Catherine Reenaers is research fellow at the FNRS Belgium.
Assessment of endoscopic activity index and biological inflammatory markers in clinically active Crohn's disease with normal C-reactive protein serum level
Article first published online: 16 MAY 2007
Copyright © 2007 Crohn's & Colitis Foundation of America, Inc.
Inflammatory Bowel Diseases
Volume 13, Issue 9, pages 1100–1105, September 2007
How to Cite
Denis, M.-A., Reenaers, C., Fontaine, F., Belaïche, J. and Louis, E. (2007), Assessment of endoscopic activity index and biological inflammatory markers in clinically active Crohn's disease with normal C-reactive protein serum level. Inflamm Bowel Dis, 13: 1100–1105. doi: 10.1002/ibd.20178
- Issue published online: 3 AUG 2007
- Article first published online: 16 MAY 2007
- Manuscript Accepted: 6 APR 2007
- Manuscript Received: 5 FEB 2007
- AstraZeneca Belgium
- Crohn's disease;
Background: Patients with clinically active Crohn's disease (CD), defined by a Crohn's Disease Activity Index (CDAI) >150, may have normal C-reactive protein (CRP) serum levels. In such cases, it is difficult to know whether these patients have really active disease or rather functional symptoms. This distinction is important to decide the most appropriate treatment. The aim of our work was to assess intestinal and colonic lesions in such patients and to look for biological markers potentially associated with endoscopic activity of the disease.
Methods: We included 28 consecutive CD patients with CDAI >150 and a normal CRP level. These patients underwent a full colonoscopy with Crohn's Disease Endoscopy Index of Severity (CDEIS) calculation, fecal calprotectin, blood fibrinogen, acid α-1 glycoprotein, and erythrocyte sedimentation rate measurement. The Harvey–Bradshaw score was also calculated. Serum IL1 beta, IL6, IL8, sIL2R, and sTNFR2 were measured.
Results: The median CDAI was 181 (151–485). Almost all (92.9%) these patients had endoscopic lesions, but the majority had only mild lesions (CDEIS ≤6). No correlation was found between CDEIS and any of the clinical or biological markers. However, all the patients with significant endoscopic lesions (defined by a CDEIS >6) had previous surgical intestinal resection and lesions involving the anastomosis.
Conclusions: Patients with elevated CDAI and normal CRP have only mild mucosal lesions of CD. Most significant lesions may be observed at the anastomosis and proximal to it in previously operated patients. None of the biological markers tested was associated with these endoscopic lesions.
(Inflamm Bowel Dis 2007)
C rohn's disease (CD) is a chronic intestinal disorder that may affect every segment of the gastrointestinal (GI) tract. It is characterized by an inappropriate immuno-inflammatory reaction inducing chronic inflammatory lesions in the bowel wall.1 Although several genetic factors were identified2–6 and pathogenic mechanisms elucidated,7 the etiology remains unknown. The clinical course is characterized by a succession of relapses and remissions. The recognition of these different phases is important but sometimes difficult. The activity of the disease may be assessed using clinical,8, 9 endoscopic,10, 11 or biological12 scores. Recently, fecal calprotectin measurement also showed its interest in assessing bowel inflammation13, 14 and predicting clinical relapses.15 This highly sensitive marker, however, has lower specificity, since high calprotectin rates are found not only in CD and ulcerative colitis (UC) but also in infections, polyps, and neoplasms.14 Among biological markers, C-reactive protein (CRP), an acute-phase reactant member of the pentraxin family, has been the most widely used and is considered to be a reliable marker of disease activity and response to treatment.12
It is known that the correlation between these various markers is globally weak and inconstant at an individual level.16, 17 Therefore, it may sometimes be difficult to choose the best treatment according to these activity scores or markers. Particularly, some patients may have a clinically active disease, defined by a Crohn's Disease Activity Index (CDAI)8 higher than 150, and a normal CRP serum level. In such cases it is difficult to know whether these patients have really mildly active disease or rather functional symptoms, and which treatment can be proposed.
The aim of this study was to assess endoscopic activity and various biological inflammatory markers in patients with clinically active CD and normal CRP serum levels. We further aimed at identifying clinical or biological factors predictive for significant mucosal lesions in these patients.
MATERIALS AND METHODS
Between January 2005 and February 2006 we included consecutive CD patients with clinically active disease, defined by a retrospective CDAI higher than 150 and with normal CRP serum levels (≤6 mg/L) (Roche/Hitachi, Roche Diagnostic, Mannheim, Germany). The study was accepted by the Liège University Ethical Committee and all patients gave informed consent.
In addition to the CDAI, the Harvey–Bradshaw score9 was also calculated. Demographic and clinical characteristics of the patients were retrieved from medical notes and patient interview. Disease classification was performed using the Montreal classification.18 Patients with known active disease in the upper part of the GI tract were excluded (proximal to the terminal ileum).
