers in the diagnosis and therapy of inflammatory bowel disease. Gastroenterology 2011; 140:1817–1826.e2. Zhang Z, Li C, Zhao X, Lv C, He Q, Lei S, et al. Anti-Saccharomyces cerevisiae antibodies associate with phenotypes and higher risk for surgery in Crohn’s disease: a meta-analysis. Dig Dis Sci 2012; 57:2944–2954. Kaul A, Hutfless S, Liu L, Bayless TM, Marohn MR, Li X. Serum anti-glycan antibody biomarkers for inflammatory bowel disease diagnosis and progression: a systematic review and meta-analysis. Inflamm Bowel Dis 2012; 18:1872–1884. Whiting PF, Weswood ME, Rutjes AW, Reitsma JB, Bossuyt PN, Kleijnen J. Evaluation of QUADAS, a tool for the quality assessment of diagnostic accuracy studies. BMC Med Res Methodol 2006; 6:9. Zamora J, Abraira V, Muriel A, Khan K, Coomarasamy A. Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol 2006; 6:31. Deville WL, Buntinx F, Bouter LM, Montori VM, de Vet HC, van der Windt DA, et al. Conducting systematic reviews of diagnostic studies: didactic guidelines. BMC Med Res Methodol 2002; 2:9. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7:177–188. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005; 58:882–893. O’Donnell S, O’Sullivan M, O’Morain CA, Ryan BM. The clinical significance of antimicrobial serologic responses within an Irish Crohn’s disease population. Eur J Gastroenterol Hepatol 2013; 25:1464–1469. Papadakis KA, Yang H, Ippoliti A, Mei L, Elson CO, Hershberg RM, et al. Anti-flagellin (CBir1) phenotypic and genetic Crohn’s disease associations. Inflamm Bowel Dis 2007; 13:524–530. Ryan JD, Silverberg MS, Xu W, Graff LA, Targownik LE, Walker JR, et al. Predicting complicated Crohn’s disease and surgery: phenotypes, genetics, serology and psychological characteristics of a population-based cohort. Aliment Pharmacol Ther 2013; 38:274–283. Suzuki H, Fukuda Y, Koizuka H, Tomita T, Hori K, Suzuki M, et al. Dietary antigens in Crohn’s disease: antibodies to porcine pancreatic amylase. Am J Gastroenterol 2008; 103:656–664. Michielan A, Basso D, Martinato M, Pathak S, Banerjee A, Oliva L, et al. Increased antibody response to microbial antigens in patients with Crohn’s disease and their unaffected first-degree relatives. Dig Liver Dis 2013; 45:894–898. Papp M, Norman GL, Vitalis Z, Tornai I, Altorjay I, Foldi I, et al. Presence of anti-microbial antibodies in liver cirrhosis – a tell-tale sign of compromised immunity? PloS One 2010; 5:e12957. Kamat A, Ancuta P, Blumberg RS, Gabuzda D. Serological markers for inflammatory bowel disease in AIDS patients with evidence of microbial translocation. PLoS One 2010; 5:e15533. Lichtenstein GR, Targan SR, Dubinsky MC, Rotter JI, Barken DM, Princen F, et al. Combination of genetic and quantitative serological immune markers are associated with complicated Crohn’s disease behavior. Inflamm Bowel Dis 2011; 17:2488–2496. Landers CJ, Cohavy O, Misra R, Yang H, Lin YC, Braun J, et al. Selected loss of tolerance evidenced by Crohn’s disease-associated immune responses to auto- and microbial antigens. Gastroenterology 2002; 123:689–699. Muller S, Styner M, Seibold-Schmid B, Flogerzi B, Mahler M, Konrad A, et al. Anti-Saccharomyces cerevisiae antibody titers are stable over time in Crohn’s patients and are not inducible in murine models of colitis. World J Gastroenterol 2005; 11:6988–6994. Eser A, Papay P, Primas C, Pernicka E, Harrer M, Dejaco C, et al. The impact of intestinal resection on serum levels of anti-Saccharomyces cerevisiae antibodies (ASCA) in patients with Crohn’s disease. Aliment Pharmacol Ther 2012; 35:292–299. Targan SR, Landers CJ, Yang H, Lodes MJ, Cong Y, Papadakis KA, et al. Antibodies to CBir1 flagellin define a unique response that is associated independently with complicated Crohn’s disease. Gastroenterology 2005; 128:2020–2028. Amre DK, Lu SE, Costea F, Seidman EG. Utility of serological markers in predicting the early occurrence of complications and surgery in pediatric Crohn’s disease patients. Am J Gastroenterol 2006; 101:645–652. Papp M, Altorjay I, Dotan N, Palatka K, Foldi I, Tumpek J, et al. New serological markers for inflammatory bowel disease are associated with earlier age at onset, complicated disease behavior, risk for surgery, and NOD2/CARD15 genotype in a Hungarian IBD Cohort. Am J Gastroenterol 2008; 103:665–681. Arnott ID, McNeill G, Satsangi J. An analysis of factors influencing shortterm and sustained response to infliximab treatment for Crohn’s disease. Aliment Pharmacol Ther 2003; 17:1451–1457.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
742 European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
52 Esters N, Vermeire S, Joossens S, Noman M, Louis E, Belaiche J, et al. Serological markers for prediction of response to anti-tumor necrosis factor treatment in Crohn’s disease. Am J Gastroenterol 2002; 97:1458–1462. 53 Papp M, Foldi I, Altorjay I, Palyu E, Udvardy M, Tumpek J, et al. Anti-microbial antibodies in celiac disease: trick or treat? World J Gastroenterol 2009; 15:3891–3900. 54 van Schaik FD, Oldenburg B, Hart AR, Siersema PD, Lindgren S, Grip O, et al. Serological markers predict inflammatory bowel disease years before the diagnosis. Gut 2013; 62:683–688. 55 Harrer M, Reinisch W, Dejaco C, Kratzer V, Gmeiner M, Miehsler W, et al. Do high serum levels of anti-Saccharomyces cerevisiae antibodies result from a leakiness of the gut barrier in Crohn’s disease? Eur J Gastroenterol Hepatol 2003; 15:1281–1285.
