Utility of faecal calprotectin analysis in adult inflammatory bowel disease

Abstract

The inflammatory bowel diseases (IBD), Crohn’s disease and ulcerative colitis, are chronic relapsing, remitting disorders. Diagnosis, along with assessment of disease activity and prognosis present challenges to managing clinicians. Faecal biomarkers, such as faecal calprotectin, are a non-invasive method which can be used to aid these decisions. Calprotectin is a calcium and zinc binding protein found in the cytosol of human neutrophils and macrophages. It is released extracellularly in times of cell stress or damage and can be detected within faeces and thus can be used as a sensitive marker of intestinal inflammation. Faecal calprotectin has been shown to be useful in the diagnosis of IBD, correlates with mucosal disease activity and can help to predict response to treatment or relapse. With growing evidence supporting its use, over the last decade this faecal biomarker has significantly changed the way IBD is managed.

Key Words: Faecal calprotectin; Inflammatory bowel disease; Crohn’s disease; Ulcerative colitis; Faecal biomarker

INTRODUCTION

It is desirable to have simple diagnostic tests and disease markers for use in the assessment and follow up of chronic diseases such as inflammatory bowel disease (IBD). These markers should be easy to perform, acceptable to both patients and health workers, economical, ideally non-invasive and have a high sensitivity and specificity. Calprotectin, first described in 1980[1], is a protein found in the cytosol of neutrophils and macrophages composed of two subunits S100A8 and S100A9. It can be detected in plasma, urine, cerebrospinal fluid, faeces, saliva, synovial fluid and colonic biopsies[2]. It is stable in faeces for up to seven days at room temperature and has a homogenous distribution in faeces[3], properties which lend it to testing spot faecal samples. There has been recent emphasis of the involvement of the innate immune system in the pathogenesis of IBD[4]. Calprotectin is classed as a damage associated molecular pattern protein (DAMP) having antimicrobial protective properties. DAMPs are released by the innate immune system from damaged or activated cells, initiating and perpetuating the immune response. The extracellular release of calprotectin during times of cell stress/damage makes it an accurate marker of intestinal inflammation.

As early as 1992 it was shown that faecal calprotectin is elevated in patients with both ulcerative colitis (UC) and Crohn’s disease (CD)[3] and this has been confirmed by subsequent studies[5-9]. Calprotectin levels in faeces correlate with faecal excretion of 111-indium labelled leucocytes, deemed to be the gold standard for measuring intestinal inflammation (r = 0.80, P < 0.0001)[5]. Disadvantages of this gold standard test compared with calprotectin are the exposure to radiation, cost and the three day collection of faeces. Faecal calprotectin levels have also been shown to correlate well with radiolabelled white cell scanning, another method of assessing intestinal inflammation, in adults with CD[10]. Spot faecal samples of < 5 g have been shown to be as reliable as 24 h collection samples for measuring calprotectin levels[3] indicating that calprotectin is evenly distributed throughout the faeces.

An elevated faecal calprotectin is not specific for IBD. Any inflammatory process within the gastrointestinal tract will result in the activation of the innate immune response and release of calprotectin. Faecal calprotectin concentration has been shown in studies to be elevated in many conditions including infection, colorectal cancer, untreated coeliac disease, microscopic colitis and diverticulitis[11-13]. Non-steroidal anti-inflammatory drugs (NSAIDs) have been shown to cause significant increases in faecal calprotectin levels within seven days due to NSAIDs induced intestinal inflammation with endoscopic correlation[14,15]. Proton pump inhibitors (PPIs) have been associated with significantly elevated faecal calprotectin levels, regardless of reason for PPI[16].

Initially faecal calprotectin concentration was reported in mg/L, but more recent assays (post 2000) usually report faecal calprotectin concentration as μg/g. To compare these results, faecal calprotectin concentrations obtained using assays pre-2000 need to be multiplied by a factor of five.

