Background: Clinical and laboratory assessment of activity in Crohn's disease (CD) correlate poorly with endoscopic findings. Calprotectin is a calcium-binding protein abundant in neutrophil cytosol, and extremely stable in faeces. Faecal calprotectin (FC) is an excellent surrogate marker of neutrophil influx into the bowel lumen.
Aim: To assess whether FC concentration from a spot stool sample reliably detects active inflammation in patients with CD.
Design: Cross-sectional comparative study.
Methods: Subjects had a previously confirmed diagnosis of CD and were suspected on clinical grounds to be in the midst of a relapse. Thirty-five entered the study; they underwent radiolabelled white cell scanning (WCS) and had a stool sample collected for calprotectin measurement on the same day. A Crohn's disease activity index (CDAI) was also calculated for each. The WCS scans were scored at six standard sites to give a mean total, ‘extent‘, ‘severity’ and ‘combined extent and severity’ scores.
Results: FC was significantly and positively correlated with mean total (r = 0.73, p<0.001), ‘extent’ (r = 0.71, p<0.001), ‘severity’ (r = 0.64, p<0.001) and combined ‘extent and severity’ WCS scores (r = 0.71, p<0.001). A cut-off of faecal calprotectin >100 µg/g gave a sensitivity of 80%, specificity of 67%, positive predictive value of 87% and a negative predictive value of 64% in identifying those with and without any inflammation on WCS. There was, however, no significant correlation between CDAI and mean total WCS score (r = 0.21, p = 0.24), nor between CDAI and FC (r = 0.33, p = 0.06).
Discussion: While the CDAI does not accurately reflect inflammatory activity in CD, a one-off FC reliably detects the presence or absence of intestinal inflammation in adult patients with CD, compared to WCS.
Crohn's disease (CD) is a chronic relapsing and remitting inflammatory disorder of the gastrointestinal tract, the incidence of which is increasing in Scotland.1,,2 No cure currently exists. Of those who enter a medically induced remission, 30–60% will relapse within 1 year.3 Five years from diagnosis, about half of patients will require surgery.4
Clinical assessment of disease activity and laboratory indices of inflammation correlate poorly with endoscopic findings and histology in patients with inflammatory bowel disease (IBD).5 The search for faecal markers of intestinal inflammation in CD has been fuelled both by the difficulty in endoscopically imaging the distal small bowel and by the invasive nature of repeated endoscopic examinations.
Calprotectin, first described in 1980,6 is released from the cytoplasm of activated neutrophils, where it accounts for 60% of the protein in the cytosol. Lower concentrations are found in monocytes and macrophages. Its name is derived from its calcium binding and antimicrobial actions.7,,8 The stability of the protein to degradation keeps it stable in faeces for up to 7 days9 at room temperature, making it an ideal analyte. It was hoped that measurement of faecal calprotectin (FC) would represent a surrogate marker of neutrophil influx into the bowel lumen and in turn act as a marker of intestinal inflammation. Increased levels of FC are found in IBD,9,,10 colonic cancer11 and non-steroidal anti-inflammatory drug (NSAID) treatment,12,,13 suggesting it is a sensitive, but non specific, marker of intestinal inflammation.
The Crohn's disease activity index (CDAI) is a validated clinical/laboratory scoring system, used largely in trial settings to assess the activity of disease and the response to treatment.14 Despite limitations,15–17 the score has been widely used to evaluate drug efficacy in CD over a 25-year period; secondly, and perhaps more importantly, it reflects and correlates with clinical assessment and patient well-being, which is the general basis on which treatments for CD are currently justified.
Another tool for Crohn's disease activity assessment is the radiolabelled white cell scan (WCS), which gives information on the approximate location and severity of bowel inflammation.18 Autologous labelled leukocytes are returned to the host, and migrate preferentially to areas of inflammation. Patients often prefer this investigation to repeated colonoscopies. In addition, WCS is useful in differentiating inflammatory from fibrotic strictures in CD,19 in distinguishing functional symptoms from those secondary to inflammation,20 and in assessing the efficacy of treatment in IBD.21,,22 However, WCS is expensive, incurs a radiation dose and cannot define complications of CD such as fistulation.