56
Vermeire S, Peeters M, Vlietinck R, Joossens S, Den Hond E, Bulteel V, et al. Anti-Saccharomyces cerevisiae antibodies (ASCA), phenotypes of IBD, and intestinal permeability: a study in IBD families. Inflamm Bowel Dis 2001; 7:8–15. 57 Blumberg RS, Saubermann LJ, Strober W. Animal models of mucosal inflammation and their relation to human inflammatory bowel disease. Curr Opin Immunol 1999; 11:648–656. 58 De Winter H, Cheroutre H, Kronenberg M. Mucosal immunity and inflammation. II. The yin and yang of T cells in intestinal inflammation: pathogenic and protective roles in a mouse colitis model. Am J Physiol 1999; 276:G1317–G1321. 59 Strober W, Fuss IJ, Blumberg RS. The immunology of mucosal models of inflammation. Annu Rev Immunol 2002; 20:495–549.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Serum antibodies to microbial antigens for Crohn’s disease progression: a meta-analysis Yao Xionga,b, Gou-Zhen Wanga,b, Jie-Qiong Zhoua,b, Bing-Qing Xiaa,b, Xin-Ying Wanga,b and Bo Jianga,b Objectives This meta-analysis evaluated the stratification powers of four well-studied serum antibodies to microbial antigens [ASCA (anti-Saccharomyces cerevisiae), anti-OmpC (anti-outer-membrane protein C), anti-I2 (anti-Pseudomonas fluorescens-associated sequence I2), and anti-CBir1 (anti-bacterial flagellin)] in characterizing progression of Crohn’s disease (CD). Methods Pooled sensitivity, specificity, and diagnostic odds ratio (DOR) with 95% confidence intervals (CI) for individual antibodies and antibody combination were used to evaluate and compare their stratification powers for CD-related complications and the need for surgery. Results Eleven studies were included in this metaanalysis. In terms of the outcomes for CD complication and surgery, ASCA had the highest sensitivities at 0.66 (CI 0.63–0.69) for complications and 0.66 (CI 0.63–0.68) for surgery, whereas anti-OmpC had the highest specificities at 0.83 (CI 0.80–0.85) for complications and 0.81 (CI 0.79–0.83) for surgery. Anti-OmpC had the highest DORs at 2.61 (CI 2.16–3.15) for complications and 2.93 (CI 2.48–3.47) for surgery, and a combination of at least
Introduction Crohn’s disease (CD) is a heterogeneous disorder manifesting with diverse clinical presentations. Although some patients will experience a benign clinical course, B80% can expect to develop complications including stricturing and penetrating disease behaviors and almost 50% will require surgery over a 10-year period [1,2]. The traditional treatment paradigm for CD includes a ‘stepup’ approach (e.g. topical steroids, ‘stepping up’ to systemic steroid use, followed by the use of immunosuppressants and biologicals if necessary). Although this approach may be effective in the near term, it may not prevent overall disease progression [3,4]. Recent evidence suggests that early effective therapies, such as the use of biologics and immunomodulators (referred to as the ‘top-down’ approach), might minimize the progression of disease complications [5,6]. Biologics, however, are expensive and are associated with rare but serious events including opportunistic infections and malignancies [7]. Therefore, it is important for physicians to identify factors that can stratify patients into at-risk populations for disease progression who would most benefit from an early and effective intervention. c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0954-691X
two antibodies presented pooled DORs at 2.93 (CI 2.42–3.56) for complications and 3.39 (CI 2.73–4.20) for surgery, superior to any single antibody. Conclusion Anti-OmpC had the highest stratification power among the four antibodies screened for the risk of both complications and surgery in CD patients, and the power became stronger when antibodies were assessed c in combination. Eur J Gastroenterol Hepatol 26:733–742 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. European Journal of Gastroenterology & Hepatology 2014, 26:733–742 Keywords: complication, Crohn’s disease, meta-analysis, microbial antigens, serum antibodies, surgery a Department of Gastroenterology, Nanfang Hospital, Southern Medical University and bGuangdong Provincial Key Laboratory of Gastroenterology, Guangzhou, China
Correspondence to Xin-Ying Wang and Bo Jiang, PhD, Department of Gastroenterology, Nanfang Hospital, Southern Medical University, Guangzhou 510515, China Tel: + 86 206 164 1540; fax: + 86 206 278 7385; e-mails: [email protected] and [email protected] Received 11 January 2014 Accepted 14 March 2014
Although current prognostic markers of CD rely on clinical risk factors such as the need for cortocosteroids, disease onset before the age of 40 years, smoking, the presence of perianal lesions at diagnosis, small bowel localization, and stricturing disease [8,9], these factors are neither sufficient nor accurate [10–13]. Identification of genetic markers has also been proposed for establishing the prognosis of CD [12]. However, the predictive power for the disease course of the NOD2/CARD15 genotype (the first and most promising genes associated with CD risk stratification) was weak, with poor area under the receiver operating characteristic curve, as shown recently by Adler et al. [14]. With emerging data showing that serologic antibodies to microbial antigens are associated with more aggressive CD phenotypes as well as a higher risk for surgery [15–20], these antibodies have become attractive prognostic markers. More than 10 kinds of antimicrobial serologic antibodies have been identified to be relevant to CD, including ASCA (antibodies directed against the oligomannan component of Saccharomyces cerevisiae, one of the members of anti-glycan antibodies), anti-OmpC (antibodies to the outer-membrane porin C of Escherichia DOI: 10.1097/MEG.0000000000000102
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
734
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
coli), anti-I2 (antibodies to the Pseudomonas fluorescensassociated sequence I2), anti-CBir1 (antibodies against bacterial flagellin), and other newly discovered antiglycan antibodies (AMCA, ALCA, ACCA, Anti-L, and Anti-C), and anti-flagellin antibodies (antibodies to A4Fla2 and Fla-X). Anti-glycan antibodies target cell wall carbohydrate epitopes found in microbiota such as yeasts and bacteria [21]. The most prominent member of this group of antibodies is ASCA, which targets the 200 kDa phosphopeptidomannan, a cell wall mannan of the common baker’s or brewer’s yeast S. cerveisiae [22]. Anti-OmpC is an antibody to an outer-membrane porin isolated from E. coli, and OmpC has been shown to help adherent-invasive E. coli thrive in the GI tract [23]. AntiI2 is an antibody to a bacterial DNA fragment that is a homolog of the bacterial transcription factor family derived from P. fluorescens [24]. Anti-CBir1 is an antibody to flagellin that is the primary structural component of bacterial flagella, and flagellin CBir1 has been shown to induce colitis in immunodeficient mice [25]. All of these antibodies have been shown to possess a certain degree of association with complicated disease behaviors or surgery in CD patients, but the results and conclusions of the existing studies [15–20] differ for each antibody. Therefore, before serological markers can aid in determining disease prognosis, their respective stratification powers and their specificities and sensitivities have to be defined clearly for their abilities to determine disease progression. This study describes a meta-analysis of serum antibodies to microbial antigens as stratification indicators of the progression and severity of CD. Here, we focused on four antibodies: ASCA, anti-OmpC, anti-I2, and anti-CBir1 because they have been the most well-characterized antibodies over the last 10 years [26] and represent antibody types specific for different microbial targets. As many studies [15,27,28] have found an association between the number of antimicrobial antibodies and CD phenotypes and disease progression (the so-called ‘serology dosage’ effect), we also assessed these antibodies in combination. To our knowledge, this is the first meta-analysis to evaluate the stratification powers of the four well-studied serum antibodies in determining complications and surgery in CD patients, although some comments and reviews [12,26,29] have been published on this subject and two meta-analyses have been carried out on the association between ASCA [30] and antiglycan antibodies [31] with progression of CD.