USE OF FAECAL CALPROTECTIN IN DIAGNOSIS OF IBD

Diagnosis of IBD has historically been based on a combination of clinical history and examination, blood parameters, radiology and endoscopy. The addition of a faecal biomarker able to reduce the need for invasive endoscopic procedures or exposure to radiation is advantageous.

Limburg et al[12], in 2000, published a study of 110 patients attending for colonoscopy for the investigation of chronic diarrhoea showing that increased faecal calprotectin levels were significantly (P = 0.0001) associated with the presence of colorectal inflammation (CD, UC, microscopic colitis or diverticulitis). Within the colonic inflammation subgroup, calprotectin concentrations were highest amongst subjects with IBD. The negative predictive value of faecal calprotectin in this dataset was 93%.

IBD and irritable bowel syndromes (IBS) can present in a similar clinical fashion with symptoms such as diarrhoea and abdominal pain. Routine colonoscopy in these patients is costly, invasive and has associated morbidity and mortality. Serum markers of inflammation such as C reactive protein (CRP) and erythrocyte sedimentation rate (ESR) in isolation are not sufficiently sensitive or specific for the diagnosis of IBD[7]. The use of faecal calprotectin to distinguish between IBD and IBS has been analysed in several studies. In 2000 Tibble et al[7] presented results of a prospective study of 220 consecutive patients in whom the principal differential diagnosis was that of either IBS or CD. They excluded patients with UC on sigmoidoscopy and biopsy. A diagnosis of CD was made from a combination of radiological, endoscopic and histological investigations. A diagnosis of IBS was made on basis of normal investigations and a compatible history fulfilling the Rome criteria. All patients subsequently diagnosed with CD had significantly higher faecal calprotectin concentrations than those with IBS. The investigators found that using a cut-off point of 30 mg/L faecal calprotectin had a 100% sensitivity and 97% specificity in discriminating between active CD and IBS.

Schoepfer et al[17] looked at the accuracy of faecal biomarkers alone and in combination with the IBD antibodies, antineutrophil cytoplasmic antibody (ANCA) and anti-Saccharomyces Cerevisiae manna antibody (ASCA), in discriminating IBD from IBS. They found that the overall accuracy of faecal calprotectin for discriminating between IBD and IBS was 89% (sensitivity 83%, specificity 100%). There was only a marginal increase in overall accuracy when faecal calprotectin was combined with IBD antibodies to 91%.

Faecal calprotectin has been studied as a tool to predict abnormal small-bowel radiology[18]. The study looked at 73 consecutive patients attending for small bowel follow through whose presenting symptoms were consistent with a possible diagnosis of IBD. The control group consisted of 25 patients with IBS, 25 normal volunteers and 25 patients with active CD. A faecal calprotectin level above 60 μg/g predicted all abnormal barium follow through results. The negative predictive value of a single calprotectin result below 60 μg/g of stool was 100% compared with 91% each for erythrocyte sedimentation rate cut off of 10 mm and C-reactive protein of 6 mg/L. Somewhat in contrast to this Sipponen et al[19] found that faecal calprotectin had a low utility for predicting the presence of small bowel CD on wireless capsule endoscopy, sensitivity was low at 59% with a moderate specificity of 71% using a cut-off of 50 μg/g.

A recent meta-analysis analysed 30 prospective studies comparing the diagnostic precision of faecal calprotectin against a histological diagnosis[20]. Summary receiver operating characteristic curve analysis showed a sensitivity of 0.95 (95% CI: 0.93-0.97), specificity of 0.91 (95% CI: 0.86-0.91), and an area under the curve (AUC) of 0.95 for the diagnosis of IBD. The diagnostic precision of faecal calprotectin for IBD was higher in children than adults with better accuracy at a cut-off level of 100 μg/g vs 50 μg/g. This meta-analysis also showed that faecal calprotectin was superior to CRP, ESR, ASCA, perinuclear anti-neutrophil cytoplasmic antibodies and anti-Escherichia coli outer membrane porin C antibody in diagnosis of IBD.

