The recent survey by the British Paediatric Surveillance Unit (1) suggests that the annual incidence of inflammatory bowel disease (IBD) in children under the age of 16 years is 5.3/100,000 children. This translates to approximately 700 new cases per annum in the United Kingdom and Republic of Ireland, and this seems to be increasing (2).
The definitive diagnosis of IBD relies on several invasive investigations, including blood tests, radiological investigations, and endoscopy (2). Blood inflammatory indices have shown a weak correlation with colonoscopic and histological results in adults and children (3,4), and radiological investigations involve significant radiation doses (5).
Because the early manifestations of pediatric IBD can be relatively nonspecific, there is a need for a reliable noninvasive test to help in the diagnosis of IBD. Several groups have attempted to develop fecal markers, such as neutrophil elastase (6–8), tumor necrosis factor-α (9,10), and eosinophilic cationic protein (11,12). Although all of these have been shown to be excreted in increased amounts in the feces of patients with IBD, and concentrations are higher in active than in inactive disease, their instability in feces precludes routine clinical use.
Recent studies have suggested that the measurement of fecal calprotectin (FC) and fecal lactoferrin (FL) may help in the assessment and monitoring of patients with IBD. Both FC and FL are neutrophil proteins and are resistant to proteolysis and unaffected by multiple freeze-thaws, providing useful markers in feces as indicators of inflammation (13). Calprotectin belongs to a group of Ca2+-binding proteins in the S100 family (14). It is released from the cytoplasm of activated neutrophils, where it accounts for 60% of the protein in the cytosol (15). FC has been shown to be valuable in the diagnosis and management of IBD in adults (16–18). To date, there have been relatively few studies on the use of FC in children with IBD. Bunn et al (5,19) have suggested that FC correlates closely with disease activity and that FC concentration is higher in pediatric patients with IBD than in control individuals. More recently, other studies have also shown the value of this novel marker in children with gastrointestinal (GI) problems (20–23) and IBD (24–26).
Fecal lactoferrin is secreted by most mucosal membranes and is a major component of the secondary granules of polymorphonuclear leukocytes. During intestinal inflammation, leukocytes infiltrate the mucosa, resulting in an increase in the FL concentration in the feces (27,28). FL has also been evaluated for monitoring disease activity in IBD and pouchitis (29–31) and for diagnosing the inflammatory causes of diarrhea in adult patients (32). There are few data from children, but recent studies have suggested that FL may be useful in assessing disease activity in children and young adults with IBD (33) and may be a helpful noninvasive tool for monitoring therapeutic efficiency in severe Crohn disease (CD) (34).
Both FC (35) and FL (29,36) are stable in stool for as long as 7 and 5 days, respectively, allowing samples to be sent to the laboratory by mail. Both FC and FL are directly associated with the inflammatory process, are easy to measure, and therefore may have the potential to act as noninvasive markers for active IBD in children and hence reduce invasive investigations. We therefore prospectively evaluated whether FC and FL could be used as noninvasive markers in children with active IBD.
PATIENTS AND METHODS
Patients and Fecal Collection
Children and young people (4–17 years) were consecutively recruited from the Paediatric Gastroenterology and General Paediatric Unit. There were 3 groups:
- The IBD group was composed of those with active IBD (both new diagnoses and known cases in relapse). The diagnosis of active IBD was made on the basis of clinical features and the usual diagnostic techniques, including barium study, leukocyte scan, endoscopy, and biopsy.
- The GI control children were for the most part referred with recurrent abdominal pain, constipation, intermittent diarrhea, or alternating diarrhea and constipation. Three children had intermittent rectal bleeding, and several underwent endoscopy and biopsy and other appropriate investigations.
- The non-GI control children were those with no GI disease. This group comprised preoperative patients (undergoing non-GI surgery) and some general pediatric patients with asthma, urinary tract infection, and upper respiratory tract infections.
A single stool sample was collected from patients whose parents consented to their participation in the study. A portion of stool was frozen at −20°C before analysis of calprotectin and lactoferrin by enzyme-linked immunoassay methods (29,35,37).
Analysis of FC and FL
The FC and FL were measured by enzyme-linked immunoassay. FC was analyzed by PHICAL kits obtained from CALPRO AS, Norway, and FL by IBD-Scan kits obtained from TechLab, Blacksburg, VA, and supplied by Bio Connections, Leeds, UK. All materials were supplied with the kit. The clinician was blind to the FC and FL results until the end of the study, and all of the laboratory measurements were performed by the biochemist without knowledge of the clinical details and the results of other investigations.