All the patients underwent a full colonoscopy under anesthesia and the CDEIS (Crohn's Disease Endoscopic Index of Severity)11 was calculated, within 7 days of the blood and stool sampling and the CDAI calculation.
Routine Biological Markers
Blood fibrinogen, erythrocyte sedimentation rate (ESR), and acid α-1 glycoprotein were measured in blood samples by routine procedures. Stool samples were taken for the dosage of fecal calprotectin (Calprotectin ELISA, Novatec Immun Diagnostica, Germany).
Blood Cytokines and Soluble Receptors of Cytokines
Blood samples were taken for centrifugation and serum storage at −20°C until further use. Interleukin-6 (IL6, coated monoclonal antibody clone 677B6A2, Biosource, Belgium), Interleukin-8 (IL8, KAC1301, Biosource), Interleukin-1β (IL1β, KAC1211, Biosource), soluble Tumor Necrosis Factor receptor type 2 (sTNFR2, KAC 1771 Biosource) and soluble Interleukin-2 receptor (sIL2R, KHR0022, Biosource) were measured in duplicate using a commercial enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instruction. The recorded result was the mean of the 2 measurements. When differences between the 2 measurements were greater than 10%, a new measurement was performed.
Statistical analysis were performed using the Instat 2000 Program. Correlation analyses were performed using a nonparametric Spearman test. Comparison of proportions was carried out using Fisher's exact test. Continuous variables were compared using the Mann–Whitney test. CDEIS was categorized between clinically significant or not endoscopic lesions according to a threshold of 6 according to a recent GETAID study.19
A multivariate analysis by logistic regression was performed with the presence of significant endoscopic lesions as the dependent variable and demographic and clinical characteristics as well as biological parameters as independent variables. P-values were considered statistically significant at ≤0.05.
Clinical, Biological, and Endoscopic Characteristics
Twenty-eight patients were included, with a median age of 46 years (range, 23–68 years). Characteristics of the patients are shown in Table 1. When these patients developed a flare of the disease and were included in the present study, most of them were on stable therapy with either steroids, mesalazine, immunosuppressants, or infliximab. Two patients had a history of upper GI tract Crohn's lesions, but these were controlled by upper GI endoscopy and/or small bowel follow-through and were inactive at the time of the study.
|Duration of the disease (y)||10 (1-35)|
|Age at diagnosis (Montreal)||A1 (<16 y)||0|
|A2 (17-40 y)||23|
|A3 (>40 y)||5|
|Disease location (Montreal)||L1 (ileum)||15|
|L4 (upper GI) + L2||1|
|L4 + L3||1|
|Disease behavior (Montreal)||B1 (nonstrict., nonfist.)||16|
|B2 + P (perianal)||2|
|B3 + P||2|
|Previous intestinal resection||15|
|IL–6 (pg/mL) (normal value: <12 pg/mL)||12 (12-1231)|
|IL–1 (pg/mL) (normal value: <44 pg/mL)||44 (44-2452)|
|IL–8 (pg/mL) (normal value: <12 pg/mL)||12 (12-880)|
|sIL2R (pg/mL) (normal value: <101 pg/mL)||101 (101-1060)|
|STNFR2 (ng/mL) (normal value: <2.67 ng/mL)||5.37 (2.67-16.6)|
|CRP (mg/L) (normal value: <6 mg/L)||2.9 (0-6)|
|Fibrinogen (g/L) (normal value: 2.3-4.3 g/L)||3.73 (3-5.2)|
|ESR (mm/h) (normal value: <10 mm/h)||13 (2-31)|
|Acid alpha-1 glycoprotein (g/L) (normal value: 0.4-1.3 g/L)||0.91 (0.5-1.33)|
|Fecal calprotectin (μg/mL) (normal value: < 50 μg/mL)||91.35 (10-571.2)|
CDAI, Harvey-Bradshaw score, fibrinogen, CRP, acid-α1 glycoprotein, cytokines, and soluble cytokine receptor serum levels as well as fecal calprotectin concentrations are shown in Table 1.
Terminal ileum or neo-terminal ileum (in the case of previous surgical resection) was reached in all cases. All but 2 patients (92.9%) had endoscopic lesions. However, globally these lesions were very mild (Fig. 1): the median CDEIS was 3.4 (0–16) and only 9 patients (32.1%) had a CDEIS ≥6.19 Nine patients (32.1%) had strictures and only 2 patients (7%) had deep ulcerations. In the 12 patients without treatment or treated with mesalazine, median CDEIS was not different (2.8; range: 0–10).