Thus it can be stated that a normal faecal calprotectin result, in the absence of ‘red flag’ symptoms and in the context of positive Rome criteria, is associated with a high likelihood of subsequent non-organic diagnosis and further endoscopic or radiological evaluation may be avoided in such patients. A meta-analysis published in 2010 to assess whether the use of faecal calprotectin reduces the number of unnecessary endoscopic procedures in the investigation of suspected IBD showed that screening with faecal calprotectin would result in a 67% reduction in the number of adults requiring endoscopy. The downside of this screening strategy is delayed diagnosis in 6% of adults because of a false negative test result[13].

Faecal calprotectin appears to better reflect disease activity in UC rather than CD[21] but faecal calprotectin has not been found to be useful in distinguishing UC from CD. Quail et al[22] looked at faecal calprotectin concentrations in Scottish children with a diagnosis of IBD; there was no statistical difference in calprotectin concentrations between CD and non-Crohn’s patients (UC or IBD type unspecified).

FAECAL CALPROTECTIN IN DISEASE ACTIVITY AND RESPONSE TO TREATMENT

In IBD, the presence of active gut inflammation is associated with migration of leucocytes, including neutrophils, to the gut mucosa[23]. As a result the faecal stream contains increased levels of these inflammatory proteins including calprotectin. Faecal calprotectin has been shown to differentiate quiescent from active disease in both patients with CD and UC[10,24-26]. Correlation of faecal calprotectin tends to be higher with endoscopic activity than clinical activity indices[24,27] and indeed some studies have demonstrated no significant correlation between faecal calprotectin and clinical indices[10,28]. In general faecal calprotectin correlates better with colonic CD rather than ileal disease[27-31] and an inflammatory rather than a structuring/penetrating phenotype[27,31]. Sipponen et al[24] showed that in active disease (CD endoscopic index of severity, CDEIS ≥ 3), faecal calprotectin concentrations were significantly higher in colonic than in ileal CD. Also, in limited ileal disease faecal calprotectin failed to correlate with endoscopic activity.

In UC Ricanek et al[32] showed that the median faecal calprotectin concentration was higher in patients with extensive and left sided disease distribution compared with proctitis (740 μg/g, 2106 μg/g, 86 μg/g respectively; P = 0.007 and P = 0.009). There was no significant difference in faecal calprotectin concentration between extensive and left sided disease distribution.

There have been several studies looking at the use of faecal calprotectin to predict or monitor response to treatment. In a study looking at 11 patients with relapsing IBD[33] (11 CD and 27 UC) faecal calprotectin was analysed at inclusion and after 8 wk of treatment (end of study). Treatment was individualised medical therapy. A normalised faecal calprotectin concentration at 8 weeks predicted a complete response in 100% patients. There was a significant decline in faecal calprotectin levels (P < 0.001) in patients with UC responding to treatment defined as normalisation of clinical and endoscopic scores. Within the small subgroup of patients with CD although 81% of patients achieved a complete clinical response defined clinically as a Harvey Bradshaw Index[34] (HBI) ≤ 5 there was no significant decline in calprotectin levels. This study was limited by small numbers and also the lack of endoscopic evidence of disease activity or remission in CD patients, it is possible that these patients had ongoing subclinical inflammation. In fact it has been shown that in patients with steroid induced clinical remission faecal calprotectin levels can remain elevated[35,36]. This finding is in keeping with earlier studies showing incomplete mucosal healing in patients treated with corticosteroids[37]. Sipponen et al[35] were able to show a significant decrease in faecal calprotectin (P = 0.005) in patients with CD who responded both clinically and endoscopically (using CDEIS) to an individualised escalation of treatment. There was no significant change in faecal calprotectin concentration in patients without endoscopic response.

