Ulcerative colitis (UC) and Crohn's disease (CD) are classified as idiopathic inflammatory bowel diseases (IBD) and could be different manifestations of the same group of disorders. UC and CD can be distinguished from each other in children at the onset of the disease in ≈85% of cases. The remaining 15% are conditionally diagnosed as unclassified enterocolitis (UE) (1).
Antibodies directed against components of the neutrophil cytoplasm [antineutrophil cytoplasmic antibodies (ANCA)] were first found in patients with segmental necrotizing glomerulonephritis (2), and their great specificity for Wegener's granulomatosis was reported in 1985 (3). The nomenclature and a standard indirect immunofluorescent assay using ethanol-fixed leukocytes were established by international workshops on ANCA (4,5). Many studies have demonstrated that ANCA are present in the sera of 60-70% of adult patients with UC and in 5-20% of those with CD (6-9). Two reports on the situation in children are contradictory. One suggests that the presence of perinuclear ANCA (p-ANCA) does not definitely distinguish pediatric patients with UC from those with CD (10); the other states that ANCA are present in 83% of children with UC and that the combination of a positive ANCA and a perinuclear indirect immunofluorescence pattern is 97% specific for UC (11).
The present study was therefore undertaken to evaluate the specificity of ANCA in children with IBD and to determine the value of this test for the differential diagnosis of UC, CD, and UE. Patients with UC were also examined to determipne the relationship between the presence or absence of these antibodies and clinical parameters such as disease activity, extent of lesions, and medical or surgical treatment.
Population and Serum Samples
The population studied consisted of 102 patients with IBD recruited from the Robert Debré Hospital in Paris (n = 75) and Purpan Hospital in Toulouse (n = 27). A total of 26 first-degree relatives with no history of IBD and 20 children with no evidence of chronic inflammatory intestinal disease were used as controls, and were investigated in Toulouse.
The diagnosis and severity of IBD were based on conventional clinical, radiological, endoscopic, and histologic criteria (1). Considering the course of disease, patients with UC or CD were assigned to one of three groups (12):
- Quiescent: less than one clinical episode a year without any specific treatment (corticosteroids, immunosuppressive drugs, or nutritional support)
- Mild: clinical symptomatic periods that responded to specific treatment (more than one episode a year) separated by symptom-free period(s) of at least 1 month with or without specific treatment.
- Severe: continuous clinical manifestations despite specific treatment, or corticodependence (>1 mg/kg/day), corticoresistance, prolonged (>1 year) nutritional support (enteral or parenteral nutrition), or surgical procedures.
The clinical characteristics of patients with IBD are summarized in Table 1.
There were 33 patients with UC. The median age at the onset of symptoms was 9.4 ± 4.2 years (range, 1-16.3) and 17 were girls (sex ratio = 0.9). Colitis began before 5 years of age in 7 children (21%), between 5 and 10 years of age in 9 children (27%), and after 10 years of age in 17 children (52%). The disease duration was 3 months-16 years (median duration, 4.6 years). The families of four children (12%) had histories of IBD. The endoscopic patterns at the time of serum sampling were graded as: 0 = no bleeding (n = 5), 1 = bleeding only after contact (n = 18), and 2 = spontaneously bleeding (n = 10) (13). Seven children also had extraintestinal manifestations (arthritis, mouth ulcerations) (21%) and five children had associated diseases: sclerosing cholangitis (n = 3), autoimmune hepatitis (n = 1), and periodic disease (n = 1). Eight children were corticodependent (24%) and six were corticoresistant (18%). Parenteral nutrition was indicated in 13 children (duration 1-7 months); serum samples were collected before (n = 2), during (n = 4), and after (n = 7) supplementary nutrition. Eleven children had total colectomies (33%), together with a proctectomy in eight patients: Seven children were tested for ANCA before and after surgery.
Sixty-four children had CD. Their median age at the onset of symptoms was 10.9 ± 2.1 years (range, 6-15.9 years) and 36 were boys (sex ratio = 1.29). The disease began in 22 of the children before 10 years of age (34%), and in the remaining 42 after 10 years of age (66%). The disease had continued for 7 months-16 years (median duration, 4.8). The families of 14 children (22%) had histories of IBD. Fifty-one children were given nutritional support: enteral feeding (n = 34) or parenteral nutrition (n = 30). ANCA were tested before (n = 10), during (n = 14), and after (n = 27) nutritional therapy. Surgery was indicated for 23 children: of these, 7 had total colectomies, 8 had a distal ileal resection and right colectomy, 4 had a distal ileal and cecal resection, whereas 4 underwent distal ileal resection. Eight serum samples were collected before surgery and 15 after; 5 children were tested before and after surgery.
