CRP levels are raised in both Crohn’s disease (CD) and ulcerative colitis (UC), but many studies have demonstrated that CRP levels are significantly higher in CD than UC for all categories of disease severity. Although the reason is unknown, an explanation could be that UC is limited to the mucosa whereas CD involves a transmural inflammation of the gut wall. At IBD diagnosis, 25 % of patients with CD and 71 % with ulcerative colitis have CRP levels within the normal range. Indeed, CRP generation is extremely variable between individuals and may be related to variation in CRP genotype; a recent study demonstrated that elevated CRP levels could be associated with nucleotide polymorphisms in CRP genotype (rs1205, rs1130864, and rs1417938) at IBD diagnosis [3]. Hence, in the case of nonspecific digestive symptoms, CRP alone is not a good biological marker to differentiate IBD from functional disorders.

Measurement of clinical activity in IBD using the Crohn’s disease activity index (CDAI) or the Harvey–Bradshaw index (HBI) can suffer from subjective interpretation. Some studies have evaluated whether noninvasive markers such as CRP could be tools to help clinicians to measure disease activity objectively in CD patients. Solem et al. aimed to correlate CRP with clinical, endoscopic, histologic, and radiographic activity in CD: in this retrospective analysis of 104 patients, moderate-to-severe clinical activity, as well as active lesions at colonoscopy or histologically active inflammation, were significantly associated with elevated CRP levels. Conversely, abnormal radiologic findings in the small bowel were not significantly correlated with elevated CRP levels [4]. These results add weight to those of studies finding low CRP levels for an ileal location of CD in comparison to an ileocolonic or colonic location [5, 6].

A recent meta-analysis (19 studies) evaluated the diagnostic accuracy of CRP and other biological markers for the assessment of endoscopically detectable activity in symptomatic IBD patients. This study, which included 2499 IBD patients, showed that CRP ≥5 mg/L has a relatively high specificity (92 %) but poor sensitivity (49 %) for endoscopic activity. This result means that a negative value did not reliably exclude the possibility of active inflammation [7]. Similarly, Denis et al. found that in CD patients with clinically active disease (CDAI >150) and normal CRP levels (<5 mg/L), colonoscopy revealed endoscopic lesions, although these lesions were only mild (CDEIS ≤6) [8].

A high-level systemic inflammation in CD patients treated with anti-TNF alpha is associated with a positive clinical response to infliximab (IFX). Louis et al. found that the response rate was significantly higher in patients with an elevated (>5 mg/L) than a normal (<5 mg/L) CRP value before treatment (76 % versus 46 %; P = 0.004) [9], and the authors suggested that CRP level may help to identify better candidates for IFX treatment. In the same way, Jürgens et al. demonstrated that more patients with high baseline CRP levels responded to IFX than those with normal levels (P = 0.014). Moreover, early normalization of CRP levels correlated with sustained long-term response (P < 0.001) [10] without the need for a therapeutic adjustment [11].

Some studies have shown an increased risk of relapse in CD patients with elevated CRP during follow-up [12, 13]. Moreover, in a prospective study, Henriksen et al. demonstrated that high CRP levels at diagnosis were significantly associated with increased risk of subsequent surgery in IBD patients, especially those with ileal location (L1) and CRP >53 mg/L (OR = 6.0; P = 0.03) for CD patients and with extensive colitis and CRP >23 mg/L (OR = 4.8; P = 0.02) for UC patients [14].

A recent study indicated that following initiation of infliximab therapy, CRP could be predictive of loss of response in CD patients. Indeed, at week 22 of initiation, a CRP level >5 mg/L and a trough level (TL) >5.5 μg/mL, and the presence of anti-drug antibody (ADA) could predict a loss of response in 50 % of patients within 20 months [15].

CRP has been identified as a predictive factor for relapse in CD patients after anti-TNF withdrawal. Louis et al., in the prospective STORI study conducted by the GETAID group, investigated anti-TNF withdrawal in CD patients with longstanding remission of at least 6 months treated with combination therapy for at least 12 months (i.e., anti-TNF plus azathioprine or methotrexate). In this study, relapse occurred in 50 % of these patients within 18 months of IFX withdrawal. Based on multivariate analysis, the study demonstrated that a CRP level ≥5 mg/L before anti-TNF discontinuation was one of the risk factors for subsequent relapse [16]. In a post hoc analysis of this study, Meuwis et al. also found that the median CRP level, measured every 2 months until follow-up or relapse, was higher in relapsers compared with nonrelapsers (3.9 vs 2.8 mg/L; P = 0.07) [17].