The South East Wales Local Research Ethics Committee approved this study. Verbal and written explanation of the study was given to parents and children, and written informed consent was obtained from the parents.
Stratum-specific analysis and a combined analysis were conducted. The combined analysis compared individuals with IBD with all control individuals, and the stratum-specific analysis compared individuals with IBD with GI control individuals, then with non-GI control individuals.
Owing to the inherent skewness within the distribution, the Wilcoxon rank sum procedure was used to detect differences in biomarker distributions.
The effect of FC and FL and the interaction on the odds of IBD were investigated by use of logistic regression, with robust standard errors on loge transformed data to indicate possible explanatory variables for use in the diagnostic decision making process. Logistic regression models were built sequentially by use of Wald tests as tests of significance.
Receiver operator characteristic (ROC) curves enabled easy identification of the most efficient diagnostic tests and the optimum cutpoints, producing the greatest sensitivity and specificity for a given value. The point on the ROC curve that maximizes sensitivity and specificity was used to define the cutpoints. The area under the ROC curve (AUC) indicated the probability that the test result from a randomly selected pair was correctly assigned to the correct diagnosis, with the greater AUC signifying the better diagnostic test. All statistical analyses were performed by use of STATA 9.1
The number of patients for the FC assay was IBD = 26 children, GI control children = 30, and non-GI control children = 25. For the FL assay the number was IBD = 24, GI control children = 26, and non-GI control children = 24. The number of patients in the FL assay was slightly smaller in view of the nonavailability of FL assay kits at the outset. The number of boys in each group was IBD = 20, GI control boys = 14, and non-GI control boys = 14.
The mean age of the active IBD group was 13.23 years, whereas mean ages of the GI and non-GI control groups were 10.93 and 11.4, respectively. There was some indication that the GI control children were slightly younger than the IBD children (2.29 years t = −2.41, P = 0.018); otherwise, there was no evidence of differences between the other groups.
The patients with IBD comprised 17 with CD (large intestine = 4, small intestine = 3, large and small intestine = 10) and 9 with ulcerative colitis (pancolitis = 5, left-sided = 2, proctitis = 2).
Comparison of Medians
The median concentrations of FC and FL were plotted and compared (Fig. 1).
The median FC concentration was 251 μg/g of stool in the active IBD group, compared with 35.4 in the GI control group and 14.1 in the non-GI control group. FC concentrations in patients with active IBD were significantly elevated compared with both the control groups (z = 5.898 and z = 5.274, respectively, P < 0.001).
The median FL concentration was 21 μg/g in the active IBD group, compared with 0.305 in the GI control group and 0.11 in the non-GI control group. The FL concentration in patients with active IBD was significantly elevated compared with both the control groups (z = 5.225 and z = 4.624, respectively, P < 0.001).
Interaction of FC and FL
The median value for the product of FC and FL in the IBD group was 5622, whereas the values for the GI control and non-GI control groups were 10.92 and 0.73, respectively. Thus, the concentration of product of these 2 biomarkers was significantly elevated in the IBD group compared with the 2 control groups (z = 5.862 and z = 5.029, respectively, P < 0.001).
Singularly logistic regression modeling revealed that the odds of a diagnosis of IBD increased by 70% (z = 4.19, P < 0.0001) for a doubling of FC concentration, and 27% (z = 3.66, P < 0.0001) for a doubling of FL concentration. However, the relation between FC and IBD was positively confounded by the relation between FL and IBD, changing the odds of a diagnosis for a doubling of FC and FL to 53% (z = 2.46, P = 0.014) and 19% (z = 2.49, P = 0.013), respectively. In addition, there was a nontrivial interaction between the 2 biomarkers (z = 1.84, P = 0.066), suggesting that the odds of a diagnosis of IBD was better when the interaction between the 2 biomarkers was used. The interaction bordered on the nominal level of significance of P = 0.05. However, further evidence of the importance of the interaction can be crudely observed by inspection of the correlation coefficients of FC and FL in each stratum (IBD rp = 0.4909, P = 0.0149; GI control children rp = −0.73, P = 0.72; non-GI control children rp = −0.788, P = 0.7144), further suggesting the importance of the interaction between FC and FL as a diagnostic indicator.
Cutpoint and ROC Curves
The cutpoints for each analysis are shown in Table 1. It is clear that using any of the cutpoints suggested will yield a test better than chance for the diagnosis of IBD. From the AUC in Table 1 and visual inspection of the plots (Figs. 2 and 3), it is apparent that the interaction of the 2 biomarkers is consistently higher; however, it is not clear whether this elevation is statistically meaningful or clinically important. Table 2 compares the AUC statistic of each division. Whereas no single comparison reaches the nominal threshold of P = 0.05, it is clear that the interaction between FC and FL is superior to FL alone when compared with the combined control group; however, this difference is much less marked in the stratum-specific control groups.
