Calprotectin and lactoferrin in the assessment of intestinal inflammation and organic disease
- First Online:
- Cite this article as:
- D’Incà, R., Dal Pont, E., Di Leo, V. et al. Int J Colorectal Dis (2007) 22: 429. doi:10.1007/s00384-006-0159-9
- 458 Views
Background and aims
Calprotectin and lactoferrin are specific neutrophil-derived proteins, which can be measured in the feces because they are released by cells in inflammatory conditions. We evaluated the efficacy of calprotectin and lactoferrin in detecting organic disease as assessed by colonoscopy.
The study comprised 144 patients undergoing colonoscopy for lower gastrointestinal symptoms (abdominal pain, altered bowel habits, and bloody stools) (67), or inflammatory bowel disease activity, or surveillance for dysplasia (77). A single stool sample was assayed for calprotectin and lactoferrin. The proportion of patients correctly diagnosed with each test and the relationship with endoscopic and histological findings were measured.
Fecal excretion of calprotectin significantly correlated with the finding of colonic inflammation at endoscopy, both in ulcerative colitis and in Crohn’s disease (p<0,001 and p<0,008, respectively), while lactoferrin excretion significantly correlated with histological inflammation (p=0.001 and p=0.009 respectively). Recommended cut-off values need to be adjusted in the inflammatory bowel disease group. Overall sensitivity, specificity, positive predictive value, and diagnostic efficacy were 78, 83, 86, and 80% for calprotectin and 80, 85, 87, and 81% for lactoferrin, respectively.
Fecal calprotectin and lactoferrin appear to be equally recommendable as inflammatory disease markers in patients with lower gastrointestinal symptoms. Both tests are needed to accurately discriminate activity in inflammatory bowel disease patients.
KeywordsCalprotectinColon cancerCrohn’s diseaseLactoferrinUlcerative colitis
Calprotectin and lactoferrin are produced in significant amounts in white blood cells. The mucosal barrier is altered in intestinal inflammation, allowing white cells to cross the intestinal wall. Activated leukocytes infiltrate the mucosa and can be detected in feces due to shedding in the intestinal lumen [1, 2].
Neutrophil determination in the stools is inefficient because its brief lifetime means that the sample should be examined within a few hours of its collection . Lactoferrin is an iron-binding glycoprotein found concentrated in the secondary granules of the neutrophils . Granule proteins can be released from living cells , while cell death does not appear to increase this phenomenon. Calprotectin is a cytoplasmic antimicrobial component prominent in granulocytes, monocytes, and macrophages. It accounts for approximately 60% of the total protein of the cytosol. The release of calprotectin is most likely a consequence of cell disruption and death . Both lactoferrin  and calprotectin  are stable in stools for more than 7 days at different temperatures and resist proteolysis even after transportation and storage.
Calprotectin and lactoferrin can inhibit bacterial proliferation, the first as a component of the innate immune response and the second through its iron-binding capacity. High fecal neutrophil levels are detected in inflammatory bowel disease patients  and may be used as surrogate markers of active disease instead of repeated endoscopic examinations . The development of new markers of intestinal inflammation would be of great value to the clinician for discriminating between patients with organic vs functional processes, for monitoring inflammatory disease activity, and for measuring the effects of treatment. Both lactoferrin and calprotectin have been found to correlate with colorectal and intestinal inflammation in two studies involving patients undergoing colonoscopy for surveillance or gastrointestinal symptoms [11, 12]. Comparative studies with multiple fecal leukocyte proteins in inflammatory bowel disease and irritable bowel syndrome found that polymorphonuclear elastase and calprotectin, but not lactoferrin, correlated with endoscopically classified severity of inflammation . Both calprotectin and lactoferrin were considered useful as markers of clinical disease activity in patients with ulcerative colitis . Moreover, both markers significantly increased after acute radiation proctitis , but the study did not compare the findings with symptom scores, biopsies, and blood tests.
The aim of our study was to compare the performance of the two fecal markers, the calprotectin and the lactoferrin, in detecting organic disease in patients undergoing colonoscopy for gastrointestinal symptoms or active disease in inflammatory bowel disease patients, by measuring their concentration in a spot stool sample.