We aimed to compare FC, CDAI and WCS in the assessment of disease activity in symptomatic patients with CD.
Participants of the study had a previously confirmed diagnosis of CD and at out-patient assessment had signs or symptoms suggestive of a relapse. Consultant and middle-grade gastroenterology staff recruited patients to the study. For each patient, a CDAI was calculated, WCS performed and sample collected for FC; the latter two occurring on the same day.
Patients were excluded from the study if they were receiving oral prednisolone, proton pump inhibitors or NSAIDs, if they had a positive stool culture, if they only had distal/perianal disease, or if they had undergone a previous surgical resection. Anti-inflammatory and immunosuppressant medications for CD were continued, and no new medications were commenced until the CDAI was obtained.
Samples for FC measurement were collected in disposable plastic containers and stored within 6 h at −20°C until analysis. FC was measured by a commercial enzyme linked immunoassay (ELISA) method (Calprest, Calprotech Ltd), following the manufacturers instructions. In brief, about 0.1 g of thawed faeces was accurately measured and diluted 1:50 (weight:volume ratio) with extraction buffer. After vigorous mixing for 30 s and homogenization on a roller for 25 min, the homongenate was centrifuged at 10 000 g for 20 min. The clear supernatant was then diluted 1:50 with dilution buffer. Diluted faecal samples, standards and quality control specimens were analysed in duplicate; 100 µl was added to wells of the microtitre plate and incubated for 45 min at room temperature. After thorough plate washing, 100 µl rabbit anti-human calprotectin labelled with alkaline phosphate was added to each well and incubated for 45 min. After a further thorough washing, 100 µl of substrate was added and the plate was incubated for 30 min in the dark at room temperature. The optical density was measured at 405 nm and results were calculated using an automatic plate reader. The FC results were expressed in µg faecal calprotectin per gram of wet faeces, with the normal reference range being <50 µg/g (Calprest).
WCS was performed at the Nuclear Medicine Department at Glasgow Royal Infirmary. Blood (40 ml) was recovered from each patient and centrifuged to obtain the white cell pellet, which was incubated with 200 mBq of 99mtechnetium hexamethyl-propylene-amine-oxime (99mTcHMPAO). The labelled cells were then re-injected into a fasting patient. Spot planar images of the abdomen, pelvis and perineum were taken at 1 and 2 h post injection. Interpretation of WCS was performed by three experienced observers independently, all blinded to the clinical information and FC results. The inflammation was scored at six standard sites (rectum and sigmoid, descending colon, transverse colon, ascending colon and caecum, terminal ileum and other small bowel) using a scoring system previously described23 (Table 1). The mean total score reflects the average of the three observers individual scores, giving a range of 0–18. Three further scores of inflammation were generated for the WCS, based on those published elsewhere.24 The ‘extent’ score reflects the total number of regions with a score ≥1 giving a score range of 0–6. The ‘severity’ score reflects the highest regional score of any of the six regions, giving a score ranging from 0–3. Finally a ‘combined extent and severity’ score, derived from summating the ‘extent’ and ‘severity’ scores, thus giving a score range of 0–9.
Abnormal activity, intensity > bone marrow, but ⩽liver
Abnormal activity, intensity > liver, but < spleen
Adapted from references 23 and 24.
On entry into the study, each patient was given a 7-day symptom recording diary for CDAI data collection. The CDAI for each study entrant was calculated by our nurse specialist using the accepted proforma.14
Statistical analyses were performed using Minitab version 13 (Minitab Inc). Correlation analysis between CDAI, WCS and FC was undertaken by calculating the Pearson correlation coefficient. A p value <0.05 was considered significant.
The study was granted full ethical approval by the Glasgow Royal Infirmary Research Ethics Committee (project no. 01GA005). Each participant received a verbal and written explanation of the trial outline, and written consent was obtained.