Children with abdominal pain and a change in bowel habit constitute a major proportion of visits to pediatric gastroenterologists, and it would be useful to find a reliable noninvasive test to detect those with possible IBD who will need further invasive investigations.
Pediatric and adult studies have shown a weak correlation between hematological inflammatory indices and bowel inflammation assessed by endoscopy (3,4). The measurement of levels of FC (16–18) and FL (13,27,32) may help in the assessment and monitoring of adult patients with IBD. Previous studies proposed that the suggested cutoff level of FC for adults (>50 μg/g) could be used in children 4 to 17 years old regardless of sex (38). The same group also found no significant difference in the concentration of FC between children with upper respiratory tract infection or tonsillitis and a healthy group. However, recent studies have demonstrated that FC concentration may also be higher in young infants and children up to the age of 2 years (39,40), and FC may be able to differentiate between functional abdominal pain and IBD in school-age children (40). The concentration of FC seemed to be related to disease activity (19), and FC levels declined in line with the clinical improvement in children with active IBD treated with glucocorticoids (26). Increased FC strongly predicted colorectal inflammation in children with gastrointestinal symptoms (20–23), and FC correlated with the best objective invasive measurement of bowel inflammation in children with colonic and small bowel IBD (5).
Adult patients with IBD had a significantly higher concentration of FL than did those with irritable bowel syndrome, regardless of disease activity (13). FL was 90% sensitive in detecting inflammation in IBD and had a negative predictive value of 99% (32). A recent pediatric study suggested that the upper limit of normal for FL level is 7.25 μg/mL (33) and that FL is a sensitive and specific marker of inflammation in children with IBD, with the level correlating well with both clinical disease activity indices and erythrocyte sedimentation rate. The FL levels significantly decreased to near baseline in parallel with the clinical assessment and the Pediatric Crohn Disease Activity Index after infliximab therapy (34,41).
Our study confirms that both FC and FL are significantly elevated in children with active IBD. Any subtype differences between ulcerative colitis and CD are difficult to comment on because of the number of individuals in each stratum. Whereas the use of FC and FL both had respectable operating characteristics, the interaction between FC and FL gave consistently higher or equal sensitivities in comparison with using 1 biomarker alone, in addition to the specificities being always increased. In a comparison of previously proposed cutpoints for FC (>50 μg/g) and FL (>4 μg/g) and the interaction of FC and FL (200), the results change as expected. The sensitivity is always greater in the FC and FC × FL interaction, whereas the specificity is reduced. However, the suggested threshold of FL >4 μg/g is close to our optimal suggested of 4.6 μg/g, and thus the magnitude of any difference is much reduced.
We believe that the lack of statistical significance in comparing the AUC of the diagnostic tests is due to the small number of individuals in each group. However, when coupled with the results from the logistic regression model, visual inspection of the distribution of the data, and the AUC being greater in each analysis, we believe this provides a strong indication for the interaction between FC and FL being important in enabling clinicians to distinguish between individuals with and without IBD. It seems that in each control group the biomarkers act independently with no evidence of association, whereas in the IBD group they are associated with disease status (rp = 0.491, P = 0.0149). This dependence enables the interaction of the 2 biomarkers to act as an ideal indicator of disease status. The results from the ROC curves are encouraging, inasmuch as FL has been identified as a possible biomarker that can be used as part of a diagnostic test, and more important, in conjunction with FC. The box plots illustrate effectively that the product of the 2 biomarkers interquartile ranges of patients with IBD are completely separated, with a large gap from the upper whiskers of the control groups.
In our study, a few individuals in the control groups had elevated levels of FC and FL. One had a colonic polyp, 1 had cystic fibrosis with distal intestinal obstruction syndrome, 1 had unexplained rectal bleeding, another had appendicitis, and 2 had infectious diarrhea. These positive results could be explained by the fact that both FC and FL are highly sensitive but not disease-specific markers, and they detect inflammation throughout the GI tract in children with different GI illnesses.
Both FC and FL are noninvasive, simple, and, more important, highly sensitive tests. These tests would enable health professionals, including general practitioners, to identify high-risk individuals to be fast-tracked to endoscopy, therefore increasing the speed of a definitive diagnosis and enabling faster treatment. It is important to remember that although FC and FL assays are sensitive tests of intestinal inflammation in IBD, they are not specific, inasmuch as any cause of increased intestinal neutrophils will result in increased FC and FL concentrations. Therefore, FC and FL cannot replace invasive tests such as endoscopy, which will always be necessary for definitive tissue diagnosis, but they could be useful in avoiding unnecessary invasive investigations in many individuals.