A total of five children with IBD could not be classified using previous criteria. Their median age at the onset of symptoms was 8.1 ± 5.3 years (range, 1-12.7 years). Colitis began in two of them before 5 years of age and in the other three after 10 years of age. The disease continued for 3-5 years (median duration, 4.5 years). There was no history of familial IBD or associated disease. Enteral feeding was indicated in two children, and parenteral nutrition was used twice in the same children for a total duration of 3 years. Two children underwent proctocolectomy. These children were tested for ANCA before and after nutritional support and surgery.
The 26 first-degree relatives of 11 children with IBD studied were 9 mothers, 5 fathers, and 12 siblings. Of these, 19 had children with CD, 6 had children with UC, and 1 had a child with UE. None of these relatives had IBD themselves.
The controls were 20 children without IBD, and serum samples were collected in routine clinical studies. These patients included five with abdominal pain, four with Schönlein-Henoch purpura, two with constipation, two with rectal bleeding without evidence of colitis, two with gastroesophageal reflux, one with gastroenteritis, one with cow's milk allergy, and three without intestinal disease. All sera were kept at -70°C until thawed for study.
Preparation of Neutrophils
Human neutrophils were prepared from peripheral blood using standard techniques. Briefly, blood from normal, healthy volunteers (group 0) was taken into heparinized containers. Neutrophils were separated by sedimentation with 2.5% Dextran for 30 min at 37°C followed by centrifugation and two washings with phosphate-buffered saline (PBS). Red blood cells were lysed by osmotic shock with NH4Cl (pH 7.4), and the remaining cells were washed twice with PBS. Granulocytes were then smeared on microtiter plates and fixed in 96% ethanol (5 min at 4°C). The slides were dried and stored at 4°C.
Indirect Immunofluorescence Assay
Each serum sample was tested at a dilution of 1/20; all positive sera were diluted from 1/20 to 1/2,000 in PBS, placed in a multiwell fixed neutrophil slide, and incubated for 45 min at 4°C. The slides were then washed and stained with fluorescein-conjugated polyvalent immunoglobulin G (Pasteur Diagnostics ref. 74521). The slides were washed, dried, covered with a coverslip, and examined under a fluorescence microscope by the same observer in each center. The observers were unaware of the diagnosis or the course of the disease. The fluorescence patterns were perinuclear (p-ANCA) or cytoplasmic (c-ANCA) (Fig. 1). A titer of 1/20 was considered positive. Seventeen sera from this study, together with positive and negative control sera, were verified by comparison between the two laboratories. There was complete agreement in all the cases. Positive control sera were obtained from adult patients with Wegener's disease. The intra- and interobserver variability of the assessments did not exceeded one dilution.
Student's t-test was used for parametric data. Nonparametric data were analyzed using the χ2 test. Differences were considered statistically significant at p < 0.05.
All patients were known to the investigators and were cared for in the two centers participating in this study. Informed consent was obtain from the parents or guardian of each patient.
Patients With Inflammatory Bowel Disease
ANCA were present in 24 of 33 patients with UC (73%), in 9 of 64 patients with CD (14%), and in 4 of 5 patients with UE (80%). The positive ANCA responses included perinuclear staining (p-ANCA) in 71% of cases with UC, 57% of CD, and 75% of UE. However, five children (two UC and three CD) had no specific pattern (c-ANCA or p-ANCA). A positive ANCA was 73% sensitive and 86% specific for a diagnosis of UC (p < 0.001), rather than CD, whereas p-ANCA was 46% sensitive and 94% specific for UC and c-ANCA was 25% sensitive and 95% specific. Comparing UC with other IBD (CD + UE), a positive ANCA was 73% sensitive and 81% specific. Children with UC had ANCA titers of 1/20-1/2,000; children with CD had titers of 1/20-1/500; and children with UE had titers of 1/80-1/500 (Fig. 2)
The UC children with sclerosing cholangitis and autoimmune hepatitis were all ANCA-positive. The patients with specific IBD (UC and CD) showed no correlation between a positive ANCA response and any of the various parameters studied (age at onset, sex, family history of IBD, disease duration, course of the disease, disease distribution, or extraintestinal manifestations). There was no correlation between endoscopic grade and ANCA status in UC, or between the presence or absence of perineal lesions in CD.