CDAI was significantly correlated with Harvey–Bradshaw score (r = 0.42; P = 0.05) and fibrinogen (r = 0.51; P = 0.02). Correlation with CRP was borderline for significance (r = 0.39; P = 0.07). No significant correlation was found between CDAI and acid α1-glycoprotein, ESR, or any of cytokines or cytokine-soluble receptors. There was no correlation either between CDAI and the CDEIS or fecal calprotectin. CDEIS was not significantly correlated with any of the biological parameters tested.
Markers Associated with Endoscopic Significant Lesions
When comparing demographic and clinical characteristics, clinical activity scores, and biological markers between patients having or not having significant endoscopic lesions as defined by a thresholds of 6 for the CDEIS, there was no significant difference except for the previous history of surgical intestinal resection: all the patients (9/9) with a CDEIS ≥6 had a history of intestinal resection as compared to 6/19 patients with a CDEIS <6 (Table 2). The 13 patients without previous intestinal resection thus all had a CDEIS <6. Multivariate analysis did not disclose any other parameter associated with a CDEIS ≥6. Particularly, steroid treatment was not associated with more severe lesions.
|CDEIS ≥6 (n = 9)||CDEIS <6 (n = 19)||P value|
|IL6 (pg/mL)||12 (12-1231)||12 (12-225)||0.73|
|IL1 (pg/mL)||44 (44-44)||44 (44-2452)||0.29|
|IL8 (pg/mL)||12 (12-18)||12 (12-880)||0.23|
|sIL2R (pg/mL)||101 (101-368)||101 (101-1060)||0.76|
|sTNFR2 (ng/mL)||4.7 (2.7-9.7)||5.6 (2.9-16.6)||0.29|
|Calprotectin (μg/mL)||76.8 (11.7-324.1)||122.1 (10–571.2)||0.57|
|ESR (mm/h)||13.5 (2-30)||10 (3-31)||0.81|
|Fibrinogen (g/L)||3.7 (3-4.3)||3.6 (3-5.2)||0.97|
|Acid alpha-1 glycoprotein (g/L)||0.91 (0.86-0.95)||0.92 (0.5-1.33)||0.89|
|CRP (mg/L)||2.5 (1-6)||2 (0– 6)||0.4|
|CDAI||178 (156-485)||182 (151-247.4)||0.73|
|H–B||11 (7-36)||11 (8-15)||0.97|
|Previous intestinal resection||9/9||6/19||0.0008|
|Pure ileal disease||5/9||10/19||1.00|
|Disease duration (y)||17.5 (4-23)||7 (1-35)||0.27|
The clinical activity of CD may sometimes be difficult to assess. The origin of symptoms such as diarrhea, fatigue, or abdominal pain may be multifactorial and does not necessarily correlate with the existence of prominent inflammatory lesions of the GI tract. Facing a patient with symptoms compatible with active CD, the optimal therapeutic options are different if the symptoms are mainly driven by active inflammatory lesions or not. In clinically active CD, serum CRP measurement helps to confirm and quantify the importance of the inflammation. CRP level is a good predictor of relapses, remission, and response to treatment.13, 20 In the case of clinically active CD with normal CRP serum levels, it is essential to correctly assess the existence and the importance of the specific lesions of CD along the GI tract. The present work shows that endoscopic lesions in these patients were globally very mild and that only one-third of the patients had lesions that could be considered clinically significant.19 However, apart from a previous history of surgical resection, which was associated with more severe endoscopic lesions, mainly at the anastomosis, no other demographic, clinical, or biological marker was predictive of these significant lesions. The absence of correlation between CDAI, CDEIS, and CRP or other biological markers in our study is in contrast with previously published studies carried out in patients with either elevated or low CDAI and blood inflammatory markers.17 This probably reflects the very particular nature of the population we studied, with low CRP despite elevated CDAI. The same reason may explain the relatively weak correlation between Harvey–Bradshaw score and CDAI. Indeed, the relative importance of subjective symptoms and diarrhea, which are not necessarily associated with active inflammation, is greater in Harvey–Bradshaw score than in CDAI.
Symptoms suggesting active CD in the absence of elevation of CRP could reflect several mechanisms: the absence of intestinal lesions and functional symptoms, intestinal lesions not associated with systemic inflammation, or a genetic defect in CRP production or other inflammatory pathways. Our data suggest more arguments for the first 2 hypotheses. Although CRP production may be influenced by genetic background and one can speculate on a genetically determined defect of CRP production,21 this genetic influence has mainly been shown for illnesses associated with low levels of CRP, measured with techniques detecting ultrasensitive CRP.22 In CD no relationship was found between CRP level and CRP genotypes.23 Other blood inflammatory markers including several cytokines and soluble cytokine receptors were also very low in our population, confirming that the low levels of CRP we observed were probably not due to a specific defect in CRP production or to activation of other inflammatory pathways.