At the time of writing, this is the first published study, as far as we are aware, to determine the efficacy of FL in combination with FC as noninvasive markers of IBD in children. We conclude that our study supports the previous data that FC and FL are significantly elevated in children with active IBD. Furthermore, we have shown that the interaction between FC and FL could be used to define the best initial diagnostic test, in comparison with their use in isolation, thus reducing the number of endoscopic procedures undertaken, but the clinician needs to weigh the costs and benefits of performing 2 laboratory tests rather than 1 test. These tests are safe, simple, and child friendly, and they have the potential to be used as noninvasive markers of intestinal inflammation in children with active IBD.
1. Sawczenko A, Sandhu BK, Logan RF, et al
. Prospective survey of childhood inflammatory bowel disease
in the British Isles. Lancet 2001; 357:1093–1094.
2. Jenkins HR. Inflammatory bowel disease
. Arch Dis Child 2001; 85:435–437.
3. 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 of children and adolescents with chronic inflammatory bowel disease
. Acta Pediatr Scand 1990; 79:527–534.
4. Beattie RM, Nicholls SW, Domizio P, et al
. Endoscopic assessment of the colonic response to corticosteroids in children with ulcerative colitis. J Pediatr Gastroenterol Nutr 1996; 22:373–379.
5. Bunn SK, Bisset WM, Main MJ, et al
. Fecal calprotectin
: validation of a noninvasive measure of bowel inflammation in childhood inflammatory bowel disease
. J Pediatr Gastroenterol Nutr 2001; 33:14–22.
6. Andus T, Gross V, Caesar I, et al
. PMN-elastase in assessment of patients with inflammatory bowel disease
. Dig Dis Sci 1993; 38:1638–1644.
7. Adeyemi EO, Neumann S, Chadwick VS, et al
. Circulating human leucocyte elastase in patients with inflammatory bowel disease
. Gut 1985; 26:1306–1311.
8. Bohe M, Genell S, Ohlsson K. Protease inhibitors in plasma and faecal extracts from patients with active inflammatory bowel disease
. Scand J Gastroenterol 1986; 21:598–604.
9. Braegger CP, Nicholls S, Murch SH, et al
. Tumour necrosis factor alpha in stool as a marker of intestinal inflammation. Lancet 1992; 339:89–91.
10. Nicholls S, Stephens S, Braegger CP, et al
. Cytokines in stools of children with inflammatory bowel disease
or infective diarrhea. J Clin Pathol 1993; 46:757–760.
11. Berstad A, Borkje B, Reidel B, et al
. Increased fecal eosinophil cationic protein in inflammatory bowel disease
. Hepatogastroenterology 1993; 40:276–278.
12. Bischoff SC, Grabowsky J, Manns MP. Quantification of inflammatory mediators in stool samples of patients with inflammatory bowel disorders and controls. Dig Dis Svi 1997; 42:394–403.
13. Kane SV, Sandborn WJ, Rufo PA, et al
. Fecal lactoferrin
is a sensitive and specific marker in identifying intestinal inflammation. Am J Gastroenterol 2003; 98:1309–1314.
14. Dale I, Fagerhol MK, Naesgaard I. Purification and partial characterisation of a highly immunogenic human leucocyte protein, the L1 antigen. Eur J Biochem 1983; 134:1–6.
15. Gaya DR, Lyon TDB, Duncan A, et al
. Faecal calprotectin
in the assessment of Crohn's disease activity. Q J Med 2005; 98:435–441.
16. Tibble J, Sigthorsson G, Bridger S, et al
. Surrogate markers of intestinal inflammation are predictive of relapse in patients with inflammatory bowel disease
. Gastroenterology 2000; 119:15–22.
17. Tibble J, Teahon K, Thjodleifsson B, et al
. A simple method for assessing intestinal inflammation in Crohn's disease. Gut 2000; 47:506–513.
18. Fagerhol MK. Calprotectin
, a faecal marker of organic gastrointestinal abnormality. Lancet 2000; 356:1783–1784.
19. Bunn SK, Bisset WM, Main MJ, et al
. Fecal calprotectin
as a measure of disease activity in childhood inflammatory bowel disease
. J Pediatr Gastroenterol Nut 2001; 32:171–177.
20. Fagerberg UL, Loof L, Myrdal U, et al
. Colorectal inflammation is well predicted by fecal calprotectin
in children with gastrointestinal symptoms. J Pediatr Gastroenterol Nutr 2005; 40:450–455.