Methods
combination: Crohn’s, Crohn, inflammatory bowel disease, ASCA, anti-Saccharomyces cerevisiae antibodies, I2, anti-I2, antibodies to the Crohn’s disease-related bacterial sequences, Pseudomonas fluorescens-related protein, antiEscherichia coli outer membrane porin C, anti-OmpC, anti-outer-membrane protein C, Omp, anti-flagellin, antiCBir1, and antibodies to flagellin. The reference lists of eligible studies and review articles were also checked to identify other publications relevant to this topic. Inclusion/exclusion criteria
Studies were included in the present analysis when the following criteria were fulfilled: (a) comparison in humans of at least two of the four antibodies [ASCA (or gASCA), anti-OmpC, anti-I2, anti-CBir1]; (b) analysis and comparison of associations between these markers with complications or surgery (we defined complications as stricturing and/or penetrating or non-inflammatory disease and surgery as CD-related surgery or small-bowel resections or abdominal surgery); (c) if raw numbers of each genotype–phenotype combination could be found or clearly calculated from the published data; (d) in cases of potential or suspected overlapping patient populations (on the basis of either location or authorship), the latest publication or the study with the largest numbers of patients was included in the meta-analysis; and (e) written in English. We excluded reviews, editorials, and meeting abstracts. Review processes and data extraction
Titles of all candidate articles were read, and for those that were not excluded solely on the basis of titles, the abstracts were reviewed. We then retrieved the full texts of published papers that were not excluded on the basis of the titles and abstracts alone. Each article was read to confirm whether it fulfilled the inclusion and exclusion criteria. Two reviewers assessed the studies selected on the basis of the authors, journal, year of publication, country, population source, total sample size, patient age, and serum antibodies measured. Study quality was assessed using the QUADAS (Quality Assessment of Studies of Diagnostic Accuracy Included in Systematic Reviews) checklist (maximum score, 14) [32]. In cases of discordance, a consensus was reached through discussion with the senior author. True-positive (TP), false-positive (FP), falsenegative (FN), and true-negative (TN) numbers were calculated for each study. Some included data were recomputed from original article descriptions that potentially resulted in slightly different values. Corresponding authors were contacted for missing data.
Database searches
Electronic databases including PubMed, Embase, Ovid, and Web of Science were searched to identify studies on the association between serum antimicrobial antibodies and CD complications or the need for surgery up to December 2013. This search strategy used a combination of the following MeSH headings and keywords alone or in
Statistical analysis
The primary outcome of interest for this analysis was the stratification powers of serum antibodies for CD progression, including complications and the need for surgery. Standard methods recommended for meta-analyses of diagnostic tests were used. Statistical analyses were
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Antimicrobial antibodies for CD progression Xiong et al. 735
carried out using Stata 12.0 (Stata Corporation, College Station, Texas, USA) and Meta-DiSc 1.4 [33]. P-values less than 0.05 were considered statistically significant. Pooled sensitivities, specificities, and diagnostic odds ratios (DORs) with 95% confidence interval (CI) were calculated. Here, the sensitivity means the odds of a positive antibody state in patients with an aggressive disease course (complications or surgery); the specificity means the odds of a negative antibody state in patients without an aggressive disease course (no complications or surgery); and the DOR is defined as the odds of a positive antibody state in patients with an aggressive disease course (complications or surgery) compared with the odds of a positive antibody state in patients without an aggressive disease course (no complications or surgery). A higher DOR indicated a higher association of serum antibodies and disease progression [34]. I2 was used to assess statistical heterogeneity, with values at least 50% indicating significant heterogeneity. The P-value of the Cochran’s Q-test was calculated to detect statistically significant heterogeneity across studies. We used the fixed-effects model by Mantel–Haenszel to calculate the summary DOR and its 95% CI among studies with homogeneous results, whereas the random-effects model by DerSimonian and Laird [35] was used when significant heterogeneity was found. Forest plots were derived to show the results of individual studies and pooled analyses. Deeks’ funnel plots [36] were used to detect any publication bias, for which a P-value less than 0.1 was suggestive of significant bias.
Results Study characteristics
The initial search of the databases identified 469 articles and among these, 11 articles [15–18,27,28,37–41] fulfilled the criteria established in this meta-analysis. The detailed steps of our literature search and reasons for exclusions are shown in Fig. 1. One study by Papp et al. [19] was excluded from the meta-analysis because it included anti-Omp, which is against multiple bacterial proteins derived from two species of intestinal bacteria according to the subsequent publication by the same author [42] and thus different from anti-OmpC in this study. All the 11 studies were retrospective and had been published between 2004 and 2013. Four studies were carried out in Europe, four in the USA, two in Canada, and one in Japan. Two studies included pediatric patients [16, 27] and the others focused primarily on adult patients. The clinical characteristics and QUADAS scores of these studies are summarized in Table 1. Complicated behavior
Three to six studies with adequate quantitative details were used for each analysis of the four individual antibodies and their combination in terms of the outcomes of the complicated behavior (Table 2 and Fig. 2). For individual antibodies, ASCA and antiCBir1 had higher sensitivities [0.66 (95% CI 0.63–0.69) and 0.60 (95% CI 0.56–0.65), respectively], but lower specificities [0.57 (95% CI 0.54–0.60) and 0.55 (95% CI 0.52–0.58), respectively], whereas anti-OmpC and anti-I2 had higher specificities [0.83 (95% CI 0.80–0.85) and 0.81
Fig. 1
469 articles identified in database search after duplicates removed
220 full-text articles assessed for eligibility
Records excluded (249): Case report (6) Reviews, commentary, or editorial (45) In vitro and animal studies (9) Not about CD (189)
Records excluded (209): Review, commentary, or case report (65) In vitro and animal studies (10) Not about IBD or the four antibodies (50) Different outcome (65) Did not compare two antibodies (11) No individual values of markers given (4)
11 articles included in this meta-analysis
Overlap of patients (1) Not in English (3)
Flow chart describing the selection criteria of the studies included in the meta-analysis. CD, Crohn’s disease; IBD, inflammatory bowel disease.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
736
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
Table 1
Patient characteristics of the 11 studies included in the meta-analyses
References, country
Mean or median age/mean or median age at CD patients diagnosis (years) no.
Source, population
Arnott et al., UK [15] Mow et al., USA [18] Dubinsky et al., USA [16]
Western General Hospital, Edinburgh Cedars-Sinai Medical Center Western Regional Pediatric IBD Research Alliance Ferrante et al., Belgium [17] The University Hospital in Leuven, Belgium Suzuki et al., Japan [40] The hospital of Hyogo College of Medicine Papadakis et al., USA [38] Cedars-Sinai Medical Center Dubinsky et al., USA [27]
Western Regional Pediatric IBD Research Alliance, the Pediatric IBD Collaborative Research Group, the Wisconsin Pediatric IBD Alliance O’Donnell et al., Ireland [37] A specialist IBD clinic at a university hospital Elkadri et al., Canada [28] Mount Sinai Hospital and The Hospital for Sick Children Ryan et al., Canada [39] Manitoba IBD Cohort Michielan et al., Italy [41]
The gastroenterology outpatient unit at the University Hospital of Padua
39/27 –/–(23)a 13/12
142 303 196
35/22 34.5/24.7 –/–
738 104 428
–/12
796
38/26 32.8/22.1 38.5/–
179 391 127
36.5/–
60
Serum antibodies measured
QUADAS quality score
ASCA, anti-OmpC, anti-I2 ASCA, anti-OmpC, anti-I2 ASCA, anti-OmpC, anti-Cbir1, anti-I2 ASCA, anti-OmpC ASCA, anti-I2 ASCA, anti-OmpC, anti-Cbir1, anti-I2 ASCA, anti-OmpC, anti-Cbir1
11 10 11
ASCA, anti-OmpC, anti-Cbir1 ASCA, anti-OmpC, anti-Cbir1 ASCA, anti-OmpC, anti-Cbir1, anti-I2 ASCA, anti-OmpC
11 11 10
9 10 11 9
11
anti-CBir1, antibodies against bacterial flagellin; anti-I2, antibodies to the Pseudomonas fluorescens-associated sequence I2; anti-OmpC, antibodies to the outer-membrane porin C of Escherichia coli; ASCA, antibodies directed against the oligomannan component of Saccharomyces cerevisiae; CD, Crohn’s disease; IBD, inflammatory bowel disease; QUADAS, Quality Assessment of Studies of Diagnostic Accuracy Included in Systematic Reviews. a Median age of onset.