Patients with UC showed no correlation between the positive ANCA titers and the course of the disease or specific treatment (corticosteroids, azathioprine, or nutritional management), nor was there any correlation between children operated on for colectomy and children managed by medical treatment. There was a statistically significant difference between the ANCA titers of the children with primary sclerosing cholangitis plus UC and those of children with UC alone (p < 0.002).
Serial samples (n = 75) were taken at intervals of 4-15 months from 20 of the patients with UC. The ANCA status (positive or negative) was not modified during this time for each individual treated with drugs or given nutritional support. The seven children operated on for colectomy in UC included five (71%) who remained ANCA-positive with titers of 1/200-1/2,000 after a follow-up of 2 months-3 years.
The 26 relatives studied included 4 who were ANCA-positive. All had a characteristic c-ANCA pattern with titers of 1/50 for three and 1/100 for one. Two were siblings of an ANCA-negative child with CD and two were parents of another ANCA-negative child with CD. All the relatives of ANCA-positive children with UC (n = 4), CD (n = 2), and UE (n = 1) were ANCA-negative.
One patient with a gastroesophageal reflux was positive (1/100), with no evidence of inflammatory intestinal disease. Sera from the 19 other control subjects were all negative.
The sera of the majority of children and adolescents with ulcerative colitis contained ANCA (73%), but only a few young patients with CD had ANCA (14%). Only 6% (1 child) of the pediatric control subjects had ANCA. These data are consistent with previous reports for IBD in adults (6-9,14-18) and two pediatric studies (10,11). The majority of our IBD patients had ANCA with a perinuclear pattern, as generally described (6,9,11,17). Seventy-one percent of the ANCA-positive UC patients had a perinuclear fluorescence staining. This confirms the findings in adults, in whom 60-98% of UC patients had sera with positive p-ANCA (15,17), and in children (11). Perinuclear ANCA may be targeted against many antigens including myeloperoxidase, elastase, lactoferrin, or cathepsin (7). The perinuclear pattern seen in IBD, and particularly in UC, could be due to the leakage of nuclear antigens into the cytoplasm or to a cytoplasmic component that is associated with the nucleus (19). The perinuclear staining results from antibody binding to antigens that can be denatured, solubilized, and attracted to the nucleus by electrostatic forces. In contrast, c-ANCA are directed against the 29-kDa proteinase 3 (20). This suggests that different antigenic specificities are involved in those diseases. More sensitive techniques will be required to determine the location and identity of antigens associated with IBDs and their role as genetic markers (15). It will be important to explore these autoantibodies and their antigenic targets at a molecular level (21). The positive ANCA titers of patients with UC were generally higher than those of CD patients, but the difference was not statistically significant. Similar findings have been reported for adults (14,16) and other children (11). The relationship between a positive ANCA and several clinical indicators in the IBD patients indicated that there is no correlation between ANCA status and sex, disease activity, disease distribution, the presence or absence of extraintestinal manifestations, or a family history of IBD. The level of ANCA in the serum during the course of the disease was examined in 20 patients, for whom sequential blood samples were available. The changes in the ANCA titer of individuals did not correlate with their disease course, sedimentation rate, or corticosteroid or immunosuppressive drug treatment. These results are similar to those of Proujansky et al. (10) for a pediatric population. They found that there was no relationship between a positive ANCA value and disease activity or other clinical indicator. Nevertheless, Proujansky et al. indicated that all the ANCA-positive patients with CD had colonic disease. Our results were different, because three of the children with CD were ANCA-positive and had only ileal lesions or ileal lesions plus esophagogastric involvement.
The published data for adult patients show no correlation between ANCA-positive status and disease activity (6,7,9,14,15). However, two studies indicate that patients with active UC have higher ANCA titers than patients with inactive disease (p < 0.002  and p < 0.05 ). Disease activity is generally graded in these studies (8,10,17) using the score described by Truelove and Witts (22), based on parameters assessed at the time of evaluation, and not during a 1-year period as in our study. The Truelove and Witts classification is less sensitive and specific than colonoscopy for assessing disease activity, particularly in severe attacks of the disease, which are frequently underestimated (23,24). The endoscopic findings lead to the same conclusions as that of Broekroelofs et al. (17), who found no link between the activity score obtained by endoscopy and the presence of ANCA.