Faecal calprotectin may be able to predict colectomy in patients with acute severe UC. Ho et al[38] showed that in patients with acute severe UC requiring inpatient treatment with intravenous corticosteroids faecal calprotectin was significantly higher in patients who failed to respond to medical therapy and required colectomy than those who did not (P = 0.04). The AUC was 0.65 (P = 0.04) for faecal calprotectin to predict colectomy with a maximum likelihood ratio of 9.23 at a cut-off of 1922.5 μg/g (specificity of 97.4%). Overall in the study faecal calprotectin concentrations were high with 86% of patients having levels of > 500 μg/g (median 1020 μg/g).

Faecal calprotectin can be used to monitor response to biological therapy. Palmon et al[39] showed that faecal calprotectin concentration decreases significantly at week 2 after an infliximab (IFX) infusion in 17 patients with CD on maintenance IFX therapy. Calprotectin levels were noted to rise back to baseline values by week 4 again despite a low median HBI. There was no endoscopic assessment of disease activity in this study. The rise in faecal calprotectin at week 4 may once again indicate a subclinical recurrence of mucosal inflammation. Sipponen et al[40] assessed the role of faecal calprotectin in monitoring clinical, using the CD activity index (CDAI) and CDEIS response to anti-tumor necrosis factor-α (TNF-α) therapy (IFX or adalumimab) in 15 patients with CD. Following 12 wk of treatment faecal calprotectin levels declined significantly from baseline level (P = 0.001) and changes in faecal calprotectin correlated to endoscopic appearances as scored using CDEIS (Spearman’s rank correlation r = 0.561, P = 0.03) suggesting that faecal calprotectin is a useful non-invasive marker of mucosal response to anti-TNF-α treatment.

PREDICTING MUCOSAL HEALING

Historically clinical practice has considered interpretation of symptoms and the use of scoring systems such as the CDAI, HBI and the Rachmilewitz UC activity index[41] to determine treatment success in IBD. These indices however tend to reflect patient well-being and quality of life rather than the degree of mucosal inflammation[10,28]. In both CD and UC there is evidence that mucosal healing is associated with sustained remission and reduced need for surgery[42,43] and following ileal resection the endoscopic appearance of the neoterminal ileum mucosa at 1 year post surgery has been shown to predict symptomatic relapse[44]. Thus mucosal healing is evolving into the new goal of IBD treatment.

Røseth et al[45] have demonstrated that normalisation of faecal calprotectin concentration corresponds to endoscopic mucosal healing. Seventeen patients with CD and 28 with UC clinically in remission who had faecal calprotectin concentrations of < 50 mg/L underwent endoscopic assessment of their lower GI tract and macroscopic mucosal appearances were assessed. Biopsies were also taken to assess histological inflammation. All but one of these patients with faecal calprotectin < 50 mg/L had inactive mucosal disease on colonoscopy.

Several subsequent studies have went on to show that concentration of faecal calprotectin correlates with both histological and endoscopic disease activity in IBD and these are summarised in Table 1. Furthermore, several of the studies included in Table 1 show that correlation of faecal calprotectin with endoscopic appearances is stronger than correlation with clinical indices[28,40].

Table 1 Correlation of faecal calprotectin with endoscopic findings, histology and clinical indices, when available.

Ref.	Patient population	Endoscopic assessment score	Correlation coefficient with endoscopy (r)	Correlation coefficient with histology	Correlation with clinical index (r)
Røseth et al[6]	UC	MAYO	0.57	NA	NA
D'Incà et al[8]	CD	SES-CD	0.48	r = 0.12	NA
D'Incà et al[8]	UC	MAYO	0.51	r = 0.32	NA
Sipponen et al[24]	CD	CDEIS	0.73	NA	0.397(CDAI)
Schoepfer et al[27]	CD	SES-CD	0.75	NA	NA
Jones et al[28]	CD	SES-CD	0.45	NA	0.23 (CDAI)
Sipponen et al[31]	CD	SES-CD	0.64	0.56 (exchange ileal only disease)	0.32 (CDAI)
Sipponen et al[40]	CD	CDEIS	0.83	0.52 (exchange ileal only disease)	NA
Langhorst et al[48]	UC	MAYO	0.49	NA	NA
Langhorst et al[48]	CD	SES-CD	0.35	NA	NA
Hanai et al[63]	UC	Matts	0.81	NA	0.68 (CDAI)
Schoepfer et al[64]	UC	Rachmilewitz	0.83	NA	NA

NA: Not available; MAYO: Mayo score[65]; SES-CD: Simplified endoscopic activity score for Crohn’s disease (CD)[66]; CDEIS: Crohn’s disease endoscopic index of severity[67]; CDAI: Crohn’s disease activity index[68]; Matts: Matts grading[69]; UC: Ulcerative colitis.