Although some studies did not find a good correlation between FC and IBD clinical activity [22, 23], others have reported a significant correlation between FC levels and endoscopic activity in IBD. In CD patients, in a prospective study, Schoepfer et al. demonstrated that FC >70 μg/g was the most predictive marker of endoscopic activity (based on Simple Endoscopic Score for CD) with 80 % sensitivity and 60 % specificity, rather than CRP, blood leukocytes or the CDAI. In this study, FC was also the only marker that reliably discriminated inactive from mild, moderate, and highly active disease [24]. In the same way, Langhorst et al. demonstrated that FC was able to differentiate active IBD from inactive IBD and was consistently superior to CRP in its ability to reflect endoscopic inflammation [25].

In UC patients, Schoepfer et al. reported from a large study that FC more accurately reflects endoscopic activity (Spearman’s rank correlation coefficient, r = 0.821) and Nancey et al. also demonstrated a good correlation (r = 0.75) between FC and endoscopic activity, with levels above 250 μg/g associated with active disease (P < 0.001) [26, 27]. A recent meta-analysis evaluated the diagnostic accuracy of FC for the assessment of endoscopically defined disease activity in IBD. Including 2499 patients, this meta-analysis confirmed that FC is more sensitive than CRP to predict endoscopic activity (pooled sensitivity: 0.88 vs 0.49, respectively), both in UC and CD [7]. These data are particularly interesting in IBD patients with abdominal symptoms but with a lower clinical suspicion of active disease or relapse, in helping to decide who should be referred for further investigation.

Mucosal healing has become a therapeutic goal in IBD patients as it is associated with longstanding remission. D’Haens et al. demonstrated that FC levels ≤250 μg/g predicted endoscopic remission (defined by CDEIS ≤3) with 94.1 % sensitivity and 62.2 % specificity (PPV = 48.5 %, NPV = 96.6 %) [28]. Roseth et al. demonstrated that mucosal healing can be determined by a normalization of FC in IBD patients, showing that all patients with normal FC (median of 18 mg/L, range 1–50) had an appearance of both the colon and the terminal ileum that was completely normal endoscopically [29]. Moreover, in a recent study assessing IBD patients both in clinical remission and with mucosal healing , an elevated FC level alone was found in patients who relapsed compared to those without relapse (284 mg/kg vs 37 mg/kg; P < 0.01) suggesting the added value of FC over mucosal healing in the prediction of clinical remission [30].

Many studies have studied the interest of FC measurement to predict IBD outcome in clinical practice. Molander et al. demonstrated in a cohort of 60 IBD patients that a normalization of FC (<100 μg/g) after induction therapy with TNFα blocking agents was predictive of sustained clinical remission after 12 months compared to an elevated post-induction FC level (88 % vs 38 %; P < 0.0001) [31]. In addition, in acute, severe UC, Ho et al. found that FC concentrations remained significantly higher in patients requiring colectomy compared to those responding to medical therapy (1200 μg/g vs 887 μg/g; P = 0.04) [32].

Studies have also evaluated whether FC measurement for IBD patients in clinical remission could be helpful to predict disease outcome over time or to evaluate a possible de-escalation of treatment. Mao et al. reported a meta-analysis demonstrating that in quiescent IBD patients, the pooled sensitivity and specificity of FC to predict relapse were 78 % and 73 %, respectively, and comparable between CD and UC. Similarly, de Vos et al. studied the interest of repeated FC measurement in UC patients in remission under IFX maintenance therapy, and demonstrated that two consecutive FC measurements >300 μg/g with a 1-month interval were the best predictor of flare during a 52-week follow-up (61.5 % sensitivity and 100 % specificity) [33]. For CD patients with longstanding remission of at least 6 months (STORI study), Louis et al. demonstrated that an FC level ≥250 μg/g before anti-TNF discontinuation was a risk factor for subsequent relapse [16]. In the same way, in a recently published study, Ben Horin et al. found that after anti-TNF discontinuation (both IFX and adalimumab), an abnormal CRP and/or FC ≥50 μg/g was predictive of risk of relapse even in patients with endoscopic mucosal healing [34]. Thus, regarding these studies, discontinuation of medical therapy for patients with high FC levels appears to be a risk for relapse.