(95% CI 0.75–0.86), respectively], but lower sensitivities [0.38 (95% CI 0.36–0.41) and 0.39 (95% CI 0.32–0.46), respectively]. When looking at the DORs, anti-OmpC had the highest value of 2.61 (95% CI 2.16–3.15) and anti-CBir1 had the lowest value of 1.94 (95% CI 1.58–2.38); ASCA and anti-I2 had intermediate values of 2.39 (95% CI 2.01–2.85) and 2.27 (95% CI 1.43–3.60), respectively. Analysis of antibody combination identified that more than one antibody resulted in a pooled DOR of 2.93 (95% CI 2.42–3.56) that was higher than any individual antibody (2.93 vs. 1.94–2.61), indicating a numerical but not statistically significant difference of having a greater chance of developing a complication.
Surgery
Three to eight studies that fulfilled the inclusion criteria were used for the analyses of individual antibodies and their combination in terms of the outcomes of surgery (Table 2 and Fig. 3). Similar to the results of the complicated behavior, the analysis of the need for surgery for individual antibodies showed that ASCA and antiCBir1 had higher sensitivities [0.66 (95% CI 0.63–0.68) and 0.61 (95% CI 0.57–0.65), respectively], but lower specificities [0.58 (95% CI 0.55–0.60) and 0.52 (95% CI 0.49–0.55), respectively], whereas anti-OmpC and anti-I2 had higher specificities [0.81 (95% CI 0.79–0.83) and 0.63 (95% CI 0.57–0.69), respectively], but lower sensitivities [0.43 (95% CI 0.41–0.46) and 0.59 (95% CI 0.53–0.65), respectively]. The DOR for surgery was the highest for anti-OmpC, with a value of 2.93 (95% CI 2.48–3.47), and the lowest for anti-CBir1, with a value of 1.83 (95% CI 1.52–2.20). The pooled DOR for surgery when a combination of at least two antibodies was used
was 3.39 (95% CI 2.73–4.20), which was also superior to any single antibody examined (3.39 vs. 1.83–2.93). Publication bias
Deeks’ funnel plots were used to evaluate publication bias of the articles, with the results of P-values shown in Table 2. Except for the analysis of ASCA, the tests showed a statistically nonsignificant value (P > 0.1), indicating that there was generally no potential publication bias.
Discussion Reliable prognostic markers are needed to identify CD patients who might benefit from the timely introduction of immunomodulators or biologicals. Although antibody responses toward microbial antigens have questionable clinical benefit as diagnostic tools because of their low sensitivities, they can be used to characterize specific and adverse outcomes associated with CD, making them promising prognostic markers. Before recommending the routine use of serological markers for predicting disease course, diagnostic studies reporting sensitivities and specificities of serological markers in terms of their stratification powers for CD progression are warranted. If found to be useful and their predictive powers are further proved in prospective studies, these markers could potentially be added to clinical indicators used to predict the risk for developing adverse outcomes associated with CD. Recently, two meta-analyses investigated the association between several serum antibodies against microbial antigens and CD progression. One meta-analysis carried out by Zhang et al. [30] focused on ASCA and suggested that a positive ASCA status was a risk factor for CD-related complications and surgery. The other
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
meta-analysis carried out by Kaul et al. [31] investigated a group of anti-glycan antibody biomarkers (ASCA, AMCA, ALCA, ACCA, anti-L, and anti-C) and showed that these anti-glycan antibodies had pooled DORs of around 2, with no prominent marker associated with disease progression. The results of the latter meta-analysis also showed that ACCA had the highest association with complications, whereas both ASCA and ACCA associated equally with the need for surgery and that the combination of more than one antibody had a higher association with disease course. These two meta-analyses only focused on one antibody or antibodies belonging to the group of antiglycan antibodies. Different from them, this metaanalysis examined four antibodies (ASCA, anti-OmpC, anti-I2, and anti-CBir1) that have recently been well studied in terms of their relationship with CD progression and are specific for different microbial targets. Our analyses for individual antibodies (Table 2 and Figs 2 and 3) indicated that ASCA and anti-CBir1 had higher sensitivities, but relatively lower specificities compared with anti-OmpC, which had the highest specificity, but the lowest sensitivity for complications and surgery. These results may reflect some difference in the prevalences of these antibodies in CD patients. The meta-analysis by Kamat et al. [43] showed that the prevalences of ASCA IgA/IgG (45%) and anti-CBir1 (55.2%) in CD patients were relatively high, whereas the prevalence of antiOmpC (29.4%) was relatively low. In terms of the DORs of individual antibodies in this study, anti-OmpC was shown to have the highest stratification power, which was superior to ASCA (the traditional and best-studied serum antibody). This suggests that future studies on serum antibodies should pay more attention toward the detection of anti-OmpC.
Sensitivity (95% CI) I2%, P-value of Q-test Specificity (95% CI) I2%, P-value of Q-test DOR (95% CI) I2%, P-value of Q-test Publication bias P-value
CI, confidence interval; DOR, diagnostic odds ratio. Other abbreviations as in Table 1. a Defined as either IgA or IgG ASCA positive.