A total of 35 out-patients (22 male) were entered into the study. The ages of the participants ranged from 19 to 51 years, mean 31. CD had been previously documented in the colon only (20 patients), small bowel only (7 patients) and both colon and small bowel (8 patients). The range of CDAIs from the 35 out-patient subjects was 25–290, mean 158. Although all subjects were thought clinically to be in the midst of a relapse, CDAI was <150 in 18/35 (51%). Mean total WCS score was 2.45, range 0–7.7. An abnormal (mean total WCS score >0) WCS was seen in 24/35 (71%) of study entrants.
The faecal calprotectin results ranged from 5 to 7625 µg/g, median 195 µg/g (Figure 1). The majority had a raised faecal calprotectin (30/35, 86%) compared to the ‘healthy’ adult reference range (<50 µg/g). Faecal calprotectin correlated significantly with mean total WCS score (r = 0.73, p<0.001, Figure 2), ‘extent’ WCS score (r = 0.71, p<0.001, Figure 3), ‘severity’ WCS score (r = 0.64, p<0.001, Figure 4) and ‘combined extent and severity’ WCS score (r = 0.71, p<0.001, Figure 5). There was, however, no significant correlation between the CDAI score and either the FC (r = 0.33, p = 0.06, Figure 6) or mean total WCS score (r = 0.21, p = 0.24, Figure 7).
The sensitivity and specificity of faecal calprotectin in discriminating between a positive (mean total WCS score >0) or negative (mean total WCS score = 0) WCS is illustrated in the receiver operating characteristic (ROC) curve analysis (Figure 8). At a value of 100 µg/g, faecal calprotectin has a sensitivity of 80%, specificity of 67%, positive predictive value of 87% and negative predictive value of 60% in detecting those Crohn's disease patients with and without inflammation on WCS. The area under the ROC curve measures the probability of correct diagnostic classification (test accuracy), and was 87%.
Receiver operating characteristic (ROC) curve of the ability of faecal calprotectin to discriminate between those Crohn's disease patients with and without a positive WCS. The diagonal line from the origin indicates no discriminatory value whatsoever.
Assessment of disease activity in ulcerative colitis is relatively straightforward, as the disease is always within the reach of the flexible sigmoidoscope, and this allows subsequent histological analysis. This is not always possible in CD, because of the variable location and patchy distribution of the disease. Active small-bowel CD is often beyond the reach of the push enteroscope, and while barium studies of the small bowel are useful to exclude significant stenoses or fistulation, they are poor at detecting mucosal inflammation and/or ulceration which can be detected on capsule endoscopy.25 Thus in the absence of histology, an accurate measure of bowel inflammation in CD is controversial.
Simple clinical assessment of disease activity and laboratory indices of inflammation has been shown to correlate poorly with endoscopic findings and histology in patients with IBD.26 The CDAI, a combined clinical and laboratory score, was developed in North America; on multivariate regression analysis, the investigators found eight variables which best predicted the overall rating for each patient.14 The score was validated prior to the beginning of the National Cooperative Crohn's Disease Study.27 CDAI of <150 denotes quiescent disease; active disease scores from 150 to 450. The score has been widely used to evaluate drug treatments in Crohn's disease over a 25-year period and does reflect patient well-being,28 the basis on which current treatments are prescribed. However, numerous problems exist with the CDAI. Many variables on the score are subjective, identifying symptoms which are often suffered by patients with IBS, in addition to those with IBD. A large proportion of the score depends on the patients’ perception of the disease, and this can be influenced by the euphoric effect of corticosteroids used in the treatment of CD.15 Substantial inter-observer variability exists when different observers review the same case notes to calculate the CDAI.16 The score includes the hematocrit, which reflects disease activity poorly.17 Bowel frequency, one of the score variables, can be raised for reasons other than active disease, such as resective surgery causing bile salt malabsorption. The need for a 7-day patient symptom record precludes the use of the CDAI in everyday clinical practice. Finally, it is well known that the CDAI underestimates Crohn's disease activity when compared to excretion of autologous indium labelled granulocytes in faeces29 or with endoscopic evaluation.30 Thus we were not surprised by the lack of correlation between the CDAI and either the faecal calprotectin or the WCS result. This provides evidence that the CDAI does not measure only inflammatory activity; rather it is influenced by subjective symptoms often associated with functional disorders. Other investigators have only found a weak correlation between clinical activity scores and faecal calprotectin in inflammatory bowel disease.31 This is similar to data comparing faecal excretion of radiolabelled white blood cells and clinical activity scores.23,29,32,,33