In contrast to the evidence shown in Table 1, Denis et al[46] failed to find a significant correlation between CDEIS and faecal calprotectin concentration. This was a small study of 28 patients with CD who had CDAI > 150 but a normal serum CRP. This lack of correlation may reflect the population studied as more than half of the patients had isolated ileal disease and overall disease activity was low (median CDEIS 3.4).

Interestingly one study in paediatric patients with IBD showed that calprotectin concentration correlated more closely with histological inflammation rather than endoscopic findings suggesting that faecal calprotectin may be more sensitive than macroscopic endoscopic appearances in evaluating disease activity status[47].

USE OF FAECAL CALPROTECTIN TO PREDICT RELAPSE

Being able to identify patients at high risk of relapse, and those with sub-clinical intestinal inflammation, may allow adjustment of their treatment strategy thus preventing clinical relapse. A non-invasive method of identifying this would reduce cost and risk of morbidity and mortality to patients. As sensitivity and specificity of serum markers of inflammation correlate poorly with intestinal inflammation their ability to predict disease relapse is poor[48,49]. Several studies have looked at the use of faecal calprotectin to predict relapse in patients with IBD and have demonstrated significant differences in faecal calprotectin concentration in relapsers compared with non-relapsers. A summary of these studies is shown in Table 2. Interestingly, faecal calprotectin appears less useful for predicting relapse in patients with ileal CD compared with patients with UC or colonic/ileocolonic CD[29,30].

Table 2 The use of faecal calprotectin to predict relapse.

Author	Patient population	Cut-off calprotectin level	Sensitivity for relapse (%)	Specificity for relapse (%)	Increased risk of relapse
Costa et al[21]	UC	150 μ/L	89	82	14 fold
	CD		87	43	2 fold
D'Incà et al[29]	UC	130 mg/kg	70	70	2.4 fold
	CD		65	62	1.7 fold
García-Sánchez et al[30]	CD + UC	120 μ/g	80	60
	Ileal CD	223 μ/g	83	50
Tibble et al[49]	CD + UC	50 mg/L (250 μ/g)	90	83
Laharie et al[50]	CD post IFX	130 μ/g	61	48
		250 μ/g	43	57
Kallel et al[70]	Colonic CD	340 μ/g	80	90.7	18.8 fold
Gisbert et al[71]	CD + UC	150 μ/g	69	69

UC: Ulcerative colitis; CD: Crohn’s disease; IFX: Infliximab.

The major outlier in these datasets is the study by Laharie et al[50] which looked at the use of faecal calprotectin to predict relapse specifically in 65 patients with CD treated with IFX induction regimen and then maintained on immunomodulator alone. There was no significant difference in faecal calprotectin concentration between those patients who relapsed by 14 wk post induction and those who did not even when the analysis was restricted to patients with pure colonic disease. No endoscopic evaluation was performed after IFX induction so although the study did show a median drop in faecal calprotectin concentration of 340 μg/g following induction there may have been ongoing subclinical inflammation. Another limitation of this study is that disease relapse was defined on clinical grounds alone.

Mao et al[51] performed a meta-analysis of the predictive capacity of faecal calprotectin in IBD relapse. Analysing 6 studies they found a pooled sensitivity of 78% and specificity of 73%. Capacity to predict relapse was comparable between CD and UC. Due to the small number of patients the predictive value of faecal calprotectin in ileal only CD patients was not assessed.