0.59 (0.53–0.65) 95.7, 0.000 0.63 (0.57–0.69) 91.2, 0.000 2.71 (1.95–93.77) 0.0, 0.533 0.181
Arnott and colleagues [15,17,18,28,41] 0.47 (0.43–0.50) 96.4, 0.000 0.77 (0.74–0.80) 95.0, 0.000 3.39 (2.73–4.20) 0.0, 0.794 0.101 Dubinsky and colleagues [27,28,37,38,39] 0.61 (0.57–0.65) 74.9, 0.003 0.52 (0.49–0.55) 47.5, 0.107 1.83 (1.52–2.20) 12.0, 0.337 0.084 Arnott and colleagues [15,18,39]
Arnott and colleagues [15,17,18,27,28,37,39,41] 0.43 (0.41–0.46) 91.7, 0.000 0.81 (0.79–0.83) 91.1, 0.000 2.93 (2.48–3.47) 47.3, 0.065 0.669
Sensitivity (95% CI) I2%, P-value of Q-test Specificity (95% CI) I2%, P-value of Q-test DOR (95% CI) I2%, P-value of Q-test Publication bias P-value Surgery Studies included
Arnott and colleagues [15,17,18,27,28,41] 0.66 (0.63–0.68) 19.0, 0.290 0.58 (0.55–0.60) 86.0, 0.000 2.82 (2.03–3.91) 63.9, 0.017 0.028
0.39 (0.32–0.46) 85.7, 0.001 0.81 (0.75–0.86) 0.0, 0.765 2.27 (1.43–3.60) 4.5, 0.351 0.762
Arnott and colleagues [15,17,27,28,41] 0.42 (0.39–0.45) 96.6, 0.000 0.75 (0.72–0.78) 92.9, 0.000 2.93 (2.42–3.56) 0.0, 0.475 0.646 Dubinsky and colleagues [27,28,37,39] 0.60 (0.56–0.65) 71.9, 0.014 0.55 (0.52–0.58) 44.2, 0.146 1.94 (1.58–2.38) 5.9, 0.364 0.683 Ferrante and colleagues [17,27,28,37,39,41] 0.38 (0.35–0.41) 93.8, 0.000 0.83 (0.80–0.85) 88.1, 0.000 2.61 (2.16–3.15) 0.0, 0.546 0.779 Ferrante and colleagues [17,27,28,41] 0.66 (0.63–0.69) 63.6, 0.041 0.57 (0.54–0.60) 74.8, 0.008 2.39 (2.01–2.85) 7.4, 0.3563 0.028
Anti-OmpC ASCAa
Dubinsky and colleagues [27,39,40]
Combination (Z 2 antibodies)
Complication Studies included
Table 2
Pooled results of serum antibodies to microbial antigens
Anti-I2
Anti-CBir1
Antimicrobial antibodies for CD progression Xiong et al. 737
The present results also suggested that individual antibodies were useful stratification markers but their respective values indicated by sensitivities, specificities, and DORs were not as high as expected. The highest sensitivity, specificity, and DOR of individual antibodies for complications were 0.66, 0.83, and 2.61 (Table 2), respectively, with similar outcomes in terms of surgery. However, these findings of individual antibodies did not maximize the role that serum antibodies played in indicating the complications and surgery. As shown in Table 2 and Figs 2 and 3, the combination of at least two antibodies had an elevated DOR of 3.05 for complications and 3.16 for surgery, suggesting that analysis of the presence of antibodies in combination had a better stratification power. It is regrettable that we cannot find a way to meta-analyze the association between the number of antibodies with the disease course. However, it is important to note that an increasing number of positive antibodies [15,27,28] appeared to be associated with a higher risk for CD-related complications and surgery. Furthermore, elevated immune responses toward individual [17–19,44] or multiple [15,18,27,37] microbial
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
738
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
Fig. 2
Diagnostic OR (95% CI) (a)
2.34 (1.73 − 3.18) Ferrante et al. [17] Dubinsky et al. [27] 2.23 (1.62 − 3.08) Elkadri et al. [28] 2.43 (1.59 − 3.69) Michielan et al. [41] 6.57 (2.11 − 20.48) Fixed-effects model Pooled diagnostic odds ratio = 2.39 (2.01 − 2.85) Cochran-Q = 3.24; d.f. = 3 (P = 0.3563) Inconsistency (I 2) = 7.4%
ASCA
0.01
1
100.0
Anti-OmpC
1
0.01 (c)
Diagnostic OR (95% CI)
2.11 (1.49 − 2.97) Ferrante et al. [17] 2.73 (1.86 − 4.01) Dubinsky et al. [27] 1.16 (0.07 − 18.88) Ryan et al. [39] 3.15 (2.06 − 4.80) Elkadri et al. [28] 2.11 (0.70 − 6.35) Michielan et al. [41] 3.90 (1.94 − 7.83) O’Donnell et al. [37] Fixed-effects model Pooled diagnostic odds ratio = 2.61 (2.16 − 3.15) Cochran-Q = 4.03; d.f. = 5 (P = 0.5456) Inconsistency (I 2) = 0.0 %
(b)
100.0 Dubinsky et al. [16] Suzuki et al. [40] Ryan et al. [39]
Anti-I2
Diagnostic OR (95% CI) 2.68 (1.37 − 5.25) 3.67 (0.96 − 14.03) 1.32 (0.54 − 3.27)
Fixed-effects model Pooled diagnostic odds ratio = 2.27 (1.43 − 3.60) Cochran-Q = 2.09; d.f. = 2 (P = 0.3509) Inconsistency (I 2) = 4.5 % 1
0.1 (d)
10.0 Dubinsky et al. [27] Ryan et al. [39] Elkadri et al. [28] O’Donnell et al. [37]
Anti-CBir1
0.1
1
(e)
Combination (≥ 2 antibodies)
Diagnostic OR (95% CI) 1.77 (1.28 − 2.47) 1.26 (0.61 − 2.58) 2.50 (1.66 − 3.76) 2.05 (1.12 − 3.76)
Fixed-effects model Pooled diagnostic odds ratio = 1.94 (1.58 − 2.38) Cochran-Q = 3.19; d.f. = 3 (P = 0.3637) Inconsistency (I 2) = 5.9 % 10.0 Diagnostic OR (95% CI) Arnott et al. [15] 2.38 (1.14 − 4.99) Ferrante et al. [17] 3.23 (2.06 − 5.05) Dubinsky et al. [27] 2.43 (1.74 − 3.39) 3.89 (2.56 − 5.93) Elkadri et al. [28] 2.51 (0.76 − 8.37) Michielan et al. [41] Fixed-effects model Pooled diagnostic odds ratio = 2.93 (2.42 − 3.56) Cochran-Q = 3.52; d.f. = 4 (P = 0.4748) Inconsistency (I 2) = 0.0 %
0.01
1
100.0
Forest plots of pooled diagnostic odds ratios (DORs) for complication. (a) ASCA, (b) anti-OmpC, (c) anti-I2, (d) anti-CBir1, and (e) combination (Z 2 antibodies). Plots show pooled DOR probabilities of included studies with corresponding 95% confidence interval (CI), with circles indicating the area proportional to the study weight in the meta-analysis.