In the special situation of surgery, a recent study evaluated the accuracy and usefulness of FC to identify asymptomatically operated patients in postoperative endoscopic remission or recurrence; at a level of 100 μg/g, FC distinguished between endoscopic remission and recurrence with 95 % sensitivity and 54 % specificity. Due to an excellent negative predictive value (93 %), the authors concluded that 30 % of colonoscopies could be replaced by a simple FC measurement to exclude postoperative recurrence [35].

Lactoferrin (LF) is an iron-binding protein expressed by activated neutrophils and secreted by mucosal membranes. Some studies have reported similar performance of FC and LF tests [25, 36]. The mean sensitivity and specificity of fecal LF determination for the diagnosis of IBD are 80 % and 82 %, respectively [37]. Regarding Langhorst’s study, fecal markers (LF and FC) were able to differentiate IBD patients with endoscopically assessed inflammation from IBD patients without inflammation, and from IBS. In this study, neither of the two stool markers investigated was clearly superior in the ability to reflect endoscopic inflammation but they were both superior to CRP in diagnostic accuracy [25].

Recently, some studies have demonstrated that these antibodies can be associated with disease location. Indeed, ASCA+ is more frequent in CD patients with upper disease or with a pure small bowel involvement compared with CD patients with a pure colonic location (81 % and 68 % vs 38 %) or ileocolonic location [44, 47, 48]. In the same way, Targan et al. found that CBir1 was associated with small bowel disease (OR = 2.16; P = 0.009) [49].

Some studies have shown that serological markers, and hence high-level immune responses towards microbial antigens, are associated with a more severe disease course. Numerous studies indicate that ASCA positivity is correlated with an earlier disease onset [50, 51], which is a recognized factor of disease aggressivity. Moreover, ASCA positivity in CD is associated with a higher risk of complicated disease, such as stricturing or penetrating disease, and of small bowel resection [39, 52, 53]. Indeed, Vasiliauskas et al. demonstrated that all patients with a high ASCA level (>50 U/mL) and the ANCA negative immune marker subprofile developed fibrostenosis (14/14, 100 %) and most of them experienced internal penetrating complications (11/14, 79 %). In this study, taking into account the prevalence and number of surgical procedures per patient with small bowel involvement, analysis revealed that surgery was required by 86 % (12/14) of the high-level-ASCA/ANCA-negative subgroup [53]. By contrast, in patients with UC, the increased frequency of ANCA in treatment-resistant left-sided ulcerative colitis suggests a possible association between these antibodies and relative resistance to medical therapy [52]. Moreover, patients with very high levels of ANCA had a much greater risk of developing pouchitis following ileal pouch-anal anastomosis [52, 54].

In the same way, it has been demonstrated that the newly discovered antibodies could also be associated with disease severity. In the study of Targan et al., CBir1 was associated with complicated CD (internal-penetrating, fibrostenosing disease features) in 61 %, compared to 42 % of patients with inflammatory-only CD (P = 0.002) [49]. Anti-OmpC antibody has been detected in 55 % of CD patients [55] and its measurement could be useful at diagnosis in ASCA-negative patients; Mow et al. demonstrated that CD patients with anti-OmpC antibody were more likely to have internal perforating disease (50.0 % vs 30.7 %; P = 0.001) and to require small bowel surgery (61.4 % vs 44.2 %; P = 0.003), whereas anti-I2 is an independent marker of fibrostenosis (64.4 % vs 40.7 %; P < 0.001) and is also associated with small bowel surgery (62.2 % vs 37.4 %; P < 0.001) [39].

A very recent study by Paul et al. confirmed the usefulness of anti-glycan antibodies (AMCA, ACCA, and ALCA), alone or combined with ASCA or ANCA, for determination of the course of IBD. Measuring a large panel of anti-glycans in a cohort of 195 IBD patients (107 CD, 85 UC), (1) a severe CD course was significantly more likely in the case of high levels of AMCA, ASCA, and ACCA (OR = 4.3, 3.5, and 2.8, respectively), and (2) a severe UC course was significantly associated with AMCA and ACCA (OR = 3.4 and 3.0, respectively) [40]. However, tests for these anti-glycan antibodies are not widely available in clinical practice.