Most of our patients (five of seven) with a colectomy were ANCA-positive. The exact prevalence of ANCA in patients with proctocolectomy having had an ileoanal anastomosis is still unknown. Duerr reported that ANCA were present in 17 of 27 (63%) patients who had undergone total colectomy with ileostomy (6). Winter et al. (11) reported that the ANCA binding of sera became normal in three of four children after colectomy and decreased from 52 to 15% (normal < 12%) in the fourth. A Finnish study also found that proctocolectomy decreased the p-ANCA titers in most patients (25). The authors suggested that a reduction in the inflammation or in the available antigenic material altered the immune disturbances of UC.
The four children with UC plus liver disease (three with sclerosing cholangitis and one anti-actin autoimmune hepatitis) were all p-ANCA-positive. Other reports for children and adults indicate that both disorders are autoimmune (26,27).
There were five children who could not definitively be diagnosed as UC or CD from their clinical, endoscopic, and histologic criteria. Of these, four children were ANCA-positive, and three had colitis with rectal involvement, whereas one of three had a perineal fistula and multiple thrombophlebitis. These children were considered to be UC patients and were managed as such. One p-ANCA-positive child had the typical colonic lesions of UC on endoscopy, plus inflammatory nongranulomatous lesions of the jejunum and the ileum. Tests for ANCA did not help differential diagnosis in this case.
Studies of the occurrence of ANCA in familial IBD and unaffected first-degree relatives have provided divergent results (16,28-32). Shanahan et al. (28) and Seibold et al. (16) concluded that ANCA could be a potential marker of genetic susceptibility to UC and suggested the possibility of genetic heterogeneity within this disease. In contrast, two other reports found that none of the relatives of patients with UC had ANCA (29,30). Lee et al. (31) found that only 2.6% of relatives were ANCA-positive, and Monteleone et al. (32) found that 6% were ANCA-positive.
This study revealed that 15% of the relatives were ANCA-positive. Although they were related to ANCA-negative children with CD, the number tested is small compared with the largest studies (30,32). It therefore seems unlikely that ANCA is a useful marker of susceptibility to UC and that their presence is more likely to be a consequence of the disease, as indicated by Yang et al. (33) for 24 monozygotic twins with IBD.
In conclusion, this study confirms that most children with UC have perinuclear antineutrophil cytoplasmic antibodies. Because the presence of p-ANCA does not reflect the disease severity or activity, serial assessment is not necessary in UC. Measurement of p-ANCA may be a useful clinical tool for classifying patients with indeterminate colitis or unusual symptoms. However, it cannot be used to separate UC from cryptogenetic enterocolitis with extension to other parts of the gut.
1. Chapoy P, Di Costanzo J. Chronic inflammatory bowel diseases in children
. In: Buts JP, Sokal EM, eds. Management of digestive and liver disorders in infants and children
. Amsterdam: Elsevier, 1993:381-401.
2. Davies DJ, Moran JE, Nial JF, Ryan GB. Segmental necrotizing glomerulonephritis with antineutrophil antibody: possible arbovirus aetiology? Br Med J
3. Van Der Woude FJ, Rasmussen N, Lobatto S, et al. Autoantibodies
against neutrophils and monocytes: tool for diagnosis and marker of disease activity in Wegener's granulomatosis. Lancet
4. Wiik A. Delineation for a standard procedure for indirect immunofluorescence detection of ANCA. APMIS
5. Mathieson PW. Proceedings of the 5th International ANCA Worshop. Clin Exp Immunol
6. Duerr RH, Targan SR, Landers CJ, Sutherland LR, Shanahan F. Anti-neutrophil cytoplasmic antibodies in ulcerative colitis
: comparisons with other colitides/diarrheal illnesses. Gastroenterology
7. Colombel JF, Reumaux D, Duthilleul P, et al. Antineutrophil cytoplasmic antibodies in inflammatory bowel diseases. Gastroenterol Clin Biol
8. Seibold F, Weber P, Klein R, Berg PA, Wiedmann KH. Clinical significance of antibodies against neutrophils in patients with inflammatory bowel disease and primary sclerosing cholangitis. Gut