Materials and methods
Characteristics of patients
Lower gastrointestinal symptoms
Altered bowel habits
Occult blood/rectal bleeding
Inflammatory bowel disease
Seventy-seven patients had inflammatory bowel disease (group 2), 46 had ulcerative colitis, and 31 had Crohn’s disease (Table 1).
Referral for colonoscopy in ulcerative colitis patients was for active disease (19 patients) or surveillance (27 patients). Endoscopic disease activity was graded as 0–3 (0 = normal mucosa, 1 = erythema, mild alterations in the vascular pattern, 2 = marked erythema, absent vascular pattern, friability, and erosions, 3 = spontaneous bleeding and ulcers), based on the most inflamed segment of the colon . During colonoscopy, at least two biopsies were taken from each of the segments explored. At histological inflammation was graded as 0–3 (0 = normal mucosa, 1 = chronic inflammatory infiltrate of the lamina propria with mild/no architectural disorder, 2 = acute inflammatory infiltrate with mucosal injury and crypt abscesses, 3 = ulcerations, mucosal injury, crypt abscesses) . Twenty-six patients had extensive colitis and 20 had left-sided disease. Four patients were on immunosuppressants while three on steroids. Only two patients were not given mesalamine due to drug intolerance.
Referral for colonoscopy in Crohn’s disease patients was for active disease (18 patients), surveillance for long-standing colonic disease (six patients), and assessment of disease extent (seven patients). Endoscopic disease activity was graded as 0–3: 0 = normal mucosa; 1 = aphthous ulcers, ulcerated surface area <10%, affected surface area <50%, single narrowings, at least one segment affected by an attenuated vascular pattern, mucosal bridging, pseudopolyps; 2 = large ulcers, ulcerated surface area 10–20%, affected surface area 50–75%, multiple narrowings, loss of vascular pattern, stricture formation; 3 = very large ulcers, ulcerated surface area <30%, affected surface area >75%, deep irregular-shaped ulcerations, longitudinal ulcerations, cobblestones, discontinuous involvement, luminal narrowing, fistulas . Specific histological findings were graded 0–3: 0 = normal mucosa; 1 = chronic inflammatory infiltrate with mild/no architectural disorder; 2 = acute inflammatory infiltrate with lymphocytes, macrophages with transmural injury; 3 = ulcerations, fibrosis, granulomas . Nine patients had terminal ileitis, 13 had colitis, and nine had ileocolitis. None of the patients had had surgery. Five patients were on immunosuppressants while three on steroids. All patients were receiving mesalamine (2.4–3.0 g/day).
Informed written consent was obtained from each patient.
Fecal samples were collected in plastic containers the day before endoscopy and stored at −20°C until assayed.
Aliquots of approximately 100 mg feces (range 40–120 mg) were homogenized for 25 min with extraction buffer. After centrifugation for 20 min at 10,000 rpm, the supernatant was diluted 1:50 with dilution buffer, and calprotectin was measured with an enzyme-linked buffer in immunoadsorbent assay. The ELISA test (Eurospital, Calprest) used polyclonal antibody against calprotectin. A 100-μl sample was added in duplicate to each well, which had been previously coated with antibody. Standards contained calprotectin at known concentrations.
After 50-min incubation at room temperature, washing was followed by the addition of 100 μl of conjugate and, after adding 100 μl substrate and incubation at room temperature for 30 min in a dark place, the optical density (OD) values were read at 405 nm. The best cut-off was chosen by receiver-operating-characteristic (ROC) curve analysis in the group as a whole and in each of the subgroups under study.
Interassay variability was 7% and intraassay variability was 6%.
Lactoferrin was assayed with an immunoadsorbent assay (IBD-CHEK, Techlab). Briefly, an aliquot of approximately 100 mg stool samples, or 50 μl in the case of liquid stools, was diluted 1:400 with a buffered protein solution and transferred to the microtiter well. The microassay plate contained 96 wells coated with purified polyclonal antibody specific for lactoferrin. A positive control consisting of diluted human lactoferrin and a negative control consisting of test diluent were evaluated with each ELISA analysis to facilitate comparisons of results between assays. After 30 min incubation at 37°C, the wells were washed and the antibody conjugate was added. A second series of washings removed any unbound material. OD was read at 450 nm. The best cut-off value was assessed by ROC curve analysis in the group as a whole and in each of the subgroups of patients.