antigens were also shown to be associated with higher frequencies of CD complications and surgery. For example, a study by Mow et al. [18] showed that only 18% of patients with the lowest-level response toward three tested microbial antigens had small bowel surgery, whereas 90% of patients with the highest-level response toward these microbial antigens had small bowel surgery. On the basis of these discussions, the combined presence and levels of a
number of serum antibodies could be the strongest stratification markers. It should be noted that the studies included in this analysis were retrospective and were mostly cross-sectional, with samples taken at various random points during the disease course. Thus, the detection of the antibodies described in this report was more an indicator for the association of, rather
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Antimicrobial antibodies for CD progression Xiong et al. 739
Fig. 3
Diagnostic OR (95% CI) (a)
Mow et al. [18] Arnott et al. [15] Ferrante et al. [17] Dubinsky et al. [27] Elkadri et al. [28] Michielan et al. [41]
ASCA
1
0.01
100.0
Anti-OmpC
(1.97 – 5.04) (3.45 – 17.48) (1.59 – 2.89) (1.54 – 3.15) (1.33 – 3.06) (2.16 – 22.10)
Random effects model Pooled diagnostic odds ratio = 2.82 (2.03 – 3.91) Cochran-Q = 13.87; d.f. = 5 (P = 0.0165) Inconsistency (I 2) = 63.9 % τ2 = 0.0968 Mow et al. [18] Arnott et al. [15] Ferrante et al. [17] Dubinsky et al. [27] Ryan et al. [39] Elkadri et al. [28] Michielan et al. [41] O’Donnell et al. [37]
(b)
3.15 7.76 2.15 2.20 2.01 6.91
Diagnostic OR (95% CI) 2.01 (1.27 – 3.19) 1.35 (0.68 – 2.70) 2.82 (2.01 – 3.95) 4.08 (2.74 – 6.06) 5.24 (0.25 – 111.44) 4.36 (2.79 – 6.83) 2.67 (0.90 – 7.94) 2.95 (1.47 – 5.90)
Fixed effects model Pooled diagnostic odds ratio = 2.93 (2.48 – 3.47) Cochran-Q = 13.29; d.f. = 7 (P = 0.0654) Inconsistency (I 2) = 47.3 % 0.01
1
100.0 Mow et al. [18] Arnott et al. [15] Ryan et al. [39]
(c) Anti-I2
Diagnostic OR (95% CI) 2.76 (1.72 – 4.43) 3.37 (1.69 – 6.72) 1.75 (0.69 – 4.39)
Fixed effects model Pooled diagnostic odds ratio = 2.71 (1.95– 3.77) Cochran-Q = 1.26; d.f. = 2 (P = 0.5332) Inconsistency (I 2) = 0.0 % 1
0.1
10.0 Papadakis et al. [38] Dubinsky et al. [27] Ryan et al. [39] Elkadri et al. [28] O’Donnell et al. [37]
(d) Anti-CBir1
Diagnostic OR (95% CI) 1.73 (1.17 – 2.55)) 2.03 (1.41 – 2.93) 1.03 (0.50 – 2.10) 2.28 (1.51 – 3.44) 1.43 (0.75 – 2.74)
Fixed effects model Pooled diagnostic odds ratio = 1.83 (1.52 – 2.20) Cochran-Q = 4.55; d.f. = 4 (P = 0.3372) Inconsistency (I 2) = 12.0 % 0.1
1
10.0 Mow et al. [18] Arnott et al. [15] Ferrante et al. [17] Elkadri et al. [28] Michielan et al. [41]
(e) Combination (≥ 2 antibodies)
Diagnostic OR (95% CI) 2.89 (1.81 – 4.61) 2.85 (1.39 – 5.85) 4.16 (2.69 – 6.44) 3.39 (2.23 – 5.16) 2.67 (0.84 – 8.44)
Fixed effects model Pooled diagnostic odds ratio = 3.39 (2.73 – 4.20) Cochran-Q = 1.68; d.f. = 4 (P = 0.7939) Inconsistency (I 2) = 0.0 % 0.01
1
100.0
Forest plots of pooled diagnostic odds ratios (DORs) for surgery. (a) ASCA, (b) anti-OmpC, (c) anti-I2 (d), anti-CBir1, and (e) combination (Z 2 markers). Plots show pooled DOR probabilities of included studies with corresponding 95% confidence interval (CI), with circles indicating the area proportional to the study weight in the meta-analysis.
than a predictor of, a severe disease course. Despite this limitation, this study paved the way for opportunities to explore and validate the predictive value of these markers.
Some studies reported stable antibody responses over time [39,45,46] or after treatment or surgery [18,47,48]; hence, if a patient was seropositive early during disease, there
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
740
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
should be a concern that this patient may at some time develop a complicated disease or may require surgery. Recently, a few studies [15,27,49] have assessed the ability of antibodies as predictors of disease course with serum usually drawn at diagnosis or before the presentation of complications or surgery. Their results strongly support the use of antibodies to microbial antigens as prediction markers. However, large prospective studies are still required to evaluate their predictive values. In addition, to make these markers further useful in choosing patients appropriate for top-down therapy, prospective studies are also needed to verify that antibody-driven risk stratification alters the progression of CD. If the presence of antimicrobial antibodies indicates treatment-refractory disease with resistance to immunomodulators/biologicals, aggressive therapy would be ineffective. However, this possibility might be small as many studies [50–52] failed to show a relationship between serotype and response to infliximab. The mechanism behind the relationship between antibody responses toward microbial antigens and severe disease phenotypes has not yet been fully understood. There are now basically two hypotheses on the mechanism. One possibility is that increased intestinal permeability or chronic bacterial translocation caused by aggressive transmural inflammation leads to immune responses and antimicrobial antibody formation. Evidences supporting this hypothesis are that significant antimicrobial antibody expression was also reported in other clinical conditions with impaired mucosa such as celiac disease [53], liver cirrhosis [42], and AIDS [43], and that antibody response could be lost with mucosal healing in celiac disease patients [53]. Some researchers, however, argue that this hypothesis is less likely because research data also suggested that antibodies could be found in preclinical settings [54], and that their presence was not associated with intestinal permeability alterations [55,56] or disease activity [45]. Another possibility is that the distinct antibody response patterns may indicate unique pathophysiological mechanisms in CD progression and are thus important determinants of disease phenotype and disease progression. This is a corollary to the current theory supported by animal models that chronic intestinal inflammation is the result of an aberrant immunologic response to gut commensal bacteria, with genetic defects as predisposing factors [57–59]. Whether antibody responses are the consequences or the triggers of severe disease phenotypes is still in dispute and their role in the pathophysiology of CD should be examined further. This meta-analysis had certain limitations. First, inevitable biases were introduced by pooling different observational studies, reflected by the statistical heterogeneity present throughout the analysis. The heterogeneity observed could be because of various sources including pharmacological treatment, disease duration, age at diagnosis, ethnic origin, and different definitions of surgery. However, we could not obtain sufficient detailed information from the original studies to further stratify the analysis. Future studies should
pay more attention to these potential confounding factors. Second, the studies included in this analysis used different panels of antibodies in determining the combined prognostic power of these markers that could lead to heterogeneity in the combination analysis. As the respective studies did not specifically name the antibody combinations used, it was impossible to know which combination worked best. Third, most of the studies included in this meta-analysis [15,16,18,27,28,37–39] have carried out serum assays in Cedars-Sinai Medical Center and Prometheus Laboratories, and may thus simply reflect the experience of these few expert laboratories. The operation and availability of such assays should be made easier for local laboratories; thus, the widespread use of serum antibody markers in clinical practice could be feasible. In conclusion, ASCA had the highest sensitivity and antiOmpC had the highest specificity for both complications and the need for surgery, and overall, the presence of antiOmpC had the highest stratification values among the four serum antibodies screened. However, analyses of at least two antibodies were more effective than any individual marker alone in the context of CD progression. To be clinically useful in identifying patients who might be ideal for ‘top-down’ therapy, detection of serum antibodies to microbial antigens and their respective levels require further evaluation in prospective studies to better define their predictive powers.