21. Bremner A, Roked S, Robinson R, et al
. Faecal calprotectin
in children with chronic gastrointestinal symptoms. Acta Pediatr 2005; 94:1855–1858.
22. Berni Canani R, Rapacciuolo L, Romano MT, et al
. Diagnostic value of faecal calprotectin
in paediatric gastroenterology clinical practice. Dig Liver Dis 2004; 36:467–470.
23. Carroccio A, Iacono G, Cottone M, et al
. Diagnostic accuracy of fecal calprotectin
assay in distinguishing organic causes of chronic diarrhea from irritable bowel syndrome: a prospective study in adults and children. Clin Chem 2003; 49:861–867.
24. Canani RB, de Horatio LT, Terrin G, et al
. Combined use of noninvasive tests is useful in the initial diagnostic approach to a child with suspected inflammatory bowel disease
. J Pediatr Gastroenterol Nutr 2006; 42:9–15.
25. de Jong NS, Leach ST, Day AS, et al
. Fecal S100A12: a novel noninvasive marker in children with Crohn's disease. Inflamm Bowel Dis 2006; 12:566–572.
26. Kolho KL, Raivio T, Lindahl H, et al
. Fecal calprotectin
remains high during glucocorticoid therapy in children with inflammatory bowel disease
. Scand J Gastroenterol 2006; 41:720–725.
27. Baveye S, Elass E, Mazurier J, et al
: a multifunctional glycoprotein involved in the modulation of the inflammatory response. Clin Chem Lab Med 1999; 37:281–286.
28. Vaishnavi C, Bhasin DK, Singh K. Fecal lactoferrin
assay as a cost effective tool for intestinal inflammation. Am J Gastroenterol 2000; 95:3002–3003.
29. Sugi K, Saitoh O, Hirata I, et al
. Fecal lactoferrin
as a marker for disease activity in inflammatory bowel disease
: comparison with other neutrophil derived proteins. Am J Gastroenterol 1996; 91:927–934.
30. Kayazawa M, Saitoh O, Kojima K, et al
in whole gut lavage fluid as a marker for disease activity in inflammatory bowel disease
: comparison with other neutrophil derived proteins. Am J Gastroenterol 2002; 97:360–369.
31. Parsi MA, Shen B, Achkar JP, et al
. Fecal lactoferrin
for diagnosis of symptomatic patients with ileal pouch–anal anastomosis. Gastroenterology 2004; 126:1280–1286.
32. Fine KD, Ogunji F, George J, et al
. Utility of a rapid fecal latex agglutination test detecting neutrophil protein, lactoferrin
for diagnosing inflammatory causes of chronic diarrhea. Am J Gastroenterol 1998; 93:1300–1305.
33. Walker TR, Land ML, Kartashov A, et al
. Fecal lactoferrin
is a sensitive and specific marker of disease activity in children and young adults with inflammatory bowel disease
. J Pediatr Gastroenterol Nutr 2007; 44:414–422.
34. Buderus S, Boone J, Lyerly D, et al
. Fecal lactoferrin
: a new parameter to monitor infliximab therapy. Dig Dis Sci 2004; 49:1036–1039.
35. Roseth AG, Fagerhol MK, Aadland E, et al
. Assessment of the neutrophil dominating protein calprotectin
in feces: a methodologic study. Scand J Gastroenterol 1992; 27:793–798.
36. Boone J. Fecal lactoferrin
stability. In: Rufo MPA (editor). Boston: 2004.
37. Ton H, Brandsnes O, Dale S, et al
. Improved assay for fecal calprotectin
. Clin Chem Acta 2000; 292:41–54.
38. Fagerberg UL, Loof L, Merzoug RD, et al
. Fecal calprotectin
levels in healthy children studied with an improved assay. J Pediatr Gastroenterol Nutr 2003; 37:468–472.
39. Olafsdottir E, Aksnes L, Fluge G, et al
. Faecal calprotectin
levels in infants with infantile colic, healthy infants, children with inflammatory disease, children with recurrent abdominal pain and healthy children. Acta Paediatr 2002; 91:45–50.
40. Rugtveit J, Fagerhol MK. Age-dependent variations in fecal calprotectin
concentrations in children. J Pediatr Gastroenterol Nutr 2002; 34:323–325.
41. Buderus S, Lohman N, Lyerly D, et al. Clinical evaluation of IBD-CHEK
test for detecting fecal lactoferrin
as indicator of intestinal inflammation in pediatric patients. Poster presentation at Digestive Disease Week, American Gastroenterology Association, Orlando; May 2003.