Interassay variability was 10% and intraassay variability was 8%.
Values are expressed as mean±SEMs. Sensitivity, specificity, positive predictive value (PPV), and diagnostic accuracy were calculated for each test. The proportion of patients correctly diagnosed by each test and the relationship with endoscopic and histological findings were evaluated by X2 test. The ROC curves were used to assess the best cut-off for identifying the presence of organic disease or intestinal inflammation. ROC curves were generated by plotting the relationship of the true positive cases (sensitivity) at various cut-off points during the test. An area under the curve (AUC) of 1.0 is characteristic of an ideal test, whereas 0.5 indicates a test of no diagnostic value. The agreement between the two tests was evaluated by the Kappa–Cohen test for ordinal data, which is a chance-adjusted measure of agreement: 1.0 = perfect agreement, 0–0.2 = poor, 0.2–0.4 = fair, 0.4–0.6 = good, >0.6 = excellent. Data analyses were performed using the Statistical Package for the Social Sciences (SPSS) for Windows software rel. 11.5 (SPSS, Chicago, IL, USA). A p-value less than 0.05 was considered statistically significant.
Sensitivity, specificity, positive predictive value (PPV), and diagnostic accuracy of calprotectin and lactoferrin for patients referring for lower gastrointestinal symptoms and for inflammatory bowel disease patients
Diagnostic accuracy (%)
Lower gastrointestinal symptoms
Inflammatory bowel disease
Group 1: lower gastrointestinal symptoms
Final diagnosis in patients with lower gastrointestinal symptoms and number (percentage) of positive calprotectin and lactoferrin tests
Calprotectin positivity N (%)
Lactoferrin positivity N (%)
Polyp <1 cm
Polyp >1 cm
Both calprotectin and lactoferrin positively correlated with intestinal inflammation (p<0.0001) with an excellent agreement between them (k=0.63).
All irritable bowel syndrome patients had normal calprotectin values; lactoferrin was positive in one patient with an OD of 0.041. Both calprotectin and lactoferrin tests were positive in seven of the eight colorectal cancer cases, while both markers were often negative in patients with adenomatous polyps and diverticular disease: only one patient with a large polyp and two with diverticular disease were positive for calprotectin. Lactoferrin was positive in three patients with a large polyp, in one with a small polyp, and in one with diverticular disease. Patients with diverticulitis were positive for both markers, as were both the patients with Crohn’s disease.
Group 2: inflammatory bowel disease activity/surveillance
For the group of inflammatory bowel disease patients as a whole, the test results are summarized in Table 2 and the AUCs are shown in Fig. 2b. Based on the ROC curves, the best cut-off for each test was calculated; it was 0.070 OD for lactoferrin and 80 mg/kg for calprotectin. The agreement between the two markers was fair (k=0.30).
Each of the eight patients negative for calprotectin had mild disease activity at endoscopy. Calprotectin was positive in three patients without active disease at endoscopy and histology. The remaining patients with endoscopic disease remission had normal mean calprotectin levels (32.72±2.71 mg/kg). Lactoferrin was falsely positive in four patients. It was found in one patient with moderate disease and in eight with mild disease at endoscopy. Both endoscopic and histological scores correlated significantly with lactoferrin (r=0.354, p=0.023 and r=0.544, p=0.0001, respectively) (Fig. 3b). Sensitivity, specificity, PPV, and efficacy were 78, 70, 90, and 97% for calprotectin and 75, 60, 87, and 92% for lactoferrin, respectively. The AUC for lactoferrin was 0.782 [95% confidence interval (CI) 0.645–0.919], with an asymptotic significance of 0.007, while the AUC for calprotectin was 0.753 (95% CI 0.599–0.906), with an asymptotic significance of 0.015.
Calprotectin was positive in two of the five patients without active disease at endoscopy, while lactoferrin was positive in one patient. Sensitivity, specificity, PPV, and diagnostic efficacy were 81, 80, 95, and 80% for calprotectin and 77, 80, 95, and 77% for lactoferrin, respectively.
