Hummel, Thalia Z.*; ten Kate, Fiebo J.W.†; Reitsma, Johannes B.‡; Benninga, Marc A.*; Kindermann, Angelika*
Inflammatory bowel disease (IBD) consists of 2 major clinical entities known as Crohn disease (CD) and ulcerative colitis (UC). In approximately 5% to 30% of children, differentiation between CD and UC is not possible and a diagnosis of indeterminate colitis (IC) is made (1–5). There is no single diagnostic test as a criterion standard that can reliably distinguish between CD and UC. A definitive diagnosis of the type of IBD is based upon a combination of medical history, endoscopic findings, histological abnormalities, and radiological features. In general, we state that the inflammatory process in UC is limited to the large bowel, whereas CD may occur throughout the entire gastrointestinal tract, including the upper gastrointestinal tract (UGT); therefore, endoscopy of the UGT may play an important role in the diagnostic assessment of CD in children. Several reports have shown that pathological changes may also have been found in the UGT in UC (6–12). Microscopic mucosal lesions have been identified in biopsies from the UGT in 64% to 90% of CD and in 38% to 70% of patients with UC (6,7,10,11). Most of these findings were nonspecific and not helpful in discriminating CD from UC. Even focally enhanced gastritis, as an isolated finding, which is often reported to be suggestive of CD, is neither sensitive nor specific for CD (13). Today, there is still no uniformity on the diagnostic criteria of IBD of the upper gastrointestinal (GI) tract (14). Only the detection of epitheloid granulomas appears to be the histological hallmark of gastric CD; therefore, the question arises whether endoscopy of the UGT is still justified in the diagnostic assessment of childhood IBD, taking into account the higher risk of complications and increased costs (15). The aims of the present study were to establish the fraction of pediatric patients with CD whose diagnosis relies on the detection of granulomatous inflammation in UGT and to detect histopathological changes in the UGT mucosa, besides granulomas, which can distinguish CD from non-CD (UC and non-IBD).
All children suspected of having IBD visiting our department of pediatric gastroenterology between January 2003 and December 2008 were included in the present study. Presenting symptoms were extracted retrospectively from the medical charts. Because of our standard procedure, all patients underwent both UGT endoscopy and ileocolonoscopy. Before performing endoscopy, a bacterial gastrointestinal infection was ruled out with a fecal culture.
All of the patients had at least 1 biopsy specimen taken from each part of the colon (cecum, ascending colon, transverse colon, descending colon, rectum), terminal ileum, duodenum, stomach (antrum and corpus), and esophagus. Biopsies were taken from macroscopically normal mucosa and from inflamed areas. The tissue was formalin fixed, paraffin embedded, and routinely processed and stained with hematoxylin and eosin. Each biopsy specimen was cut in 2 or more levels to increase the chances of detecting histopathological changes. Helicobacter pylori was detected by both histology and culture from antral biopsies. In the present study, biopsies from all of the sites were reassessed by an expert pathologist who was blinded to the clinical condition and other test results. First, a diagnosis based on ileocolonic biopsies was recorded. Second, a new diagnosis was recorded based on the histological diagnosis of the lower GI tract and the presence or absence of granulomatous inflammation in UGT. After all, the final diagnosis was made based on the reference standard procedure, which consisted of endoscopic findings and histopathological interpretation (as described above), imaging studies, and clinical follow-up data and/or repeat endoscopy (Table 1) (7,16). Imaging studies, small bowel follow-through, or contrast-enhanced magnetic resonance imaging (MRI) were performed in all of the patients without a definite diagnosis based on endoscopic findings and histopathological interpretation of UGT and ileocolonic biopsies. Clinical follow-up data were extracted retrospectively from the medical charts.
All of the histopathological changes found in the biopsies from the UGT, besides granulomas, which were a diagnostic marker, were analyzed against the final diagnosis. Focally enhanced gastritis was defined as the presence of at least 1 foveolum/gland surrounded and infiltrated by inflammatory cells. Focal inflammation of the stomach was defined as chronic and/or acute inflammatory infiltrate occurring in only a portion of a biopsy and/or in 1 biopsy but not in another taken from the same area, without infiltration of foveolae or glands. Focal duodenal cryptitis was defined as chronic and/or acute inflammatory infiltrate with at least 1 crypt surrounded and infiltrated by inflammatory cells.