Another option to assess disease activity in CD is WCS. This gives information on the approximate location and severity of bowel inflammation. Autologous labelled leukocytes are returned to the host and migrate preferentially to areas of inflammation. At our institution we frequently use 99mTc-HMPAO (Ceretec) scanning to help with the follow-up of patients with CD, and several studies have supported the view that WCS gives excellent sensitivity, specificity and accuracy in detecting disease extent and activity.18–20,34,,35 Indeed, in one study,18 WCS gave a 90% sensitivity, 97% specificity and 93% accuracy, compared to endoscopy for diagnosis. WCS is particularly useful in distinguishing functional symptoms from those secondary to inflammation,20 as well as refining treatment options in stricturing Crohn's disease where inflammatory strictures (amenable to anti-inflammatory/immunosuppressant treatment) are reliably distinguished from fibrostenotic disease (amenable to surgical resection/repair). In addition, patients often prefer this investigation to repeated colonoscopies, and data are emerging in support of this method to assess the efficacy of treatments in IBD.21,,22 Imaging must however be performed at the correct times, specifically 1–2 h post injection, as non-specific activity occurs beyond 3 h.36 The drawbacks of WCS are its cost (approximately £200 per patient), radiation dose, lack of availability in most hospitals, failure to localize the affected bowel segment, especially in previously operated patients, and the significant false positive rate, which limits its usefulness in investigation of IBD, especially when there is a low clinical probability of inflammation.37 It is also not a useful test in defining the complications of CD, such as fistulation.
Calprotectin is released by activation of leukocytes, giving increased levels in plasma, cerebral spinal fluid, synovial fluid, urine or stools as a consequence of disease in the relevant organ(s).38 Calprotectin inhibits zinc-dependent enzyme systems, as a result kills microbes,7 and induces apoptosis in normal and cancer cells.39 In the presence of calcium, calprotectin is remarkably resistant to proteolytic degradation, and so is stable in stools kept at room temperature for 7 days.9 Thus faeces has become the standard matrix for measuring this analyte, and indeed this has facilitated the participants of pivotal trials in the investigation of this protein to post their stool sample to the lab for analysis.9,,24 The routine management of out-patients with IBD in Oslo, Norway also uses the postal service in this way.40
It was hypothesized that measurement of FC would represent a surrogate marker of neutrophil influx into the bowel lumen, and in turn act as a marker of intestinal inflammation. Studies thus far support this hypothesis; increased levels being found in IBD,9,,10 colonic cancer11 and NSAID enteropathy,12,,13 suggesting that it is a sensitive, but non-specific, marker of intestinal inflammation. Clinical assessments of disease activity and laboratory indices of inflammation correlate poorly with endoscopic findings and histology in patients with IBD,25 but FC correlates well with endoscopic and histological grading of disease activity in ulcerative colitis.41 Moreover, faecal calprotectin has been shown to correlate more closely to histological grading of disease severity than to the macroscopic (colonoscopic) findings,24 suggesting it is a more sensitive test in IBD than endoscopy. A 4-day faecal excretion of indium-111-labelled granulocytes has been suggested as a ‘gold standard’ for assessing intestinal inflammation in Crohn's disease.42 However, this test is costly to perform, requires sterile labelling facilities, exposes the patient to radiation, and incurs all the practical problems of complete faecal collection over a 4-day period. Hence its use is limited to only a few research centres. The strong positive correlation between both daily FC (r = 0.87, p<0.01), and one-off FC (r = 0.8, p<0.001) and faecal excretion of the indium-111-labelled neutrophils43 supports the hypothesis that FC reflects the migration of neutrophils through the inflamed gastro-intestinal mucosa.