The proportion of patients positive for calprotectin or lactoferrin was not influenced by the extent of disease when a similar disease activity was considered. The AUC for lactoferrin was 0.823 (95% CI 0.642–1.004), with an asymptotic significance of 0.024, while the AUC for calprotectin was 0.815 (95% CI 0.612–1.019), with an asymptotic significance of 0.028.
Identifying gastrointestinal disease or inflammatory causes of intestinal symptoms is crucial to clinical decisions and therapeutic strategies . Several studies have proposed leukocyte-derived protein assay to identify inflammatory vs functional causes of diarrhea [20, 21]. In our study, we tested and compared calprotectin and lactoferrin, the two promising markers of intestinal inflammation because they are stable, they resist degradation, and they are easy to measure. Their diagnostic capacity is also considered superior to many other acute-phase reactant proteins released during inflammation in the gastrointestinal tract . Measuring fecal excretion of 111indium-labeled cells  and TNF-alpha  has been limited to research studies because of the exposure to radiation and prolonged fecal collection. A highly significant correlation was found between the 4-day fecal excretion of white cells and fecal calprotectin concentrations, and the correlation was maintained even when single stool calprotectin concentrations were used .
Among our patients complaining of gastrointestinal symptoms, calprotectin and lactoferrin tests were negative in patients with functional disorders and positive in the two patients with Crohn’s disease and diverticulitis.
Serum and fecal calprotectin levels have been found increased in patients with advanced colon cancer and in neoplastic tissue vs normal mucosa: recruitment of neutrophils to the tumor is probably due to the local production of chemotactic factors [12, 25–28]. The specificity and sensitivity of calprotectin for detecting colorectal cancer were only satisfactory in high-risk populations , however, probably because neutrophils infiltrate the neoplastic tissues as a function of the volume of the neoplasm. Our data are partially consistent with this hypothesis because most of the cancers had positive markers, while calprotectin was often negative even in the presence of large polyps. Lactoferrin was slightly more sensitive than calprotectin in detecting adenomatous colonic polyps.
Summerton et al.  likewise observed high calprotectin levels in colorectal cancer but a lower sensitivity in detecting colonic polyps. While elevated calprotectin and lactoferrin might significantly contribute to characterizing disease activity in patients with known inflammatory bowel disease, they cannot replace endoscopic procedures in patients with no previous diagnosis because inflammatory as well as cancerous processes might be expected.
Fecal calprotectin levels have been found to correlate with disease activity and endoscopic and histological inflammation in patients with inflammatory bowel disease [31, 32] and to predict relapse in patients in clinical remission [33, 34].
The overall percentage of false negative results in our patients with active inflammatory bowel disease was 16% for calprotectin and 19% for lactoferrin, encountered mainly among patients with mild disease activity. A significant correlation was found between calprotectin/lactoferrin levels and endoscopic and histological findings in ulcerative colitis. Contradictory results have been found with both markers in Crohn’s disease: biopsy sampling could be a problem due to the patchy distribution of the disease. Disease activity can also be difficult to measure in Crohn’s disease, and all the indices we use carry varying degrees of subjectivity. Small bowel mucosal lesions were not accurately ruled out and, moreover, may have influenced the fecal markers.
In conclusion, calprotectin and lactoferrin are inexpensive and easily measured and, therefore, suitable for extensive use. Both tests appear to be useful in detecting bowel inflammation in symptomatic patients, achieving a similar diagnostic accuracy. From a clinical perspective, a spot test without any extraction process would be welcomed for daily practice in unselected patients with lower gastrointestinal symptoms.
Different cut-off values are suggested for different patient categories, i.e., higher for patients with known inflammatory conditions while lower for screening purposes. Calprotectin and lactoferrin determination appears to reflect endoscopic and histological disease activity in ulcerative colitis but not in Crohn’s disease. Both tests should be used until the reason for this discrepancy has been unraveled.
Dr. R. D’Incà and Dr. E. Dal Pont contributed equally to this work. This is supported in part by a grant from MIUR ex 40% 2004 (2004063577_004).