FAECAL CALPROTECTIN POST-SURGERY

More than 80% of CD patients require surgery within 10 years of diagnosis and by 3-5 years after surgery around a third of patients will have had a clinical relapse[44,52]. The role of faecal calprotectin in predicting post-surgical recurrence of CD has been assessed. One study assessed 39 patients with CD undergoing bowel resection[53]. The majority of patients (67%) had ileocolonic disease. Measurements of faecal calprotectin, CDAI and ultrasound examination were performed at 3 mo post surgery. Endoscopy was performed at 1 year regardless of patient symptoms and was considered the ‘gold standard’ for recurrence of disease. All patients were clinically in remission at 3 mo. Using a cut-off level of 200 mg/L, in predicting endoscopic post-surgical recurrence, faecal calprotectin has a sensitivity of 63% and specificity of 75% leading the authors to suggest that patients with elevated faecal calprotectin levels at 3 mo post surgery should be assessed endoscopically for early recurrence.

Lamb et al[54] prospectively followed 13 patients for 12 mo post ileocaecal resection for symptomatic CD. Faecal calprotectin was seen to normalise after uncomplicated surgery at 2 mo and remained within normal limits in 8 patients who remained clinically in remission. In the remaining 5 patients there was an increase in faecal calprotectin concentration after initial normalisation associated with disease recurrence or post operative intra-abdominal collections. The authors’ conclusion from this small number of patients was that patients with low levels of faecal calprotectin after resection who had symptoms were unlikely to have mucosal inflammation. It should be pointed out thought that there was no scheduled endoscopic evaluation of the gut mucosa during the 12 mo follow up period.

Pouchitis is an inflammatory condition with significant associated morbidity common in patients post restorative proctocolectomy[55]. Diagnosis can be difficult histologically due to the patchy distribution of inflammation. Faecal calprotectin has been shown to reliably differentiate between inflamed and non-inflamed pouches and correlates with severity of pouchitis[55-57] thus may reduce the need for endoscopic evaluation of the pouch in some patients.

USE OF FAECAL CALPROTECTIN TO AID DECISIONS OF WITHDRAWAL OF TREATMENT

There has been focus of late on when and in which patients it is appropriate to withdraw anti TNF-α therapy driven by costs and concerns about long term safety. Faecal calprotectin levels may assist in this decision making. Louis et al[58] published results of a prospective study (STORI) of CD patients who had been in steroid free remission on IFX for at least six months. Relapse after IFX withdrawal was associated with various risk factors including a faecal calprotectin concentration of ≥ 300 μg/g. Other risk factors for relapse identified on multivariate analysis included male sex, absence of surgical resection, leucocyte counts > 6 × 10/L, haemoglobin ≤ 145 g/L and hsCRP ≥ 5.0 mg/L. Patients with no more than 2 of the above risk factors had a reduced risk of relapse within 1 year (15% compared with 43.9% overall).

POINT-OF-CARE FAECAL CALPROTECTIN TESTING

Faecal calprotectin concentration is most commonly measured using an enzyme-linked immunosorbent assay (ELISA) technique in clinical laboratories and several ELISA kits are available. Recently point-of-care or bed-side faecal calprotectin tests have become available which may be advantageous in clinical circumstances such as primary care or when a more rapid result is required. These tests can be performed in less than 30 min. Correlation between faecal calprotectin concentrations measured using an ELISA technique and a point-of-care test has been shown to be good in recent initial studies[59-62].

CONCLUSION

Faecal calprotectin is now playing a major role in the investigation and diagnosis of patients presenting to the physician with lower gastrointestinal symptoms and can obviate the need for costly invasive investigations in selected patients.

Perhaps more excitingly faecal calprotectin is dramatically changing the way IBD is managed. There is a growing body of evidence that faecal calprotectin correlates well with mucosal disease activity and this in turn makes it useful in assessing activity, monitoring response to treatment, predicting relapse and aiding in the difficult decision of whether or not to withdraw biological therapy to a degree than clinical scoring indices or other non-invasive serological markers cannot.