Statistical analysis was performed using SPSS version 18 (SPSS Inc, Chicago, IL). Because of the relatively small sample sizes within groups, we used Fisher exact t tests when comparing categorical variables between groups and nonparametric tests (Kruskal-Wallis) when comparing continuous variables, such as age and duration of disease. The criterion for statistical significance was defined as P < 0.05.
A total of 172 children suspected of IBD were enrolled (boys 54%, mean age 12.3 years, range 1.6–18.1 years). Based on all of the available data, the final diagnosis was CD in 70 patients, UC in 33 patients, IC in 1 patient, and non-IBD in 68 patients. The results of the histological reassessment of ileocolonic biopsies alone, the diagnosis based on all of the biopsies, and the final diagnosis are shown in Figure 1. Because of the small number of only 1 patient with IC, this patient was not included in the statistical analyses, but the patients also did not show histological abnormalities in the UGT. No differences were found between the 3 main groups with regard to age at endoscopy and sex. Both duration of symptoms before presentation and follow-up period were significantly different between groups (P = 0.03 and P < 0.0001, respectively). These data and the presenting symptoms are shown in Table 2.
Diagnostic Yield of UGT Endoscopy
Noncaseating epithelioid granulomas in the UGT were present in 21 (30%) children with CD. Granulomas were found in the esophagus in 3 patients, in the stomach in 19 children (antral, 10 patients; corpus, 5 patients; both, 4 patients), and in the duodenum in 2 children. In all 3 patients with granulomas in the esophagus, granulomas were also found in the colon and the stomach. Nonorganized aggregates of histiocytes and Langerhans giant cells in the UGT were seen in 10 (14%) children with CD.
In 8 children (11%), the diagnosis based on histopathological assessment of the ileocolonic biopsies alone (2 UC, 4 IC, 2 non-IBD) was changed to CD after assessing the biopsies from the UGT (Fig. 1). In these children, a granulomatous inflammation was found in UGT consisting of granulomas (n = 5) and nonorganized aggregates of histiocytes and Langerhans giant cells (n = 3), whereas the biopsies from the lower GI tract showed no specific features of CD. These 8 children would have been diagnosed incorrectly if endoscopy of the UGT had not been performed and the diagnosis was based on biopsies from the lower gastrointestinal tract alone. In 11 patients, the diagnosis after histological reassessment of biopsies from the upper and lower gastrointestinal tract differed from the final diagnosis (Fig. 1, Table 3). Seven patients were diagnosed as having IC after histological reassessment of all of the biopsies, but turned out to have CD as the final diagnosis, and 2 patients diagnosed as having UC after histological reassessment of all of the biopsies turned out to have CD. Two patients diagnosed as having IC after histological reassessment of biopsies turned out to have UC as the final diagnosis.
Histopathological Findings in the UGT
The presence of esophageal inflammation was significantly higher in patients with CD compared with children with UC (P = 0.008), but was not significantly different compared with children with non-IBD (P = 0.28). Diffuse inflammation, consistent with gastroesopagheal reflux, was similarly found in all 3 groups. The presence of focal inflammation was significantly higher in patients with CD compared with the children with non-IBD (11% vs 0%; P = 0.006) (Table 4).
The presence of microscopic gastric inflammation was significantly higher (P < 0.0001) in patients with CD compared with children with non-IBD, but it was not significantly different compared with children with UC (P = 0.19). Several patients had >1 type of inflammation in their biopsies from the antral and corpus mucosa. Focally enhanced gastritis was significantly more frequently seen in patients with CD (69%) compared with patients with UC (24%, P < 0.001) and patients with non-IBD (7%, P < 0.0001). The lesions were located in the antral mucosa in 22, in the corpus mucosa in 12, and in both in 27 patients. Specificity and positive predictive value of focally enhanced gastritis in CD were 87.1% and 78.6%, respectively. The lesions in all 3 groups consisted of a focal accumulation of chronic inflammatory cells (mononuclear cells; lymphocytes and plasma cells), active inflammatory cells (granulocytes; neutrophils and eosinophils), or both types of inflammatory cells. There was no significant difference in the composition of inflammatory infiltrates between the groups. Focal and diffuse inflammation occurred more frequently in patients with CD compared with non-IBD patients (P = 0.002 and P < 0.0001, respectively) but was found with equal frequency in patients with UC (both not significant). The composition of these inflammatory infiltrates was not significantly different between the groups. Reactive antrum gastritis, which may point to a bile-associated chemical gastritis, was frequently seen in non-IBD children (43%). H pylori infection was found in the biopsies of 10 patients with no significant difference between the groups (3 CD, 3 UC, 4 non-IBD) (Table 4).