Acknowledgements The authors thank Dr J.D. Ryan (University of Manitoba, Winnipeg, Canada), Dr Abdul A. Elkadri, and Dr Mark S. Silverberg (Mount Sinai Hospital, Toronto, Canada), Dr Dubinsky (Cedars-Sinai Medical Center, Los Angeles, USA), Dr Andrea Michielan (University of Padua, Padua, Italy), and Dr Severine Vermeire and Dr Marc Ferrante (University Hospital Gasthuisberg, Leuven, Belgium) for providing additional information about their studies. Comments of four anonymous reviewers have considerably helped to improve the manuscript. This study was funded by 2012B050600022 Science and Technology Planning Project of Guangdong Province. Conflicts of interest
There are no conflicts of interest.
References 1
2
3
4
Cosnes J, Cattan S, Blain A, Beaugerie L, Carbonnel F, Parc R, et al. Long-term evolution of disease behavior of Crohn’s disease. Inflamm Bowel Dis 2002; 8:244–250. Cosnes J, Nion-Larmurier I, Beaugerie L, Afchain P, Tiret E, Gendre JP. Impact of the increasing use of immunosuppressants in Crohn’s disease on the need for intestinal surgery. Gut 2005; 54:237–241. Faubion WA Jr, Loftus EV Jr, Harmsen WS, Zinsmeister AR, Sandborn WJ. The natural history of corticosteroid therapy for inflammatory bowel disease: a population-based study. Gastroenterology 2001; 121:255–260. Vermeire S, van Assche G, Rutgeerts P. Review article: altering the natural history of Crohn’s disease – evidence for and against current therapies. Aliment Pharmacol Ther 2007; 25:3–12.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Antimicrobial antibodies for CD progression Xiong et al. 741
5
6
7 8 9
10
11
12
13
14
15
16
17
18
19
20
21 22
23 24
25
26
27
28
Feagan BG, Panaccione R, Sandborn WJ, D’Haens GR, Schreiber S, Rutgeerts PJ, et al. Effects of adalimumab therapy on incidence of hospitalization and surgery in Crohn’s disease: results from the CHARM study. Gastroenterology 2008; 135:1493–1499. Peyrin-Biroulet L, Oussalah A, Williet N, Pillot C, Bresler L, Bigard MA. Impact of azathioprine and tumour necrosis factor antagonists on the need for surgery in newly diagnosed Crohn’s disease. Gut 2011; 60:930–936. Hanauer SB. Positioning biologic agents in the treatment of Crohn’s disease. Inflamm Bowel Dis 2009; 15:1570–1582. Beaugerie L, Seksik P, Nion-Larmurier I, Gendre JP, Cosnes J. Predictors of Crohn’s disease. Gastroenterology 2006; 130:650–656. Romberg-Camps MJ, Dagnelie PC, Kester AD, Hesselink-van de Kruijs MA, Cilissen M, Engels LG, et al. Influence of phenotype at diagnosis and of other potential prognostic factors on the course of inflammatory bowel disease. Am J Gastroenterol 2009; 104:371–383. Dias CC, Rodrigues PP, da Costa-Pereira A, Magro F. Clinical prognostic factors for disabling Crohn’s disease: a systematic review and meta-analysis. World J Gastroenterol 2013; 19:3866–3871. Cerqueira RM, Lago PM. Clinical factors predictive of Crohn’s disease complications and surgery. Eur J Gastroenterol Hepatol 2013; 25: 129–134. Lichtenstein GR. Emerging prognostic markers to determine Crohn’s disease natural history and improve management strategies: a review of recent literature. Gastroenterol Hepatol 2010; 6:99–107. Tamboli CP, Doman DB, Patel A. Current and future role of biomarkers in Crohn’s disease risk assessment and treatment. Clinic Exp Gastroenterol 2011; 4:127–140. Adler J, Rangwalla SC, Dwamena BA, Higgins PD. The prognostic power of the NOD2 genotype for complicated Crohn’s disease: a meta-analysis. Am J Gastroenterol 2011; 106:699–712. Arnott IDRMD, Landers CJPD, Nimmo EJPD, Drummond HEBS, Smith BKRBS, Targan SRMD, et al. Sero-reactivity to microbial components in Crohn’s disease is associated with disease severity and progression, but not NOD2/CARD15 genotype. Am J Gastroenterol 2004; 99:2376–2384. Dubinsky MC, Lin YC, Dutridge D, Picornell Y, Landers CJ, Farrior S, et al. Serum immune responses predict rapid disease progression among children with Crohn’s disease: immune responses predict disease progression. Am J Gastroenterol 2006; 101:360–367. Ferrante M, Henckaerts L, Joossens M, Pierik M, Joossens S, Dotan N, et al. New serological markers in inflammatory bowel disease are associated with complicated disease behaviour. Gut 2007; 56:1394–1403. Mow WS, Vasiliauskas EA, Lin YC, Fleshner PR, Papadakis KA, Taylor KD, et al. Association of antibody responses to microbial antigens and complications of small bowel Crohn’s disease. Gastroenterology 2004; 126:414–424. Papp M, Altorjay I, Norman GL, Shums Z, Palatka K, Vitalis Z, et al. Seroreactivity to microbial components in Crohn’s disease is associated with ileal involvement, noninflammatory disease behavior and NOD2/ CARD15 genotype, but not with risk for surgery in a Hungarian cohort of IBD patients. Inflamm Bowel Dis 2007; 13:984–992. Zholudev A, Young WB, Leichtner AM, Bousvaros A. Sensitivity and specificity of p-ANCA, ASCA, and anti-OmpC serologic testing for inflammatory bowel disease in children and young adults. Gastroenterology 2002; 122:A12–A12. Dotan N, Altstock RT, Schwarz M, Dukler A. Anti-glycan antibodies as biomarkers for diagnosis and prognosis. Lupus 2006; 15:442–450. Main J, McKenzie H, Yeaman GR, Kerr MA, Robson D, Pennington CR, et al. Antibody to Saccharomyces cerevisiae (bakers’ yeast) in Crohn’s disease. BMJ 1988; 297:1105–1106. Barnich N, Darfeuille-Michaud A. Adherent-invasive Escherichia coli and Crohn’s disease. Curr Opin Gastroenterol 2007; 23:16–20. Wei B, Huang T, Dalwadi H, Sutton CL, Bruckner D, Braun J. Pseudomonas fluorescens encodes the Crohn’s disease-associated I2 sequence and T-cell superantigen. Infect Immun 2002; 70:6567–6575. Lodes MJ, Cong Y, Elson CO, Mohamath R, Landers CJ, Targan SR, et al. Bacterial flagellin is a dominant antigen in Crohn disease. J Clin Invest 2004; 113:1296–1306. Prideaux L, De Cruz P, Ng SC, Kamm MA. Serological antibodies in inflammatory bowel disease: a systematic review. Inflamm Bowel Dis 2012; 18:1340–1355. Dubinsky MC, Kugathasan S, Mei L, Picornell Y, Nebel J, Wrobel I, et al. Increased immune reactivity predicts aggressive complicating Crohn’s disease in children. Clin Gastroenterol Hepatol 2008; 6:1105–1111. Elkadri AA, Stempak JM, Walters TD, Lal S, Griffiths AM, Steinhart AH, et al. Serum antibodies associated with complex inflammatory bowel disease. Inflamm Bowel Dis 2013; 19:1499–1505.