Our data revealed good correlation between faecal calprotectin and mean total, ‘extent’ ‘severity’ and ‘combined extent and severity’ WCS scores. This is similar to a report published in a paediatric population,24 but represents the first data in adults and is supportive of faecal calprotectin being a marker of both the length of bowel affected and the severity of that inflammation in CD. ROC curves (Figure 8) provide a means of investigating the discrimination of a test across a range of cut-off values and plot a score's sensitivity (i.e. true positives) on the y axis against 1-specificity (i.e. false positives) on the x axis. Tests with good discrimination have curves that tend toward the left upper corner, while curves for weaker tests lie close to the line from the origin at 45 degrees to the x axis. The best diagnostic cut-off can be assessed from the curve. In our population of CD patients, a faecal calprotectin of >100 µg/g gave a sensitivity of 80%, specificity of 67%, positive predictive value of 87%, negative predictive value of 64% and an accuracy of 87% in identifying those with and without any inflammation on a WCS. This suggests that a one-off faecal calprotectin in a patient with CD gives a reliable indication of the presence or absence of active inflammation, and an accurate reflection of the extent and/or severity of that inflammation.
In conclusion, the measurement of FC has several advantages in the assessment of disease activity in Crohn's patients. It is cheap, non-invasive, sensitive and a relatively straightforward to perform, requiring only a tiny stool sample, which may be posted to the lab. While the CDAI does not accurately reflect inflammatory activity in CD, a one-off faecal calprotectin reliably detects the presence or absence of intestinal inflammation in adult patients with CD, when compared to WCS. In addition the FC correlates with scores of both severity and extent of inflammation on WCS. In the future, a spot stool sample test for calprotectin may replace more invasive tests in the assessment of those with known CD.
Barton JR, Gillon S, Ferguson A. Incidence of inflammatory bowel disease in Scottish children between 1968 and 1983; marginal fall in ulcerative colitis, three fold rise in Crohn's disease. Gut1989; 30:618–22.
Holmquist L, Ahren C, Fallstrom SP. Clinical disease activity and inflammatory activity in the rectum in relation to mucosal inflammation assessed by colonoscopy. A study in children and adolescents with chronic inflammatory bowel disease. Acta Paediatr Scand1990; 79:327–34.
De Dombal FT, Softley A. I0IBD report no. 1: observor variation in calculating indices of severity and activity in Crohn's disease. International Organisation for the Study of Inflammatory Bowel Disease. Gut1987; 28:474–81.
Holmquist L, Ahren C, Fallstrom SP. Clinical disease activity and inflammatory activity in the rectum in relation mucosal inflammation assessed by colonoscopy. A study of children and adolescents with chronic inflammatory bowel disease. Acta Paediatr Scand1990; 79:527–34.
Teahon K, Smethurst P, MacPherson AJ, et al. Intestinal permeability in Crohn's disease and its relation to disease activity and relapse following treatment with elemental diet. Eur J Gastroenterol Hepatol1993; 5:79–84.
Scholmerich J, Schmidt E, Schumichen C, et al. Scintigraphic assessment of bowel involvement and disease activity in Crohn's disease using technetium-99m hexamthyl propylene amine oxine as a leukocyte label. Gastroenterology1988; 95:1287–93.
Yiu S, Mikami M, Yamazaki M. Induction of apoptoic cell death in mouse lymphoma and human leukaemia cell lines by a calcium-binding protein complex, calprotectin, derived from inflammatory peritoneal exudate cells. J Leukoc Biol1995; 336:763–5.
Roseth AG, Schmidt PN, Fagerhol MK. Correlation between faecal excretion of indium-111-labelled granulocytes and calprotectin, a granulocyte marker protein in patients with inflammatory bowel disease. Scand J Gastroenterol1999; 34:50–4.