The presence of duodenal inflammation was significantly higher in patients with CD compared with children with UC and non-IBD children. Focal cryptitis was seen significantly more frequently in patients with CD (19%) compared with children with UC and non-IBD children (0% and 1%; P = 0.008 and P = 0.001, respectively). Specificity and positive predictive value of focal cryptitis in CD were 99% and 93%, respectively (Table 4).
In our study, the diagnosis of CD was based solely on granulomatous inflammation of the UGT in 11% of the children. These children would not have been diagnosed correctly without UGT endoscopy. Furthermore, focal cryptitis of the duodenum and focally enhanced gastritis were found significantly more frequently in children with CD compared with children with UC and non-IBD children. Of these inflammatory changes, focal cryptitis of the duodenum showed a good specificity and positive predictive value of 99% and 93%, respectively, whereas the specificity and positive predictive value of focally enhanced gastritis in CD, 87.1% and 78.6%, respectively, were less convincing.
The value of UGT endoscopy is still a topic of debate. Performing endoscopy of the UGT increases the duration of the procedure, the risk of anesthesia, and procedural complications. Moreover, the increased number of biopsies also increases the costs. The European Society for Pediatric Gastroenterology, Hepatology, and Nutrition IBD Working Group has recommended routine UGT endoscopy at initial presentation in every child suspected of IBD (17). In contrast, a report from the working group of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition and the Crohn's and Colitis Foundation of America refrained from recommending routine diagnostic use of endoscopy of the UGT in the diagnostic assessment of children with suspected IBD (18). In the published, so far sparse, literature, the frequency of pediatric patients with CD whose diagnosis relies on detection of granulomas in the UGT ranges between 7% and 20% (7,11,19). Based on the latter data, we advise the routine use of endoscopy of the UGT, to start in an early phase as being the most appropriate treatment.
The number of children with the diagnosis of CD based on inflammation isolated to the UGT would increase if besides granulomas, other histological markers for CD could be identified. To date, few studies have evaluated UGT histology in children with suspected IBD at initial diagnostic assessment and compared the histopathological findings in the UGT of patients with CD and UC and patients without IBD (6,7,9–11). Approximately two-thirds of all patients with CD and half of those with UC have microscopic mucosal lesions in the UGT (6,10,11); however, it remains controversial whether the frequently found nonspecific microscopic UGT lesions are of clinical relevance for the patient with IBD. To our knowledge, this is the first study showing that focal duodenal cryptitis differentiates between CD and UC or non-IBD. In 2 other pediatric studies, cryptitis of the duodenum was reported in, respectively, 26% (10) and 8% (20) of children with CD but was not seen in patients with UC or non-IBD patients; but because of the small number of patients in both studies, the difference between groups did not reach statistical significance.
Furthermore, our study shows that focally enhanced gastritis supports the diagnosis of CD but does not differentiate reliably between CD and UC or non-IBD. This finding corresponds with the study of Sharif et al (13). Oberhuber et al (21) found that focally enhanced gastritis was present in 76% of adult patients with CD compared with 0.8% of healthy controls.
Our data should be interpreted in the context of the following limitations. First, data were collected retrospectively; however, IBD is a disease without well-defined diagnostic criteria and no diagnostic criterion standard. Because of the retrospective design, we were able to integrate clinical, endoscopic, histological, and radiological features and use the final diagnosis as our reference standard. We used stringent criteria to classify our patients and were able to follow all of the patients for a long period. Furthermore, the non-IBD group is not an average of the general population; therefore, the prevalence of inflammation in the UGT may be higher in the non-IBD group compared with the general healthy population. None of these patients developed IBD or other gastrointestinal disease during the follow-up period.
In conclusion, in 11% of the children with CD, the diagnosis was based solely on finding granulomatous inflammation in the UGT. These children would have been misdiagnosed if endoscopy of the UGT had not been performed and the diagnosis would have been based on biopsies from the lower GI tract alone. Focal duodenal cryptitis is a significant finding, pointing toward a diagnosis of CD. The presence of focally enhanced gastritis suggests underlying CD but is not exclusive to this condition. It does not reliably differentiate among patients with CD or UC and patients without IBD. All of the other histological findings in the UGT do not differentiate between CD and UC or non-IBD. Therefore, the present study underlines that endoscopy of the UGT is essential in the diagnostic assessment of childhood IBD.