9. Oudkerk Pool M, Ellerbroek AM, Ridwan BU, et al. Serum anti-neutrophil cytoplasmic autoantibodies
in inflammatory bowel disease are mainly associated with ulcerative colitis
: a correlation study between perinuclear antineutrophil cytoplasmic antibodies and clinical parameters, medical, and surgical treatment. Gut
10. Proujansky R, Fawcett PT, Gibney KM, Treem WR, Hyams JS. Examination of anti-neutrophil cytoplasmic antibodies in childhood inflammatory bowel disease. J Pediatr Gastroenterol Nutr
11. Winter HS, Lander CJ, Winkelstein A, Vidrich A, Targan SR. Anti-neutrophil cytoplasmic antibodies in children
with ulcerative colitis
. J Pediatr
12. Hendriksen C, Kreiner S, Binder V. Long term prognosis in ulcerative colitis
-based on results from a regional patient group from the county of Copenhagen. Gut
13. Powell-Tuck J, Day DW, Buckell NA, Wadworth J, Lennard-Jones JE. Correlations between defined sigmoidoscopic appearances and other measures of disease activity in ulcerative colitis
. Dig Dis Sci
14. Saxon A, Shanahan F, Landers CJ, Ganz T, Targan S. A distinct subset of antineutrophil cytoplasmic antibodies is associated with inflammatory bowel disease. J Allergy Clin Immunol
15. Cambridge G, Rampton DS, Stevens TRJ, McCarthy DA, Kamm M, Leaker B. Antineutrophil antibodies in inflammatory bowel disease: prevalence and diagnostic role. Gut
16. Seibold F, Slametschka D, Gregor M, Weber P. Neutrophils autoantibodies
: a genetic marker in primary sclerosing cholangitis and ulcerative colitis
17. Broekroelofs J, Mulder AHL, Nelis GF, Westerveld BD, Tohen Terraert JW, Kallenberg CGM. Anticytoplasmic antibodies (ANCA) in sera from patients with inflammatory bowel disease (IBD): relation to disease pattern and disease activity. Dig Dis Sci
18. Cellier C, Chaussade S, Noel LH, et al. Perinuclear antineutrophil cytoplasm antibodies: frequency in ulcerative colitis
, Crohn' disease and others GI tract disorders. Gastroenterology
19. Gross WL, Schmitt WH, Csernok E. ANCA and associated diseases: immunodiagnostic and pathogenic aspects. Clin Exp Immunol
20. Frischman W, Couper R. Antineutrophil cytoplasmic antibodies in inflammatory bowel disease [Selected summary]. J Pediatr Gastroenterol Nutr
21. Shanahan F. Neutrophil autoantibodies
in inflammatory bowel disease: Are they important? Gastroenterology
22. Truelove SC, Witts LJ. Cortisone in ulcerative colitis
: final report on a therapeutic trial. Br Med J
23. Gomes P, Duboulay C, Smith CL, Holdstock G. Relationship between disease activity indices and colonoscopic findings in patients with colonic inflammatory bowel disease. Gut
24. Carbonnel F, Lavergne A, Lemann M, et al. Value of colonoscopy in assessment of severity of lesions in attacks of ulcerative colitis
25. Aitola P, Miettinen A, Matilla M, Soppi E. The effect of proctocolectomy on serum antineutrophil cytoplasmic antibodies in patients with ulcerative colitis
26. Lo SK, Chapman RWG, Cheeseman P, et al. Antineutrophil antibody: a test for immune primary cholangitis in childhood? Gut
27. Targan SR, Landers C, Vidrich A, Czaja AJ. High-titer antineutrophil cytoplasmic antibodies in type-1 autoimmune hepatitis. Gastroenterology
28. Shanahan F, Duerr RH, Rotter JI, et al. Neutrophil autoantibodies
in ulcerative colitis
: familial aggregation and genetic heterogeneity. Gastroenterology
29. Reumaux D, Delecourt L, Colombel JF, Noël LH, Duthilleul P, Cortot A. Anti-neutrophil cytoplasmic autoantibodies
in relatives of patients with ulcerative colitis
30. Bansi DS, Lo S, Chapman RW, Fleming KA. Absence of antineutrophil cytoplasmic antibodies in relatives of UK patients with primary sclerosing cholangitis and ulcerative colitis
. Eur J Gastroenterol Hepatol
31. Lee JCW, Lennard-Jones JE, Cambridge G. Antineutrophil antibodies in familial inflammatory bowel disease. Gastroenterology
32. Monteleone G, Doldo P, Marasco R. Perinuclear neutrophil antibodies (p-ANCA) in unaffected relatives of patients with ulcerative colitis
(UC): suggestions against familial aggregation. Gut
33. Yang P, Järnerot G, Danielsson D, Tysk C, Lindberg E. P-ANCA in monozygotic twins with inflammatory bowel disease. Gut