The vast majority of the patients had endoscopic lesions and no mucosal healing. These lesions were thus associated with very low systemic inflammation. Several characteristics of these lesions may help to explain the low systemic inflammation: more than half of these patients were treated with steroids that can suppress systemic inflammation but induce little mucosal healing,24 those lesions were superficial in most cases, there were more ileal than colonic lesions, one-third of the patients had strictures that increase endoscopic activity score but induce little systemic inflammation, and finally, most patients with a clinically significant endoscopic score had anastomotic lesions.
Superficial lesions could be associated with a limitation of the inflammation into the mucosa, as in UC. It is interesting to note that UC is also characterized by lower CRP serum levels compared to CD.25 Consistent with this, in our patients only 2 harbored deeper ulcers. In such situations, mucosal cytokines measurement in intestinal biopsies could give interesting information about local inflammation phenomenon. Strictures that were present in one-third of our patients could also have contributed to explain the globally low systemic inflammation in this cohort and also the absence of correlation between endoscopic score and other blood inflammatory markers. However, even when calculating the CDEIS without taking into account the strictures (data not shown), no correlation was found between this modified score and all of these other inflammatory markers. More than half of our patients had pure ileal disease. Ileal lesions are probably associated with less CRP elevation than colonic lesions.26 A recent observation from Florin et al27 about clinically active CD with low CRP levels also found more ileal CD in this group of patients, although no systematic endoscopy was performed at the time of CRP measurement in that study. The reason why ileal lesions induce less CRP is not precisely known. This could also have influenced our results, although in the multivariate analysis disease location was not associated with clinically significant endoscopic lesions.
Fecal calprotectin was mildly to moderately increased in our cohort but there was no correlation with the intensity of the endoscopic lesions assessed by the endoscopic score. It has been suggested that this marker is a better marker for colonic than ileal lesions. The large proportion of patients with ileal disease in our study may have contributed to this absence of correlation. The discrepancy might also have been due to undetected jejunal or proximal ileal lesions. Two patients of our cohort had upper GI tract lesion in the past. However, those were GI lesions and these patients underwent an upper GI tract exploration that showed no significantly active disease at the time of the present study. Although the value of fecal calprotectin is well established to differentiate between irritable bowel syndrome (IBS) and inflammatory bowel disease (IBD) patients,14 the correlation between calprotectin level and intensity of mucosal endoscopic lesions in CD has not been well established.28, 29
When looking at the patients with a CDEIS threshold of 6,19 the only predictive factor of endoscopic active disease was the presence of previous intestinal resection. Indeed, all the patients without previous surgery had a CDEIS lower than 6, while all the patients with a CDEIS ≥6 had undergone previous surgery. Lesions in these patients were mainly located around and proximal to the anastomosis. A low level of CRP in these patients with previous ileal resection evolving toward a stricturing behavior has already been reported.27 Nonspecific factors due to the anastomosis, like proximal luminal hyperpressure or ischemic peri-anastomotic phenomenon, may be involved in the development of these anastomotic lesions. Still, the implication of such mild lesions in the symptoms of the patients is questionable, particularly because we did not observe any correlation between clinical activity index (CDAI) and endoscopic activity index (CDEIS) or even a biological marker of mucosal inflammation, the fecal calprotectin. Quite surprisingly, a significant correlation was found between CDAI and fibrinogen serum level as well as a borderline correlation with CRP. Therefore, despite their level within the normal range, this correlation with the systemic inflammatory markers may suggest the implication of the systemic inflammatory reaction in the elevation of the CDAI. Besides this, functional symptoms are also most probably involved. Diarrhea may be due to rapid transit time or biliary salt malabsorption, particularly in patients with previous ileal resection. Abdominal pain may be due to peritoneal adherence, quite frequent in CD patients, particularly following surgery. IBS may also explain both diarrhea and abdominal pain. A recent study showed that up to 50% of CD patients may present with IBS.30 Fatigue is a very frequent symptom in CD.31 Furthermore, this symptom is very difficult to treat and tends to persist despite remission of the disease.
In conclusion, in almost all cases, patients with clinically active CD but normal CRP had active mucosal lesions at endoscopy, although these were usually mild. Several factors may explain the low CRP in this group of patients, including ileal or stricturing lesions, as well as steroid treatment. Within the present cohort, the only clinical or biological factor associated with clinically significant endoscopic activity was a previous surgical resection. Therapeutic choice in these patients remains difficult, since there was no correlation between endoscopic activity and clinical activity of the disease and the impact of medical treatments on very mild mucosal lesions or on strictures is questionable. However, it is difficult to fully consider these patients as functional bowel disorder, not only because they still harbored mucosal lesions, but also because the clinical activity of the disease correlated with blood markers of inflammation, albeit these were within the normal range.
- 19Endoscopic remission and response in Crohn's disease: an objective definition using CDEIS. Gut. 2005; 54(Suppl VII): A50., , , et al.