1. Kugathasan S, Judd RH, Hoffmann RG. Epidemiologic and clinical characteristics of children with newly diagnosed inflammatory bowel disease in Wisconsin: a statewide population-based study. J Pediatr 2003; 143:525–531.
2. Mamula P, Telega GW, Markowitz JE. Inflammatory bowel disease in children 5 years of age or younger. Am J Gastroenterol 2002; 97:2005–2010.
3. Carvalho RS, Abadom V, Dilworth HP, et al. Indeterminate colitis: a significant subgroup of pediatric IBD. Inflamm Bowel Dis 2006; 12:258–262.
4. Heyman MB, Kirschner BS, Gold BD, et al. Children with early-onset inflammatory bowel disease: analysis of a pediatric IBD consortium registry. J Pediatr 2005; 146:35–40.
5. Hildebrand H, Fredrikzon B, Holmquist H, et al. Chronic inflammatory bowel disease in children and adolescents in Sweden. J Pediatr Gastroenterol Nutr 1991; 13:293–297.
6. Abdullah BA, Gupta SK, Croffie JM, et al. The role of esophagogastroduodenoscopy in the initial evaluation of childhood inflammatory bowel disease: a 7 year study. J Pediatr Gastroenterol Nutr 2002; 35:636–640.
7. Castellaneta SP, Afzal NA, Greenberg M, et al. Diagnostic role of upper gastrointestinal endoscopy in pediatric inflammatory bowel disease. J Pediatr Gastroenterol Nutr 2004; 39:257–261.
8. Kaufman SS, Vanderhoof JA, Young R. Gastroenteric inflammation in children with ulcerative colitis. Am J Gastroenterol 1997; 92:1209–1212.
9. Ruuska T, Vaajalahti P, Arajärvi P, et al. Prospective evaluation of upper gastrointestinal mucosal lesions in children with ulcerative colitis and Crohn's disease. J Pediatr Gastroenterol Nutr 1994; 19:181–186.
10. Tobin JM, Sinha B, Ramani P, et al. Upper gastrointestinal mucosal disease in pediatric Crohn disease and ulcerative colitis: a blinded, controlled study. J Pediatr Gastroenterol Nutr 2001; 32:443–448.
11. Lemberg DA, Clarkson CM, Bohane TD, et al. Role of esophagogastroduodenoscopy in the initial assessment of children with inflammatory bowel disease. J Gastroenterol Hepatol 2005; 20:1696–1700.
12. Sonnenberg A, Melton SD, Genta RM. Frequent occurrence of gastritis and duodenitis in patients with inflammatory bowel disease. Inflamm Bowel Dis 2011; 17:39–44.
13. Sharif F, McDermott M, Path MRC, et al. Focally enhanced gastritis in children with Crohn's disease an ulcerative colitis. Am J Gastroenterol 2002; 97:1415–1420.
14. Turner D, Griffiths AM. Esophageal, gastric, and duodenal manifestations of IBD and the role of upper endoscopy in IBD diagnosis. Curr Gastroenterol Rep 2007; 9:475–478.
15. Paerregaard A. What does the IBD patient hide in the upper gastrointestinal tract? Inflamm Bowel Dis 2009; 15:1101–1104.
16. Chong SKF, Blackshaw AJ, Boyle S, et al. Histological diagnosis of chronic inflammatory bowel disease in childhood. Gut 1985; 26:55–59.
17. IBD Working Group of the European Society for Paediatric Gastroenterology, Hepatology and Nutrition. Inflammatory bowel disease in children and adolescents: recommendations for diagnosis: the Porto criteria. J Pediatr Gastroenterol Nutr 2005;41:1–7.
18. Bousvarous A, Antonioli DA, Colletti RB, et al. Differentiating ulcerative colitis from Crohn disease in children and young adults: report of a working group of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition and the Crohn's and Colitis Foundation of America. J Pediatr Gastroenterol Nutr 2007; 44:653–674.
19. Kundhal PS, Stormon MO, Zachos M, et al. Gastral antral biopsy in the differentiation of pediatric colitides. Am J Gastroenterol 2003; 98:557–561.
20. Schmidt-Sommerfeld E, Kirschner BS, Stephens JK, et al. Endoscopic and histologic findings in the upper gastrointestinal tract of children with Crohn's disease. J Pediatr Gastroenterol Nutr 1990; 11:448–454.
21. Oberhuber G, Püspök A, Oesterreicher C, et al. Focally enhanced gastritis: a frequent type of gastritis in patients with Crohn's disease. Gastroenterology 1997; 112:698–706.