29 30
31
32
33
34
35 36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
Lewis JD. The utility of biomarkers in the diagnosis and therapy of inflammatory bowel disease. Gastroenterology 2011; 140:1817–1826.e2. Zhang Z, Li C, Zhao X, Lv C, He Q, Lei S, et al. Anti-Saccharomyces cerevisiae antibodies associate with phenotypes and higher risk for surgery in Crohn’s disease: a meta-analysis. Dig Dis Sci 2012; 57:2944–2954. Kaul A, Hutfless S, Liu L, Bayless TM, Marohn MR, Li X. Serum anti-glycan antibody biomarkers for inflammatory bowel disease diagnosis and progression: a systematic review and meta-analysis. Inflamm Bowel Dis 2012; 18:1872–1884. Whiting PF, Weswood ME, Rutjes AW, Reitsma JB, Bossuyt PN, Kleijnen J. Evaluation of QUADAS, a tool for the quality assessment of diagnostic accuracy studies. BMC Med Res Methodol 2006; 6:9. Zamora J, Abraira V, Muriel A, Khan K, Coomarasamy A. Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol 2006; 6:31. Deville WL, Buntinx F, Bouter LM, Montori VM, de Vet HC, van der Windt DA, et al. Conducting systematic reviews of diagnostic studies: didactic guidelines. BMC Med Res Methodol 2002; 2:9. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7:177–188. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005; 58:882–893. O’Donnell S, O’Sullivan M, O’Morain CA, Ryan BM. The clinical significance of antimicrobial serologic responses within an Irish Crohn’s disease population. Eur J Gastroenterol Hepatol 2013; 25:1464–1469. Papadakis KA, Yang H, Ippoliti A, Mei L, Elson CO, Hershberg RM, et al. Anti-flagellin (CBir1) phenotypic and genetic Crohn’s disease associations. Inflamm Bowel Dis 2007; 13:524–530. Ryan JD, Silverberg MS, Xu W, Graff LA, Targownik LE, Walker JR, et al. Predicting complicated Crohn’s disease and surgery: phenotypes, genetics, serology and psychological characteristics of a population-based cohort. Aliment Pharmacol Ther 2013; 38:274–283. Suzuki H, Fukuda Y, Koizuka H, Tomita T, Hori K, Suzuki M, et al. Dietary antigens in Crohn’s disease: antibodies to porcine pancreatic amylase. Am J Gastroenterol 2008; 103:656–664. Michielan A, Basso D, Martinato M, Pathak S, Banerjee A, Oliva L, et al. Increased antibody response to microbial antigens in patients with Crohn’s disease and their unaffected first-degree relatives. Dig Liver Dis 2013; 45:894–898. Papp M, Norman GL, Vitalis Z, Tornai I, Altorjay I, Foldi I, et al. Presence of anti-microbial antibodies in liver cirrhosis – a tell-tale sign of compromised immunity? PloS One 2010; 5:e12957. Kamat A, Ancuta P, Blumberg RS, Gabuzda D. Serological markers for inflammatory bowel disease in AIDS patients with evidence of microbial translocation. PLoS One 2010; 5:e15533. Lichtenstein GR, Targan SR, Dubinsky MC, Rotter JI, Barken DM, Princen F, et al. Combination of genetic and quantitative serological immune markers are associated with complicated Crohn’s disease behavior. Inflamm Bowel Dis 2011; 17:2488–2496. Landers CJ, Cohavy O, Misra R, Yang H, Lin YC, Braun J, et al. Selected loss of tolerance evidenced by Crohn’s disease-associated immune responses to auto- and microbial antigens. Gastroenterology 2002; 123:689–699. Muller S, Styner M, Seibold-Schmid B, Flogerzi B, Mahler M, Konrad A, et al. Anti-Saccharomyces cerevisiae antibody titers are stable over time in Crohn’s patients and are not inducible in murine models of colitis. World J Gastroenterol 2005; 11:6988–6994. Eser A, Papay P, Primas C, Pernicka E, Harrer M, Dejaco C, et al. The impact of intestinal resection on serum levels of anti-Saccharomyces cerevisiae antibodies (ASCA) in patients with Crohn’s disease. Aliment Pharmacol Ther 2012; 35:292–299. Targan SR, Landers CJ, Yang H, Lodes MJ, Cong Y, Papadakis KA, et al. Antibodies to CBir1 flagellin define a unique response that is associated independently with complicated Crohn’s disease. Gastroenterology 2005; 128:2020–2028. Amre DK, Lu SE, Costea F, Seidman EG. Utility of serological markers in predicting the early occurrence of complications and surgery in pediatric Crohn’s disease patients. Am J Gastroenterol 2006; 101:645–652. Papp M, Altorjay I, Dotan N, Palatka K, Foldi I, Tumpek J, et al. New serological markers for inflammatory bowel disease are associated with earlier age at onset, complicated disease behavior, risk for surgery, and NOD2/CARD15 genotype in a Hungarian IBD Cohort. Am J Gastroenterol 2008; 103:665–681. Arnott ID, McNeill G, Satsangi J. An analysis of factors influencing shortterm and sustained response to infliximab treatment for Crohn’s disease. Aliment Pharmacol Ther 2003; 17:1451–1457.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
742 European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 7
52 Esters N, Vermeire S, Joossens S, Noman M, Louis E, Belaiche J, et al. Serological markers for prediction of response to anti-tumor necrosis factor treatment in Crohn’s disease. Am J Gastroenterol 2002; 97:1458–1462. 53 Papp M, Foldi I, Altorjay I, Palyu E, Udvardy M, Tumpek J, et al. Anti-microbial antibodies in celiac disease: trick or treat? World J Gastroenterol 2009; 15:3891–3900. 54 van Schaik FD, Oldenburg B, Hart AR, Siersema PD, Lindgren S, Grip O, et al. Serological markers predict inflammatory bowel disease years before the diagnosis. Gut 2013; 62:683–688. 55 Harrer M, Reinisch W, Dejaco C, Kratzer V, Gmeiner M, Miehsler W, et al. Do high serum levels of anti-Saccharomyces cerevisiae antibodies result from a leakiness of the gut barrier in Crohn’s disease? Eur J Gastroenterol Hepatol 2003; 15:1281–1285.
56
Vermeire S, Peeters M, Vlietinck R, Joossens S, Den Hond E, Bulteel V, et al. Anti-Saccharomyces cerevisiae antibodies (ASCA), phenotypes of IBD, and intestinal permeability: a study in IBD families. Inflamm Bowel Dis 2001; 7:8–15. 57 Blumberg RS, Saubermann LJ, Strober W. Animal models of mucosal inflammation and their relation to human inflammatory bowel disease. Curr Opin Immunol 1999; 11:648–656. 58 De Winter H, Cheroutre H, Kronenberg M. Mucosal immunity and inflammation. II. The yin and yang of T cells in intestinal inflammation: pathogenic and protective roles in a mouse colitis model. Am J Physiol 1999; 276:G1317–G1321. 59 Strober W, Fuss IJ, Blumberg RS. The immunology of mucosal models of inflammation. Annu Rev Immunol 2002; 20:495–549.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
