Secondary Logo

Journal Logo

Esophageal Crohn Disease in Children: A Clinical Spectrum

Ramaswamy, Kannan; Jacobson, Kevan; Jevon, Gareth; Israel, David

Journal of Pediatric Gastroenterology and Nutrition: April 2003 - Volume 36 - Issue 4 - p 454-458
Original Articles: Gastroenterology

Background The incidence of esophageal Crohn disease (ECD) in adults ranges from 0.2% to 11.2% and in children is up to 43%. The aim of the study was to determine the clinical and endoscopic spectrum of ECD and its prevalence in our patient population.

Methods Chart review of children with Crohn disease (CD). Esophageal Crohn disease was defined by accepted endoscopic and/or histologic findings.

Results 210 children with CD were identified; 27 of those children had ECD. Nine children presented with specific upper GI symptoms; dysphagia, heartburn, nausea, vomiting, and odynophagia. Esophagoscopy in children with upper gastrointestinal symptoms revealed deep ulcers (n = 2), aphthous ulcers (n = l), erosions (n = l), edematous nodules, (n = l) and normal mucosa (n = 4). In asymptomatic children aphthous ulcers (n = 5), erosions (n = 3), deep ulcers (n = 3), and normal looking mucosa (n = 7) were seen. Twenty children also had gastric lesions, 3 children had duodenal lesions, and 3 children had both duodenal and gastric involvement. All 27 children had evidence of ileo-colonic or colonic disease. Acid suppressive medications were given only to children with upper GI symptoms and endoscopic esophageal lesions. The mean duration of follow-up from diagnosis of CD was 3.02 years (range 2 months—11.7 years). At last follow-up review, 7 children were receiving acid suppression and no children were receiving steroids. There were no complications related to ECD.

Conclusion The prevalence of endoscopic ECD is 7.6% but as many as 17.6% of our patient population had histologic evidence of ECD. The clinical and endoscopic spectrum of ECD are highly variable and poorly correlate with each other.

Division of Gastroenterology, and Pathology, BC's Children's Hospital, Vancouver, B.C. Canada

Address correspondence to Dr. David M. Israel, Division of Pediatric Gastroenterology, BC's Children's Hospital, 4480 Oak Street, Vancouver V6H 3V4, B.C. Canada (e-mail:

Crohn disease (CD) is a lifelong disease that may affect any part of the gastrointestinal tract. Involvement of the esophagus was previously thought to be uncommon (1). However, based on retrospective studies and endoscopic findings, the prevalence of esophageal Crohn Disease (ECD) ranges from 0.2% to 11.2% in adults (2,3) and 6.5% to 43% in children (4–8). In children the low estimates are based on a retrospective study in patients with upper gastrointestinal symptoms and the upper estimates include asymptomatic patients with histologic evidence of esophageal Crohn disease. Limited information is available on the clinical spectrum of ECD. Moreover, the clinical significance of histologic involvement of the esophagus remains to be determined.

The objectives of this study were to estimate the prevalence of ECD in our patient population, to better delineate the clinical course, and to correlate clinical course with endoscopic and histologic findings. We also report the therapeutic interventions used in these children and their therapeutic effect.

Back to Top | Article Outline


All the charts of children diagnosed with inflammatory bowel disease (IBD) attending British Columbia Children's Hospital, Vancouver, Canada between January 1990 and May 2001 were reviewed. Detailed clinical history, radiologic, endoscopic and histologic data, treatment and outcome were collected. Patients with Crohn disease were identified according to acceptable criteria (7). All patients included in the study had a barium swallow with small bowel follow-through and a colonoscopy. However, UGI endoscopy was not performed in all children as indications varied among physicians. All children with upper gastrointestinal symptoms had UGI endoscopy, but as a result of the variation in practice among physicians, UGI endoscopy was only performed in some children without upper gastrointestinal symptoms. The presence of aphthous ulcers, erosions, deep ulcers, or nodular lesions in the esophagus was considered endoscopic evidence of ECD if it was associated with Crohn disease elsewhere. Moreover, the histologic findings on esophageal biopsies of noncaseating granulomata or chronic deep inflammatory infiltration in the lamina propria without basal cell hyperplasia were considered evidence of ECD. Heartburn, dysphagia, and odynophagia were considered specific upper gastrointestinal symptoms. Nausea, vomiting, and abdominal pain were considered overlapping symptoms.

Back to Top | Article Outline


Two hundred and ten children with Crohn disease were identified. Of all the children identified, 144 of the children had an UGI endoscopy and 27 (15 boys and 12 girls) had endoscopic and/or histologic evidence of ECD. In 22 of the 27 children, ECD was detected at the time of initial diagnosis. The remaining 66 patients in whom a UGI endoscopy was not performed had no symptoms suggestive of upper gastrointestinal disease. The mean age at diagnosis of ECD was 11.0 years (range 9 months to 15 years). The patients were grouped according to the presence or absence of upper gastrointestinal symptoms and by the presence or absence of endoscopic involvement of the esophagus (see Figures 1 and 2).

FIG. 1.

FIG. 1.

FIG. 2.

FIG. 2.

Back to Top | Article Outline

Clinical Manifestations

All 27 children with ECD had symptoms of lower gastrointestinal tract involvement. However, clinical symptoms suggestive of upper gastrointestinal tract involvement were found in only 9 children. The symptoms included dysphagia (6 cases), heartburn (5 cases), vomiting (5 cases), and odynophagia (1 case). A combination of these symptoms was seen in 5 cases. The remaining 18 children who were found to have incidental ECD detected at endoscopic evaluation for their ileo-colonic symptoms were grouped as asymptomatic ECD. Extra intestinal manifestations identified in 25.9% of cases included erythema nodosum (2), aphthous mouth ulcers (3), episcleritis (1), and arthritis (1). Extra intestinal manifestations were found in 22.4% in children with Crohn disease without esophageal involvement. A family history of IBD was present in 44.4% of ECD patients compared to 30.9% of patients with Crohn disease and no esophageal involvement.

Back to Top | Article Outline

Endoscopy and Radiologic Studies

All 27 children with ECD had a barium small bowel follow-through study within 6 to 8 weeks of the endoscopic examination. None of the children had detectable radiologic signs of CD in the upper gastrointestinal tract. Combining radiology and colonoscopy findings identified ileo-colonic involvement in 18 children, isolated small bowel disease in 3 children, and isolated colonic disease in 6 children (Fig. 2).

In symptomatic ECD patients, endoscopy identified deep ulcers in 2 patients, erosions, aphthous ulcers, and erythematous nodules each in 1 patient, and normal-looking mucosa in 4 patients. Histological evidence of ECD was evident in all symptomatic patients. In the asymptomatic patients, 5 children had aphthous ulcers, 3 children had deep ulcers, and 3 children had erosions. The remaining 7 children had normal-looking mucosa but showed histology evidence of Crohn disease.

Gastroduodenal involvement was detected endoscopically in all symptomatic children and in 17 of 18 asymptomatic children (Fig. 2), including 20 cases of gastric CD, 3 cases of duodenal CD, and 3 cases of both gastric and duodenal CD. Endoscopic and/or histologic evidence of CD of the stomach and/or duodenum were also found in 95 of 117 (81%) children without ECD (Fig. 1).

Back to Top | Article Outline


Esophageal inflammation was evident in 25 of 27 patients. Esophageal biopsies included epithelium with lymphocytosis in 17 children and acute inflammation in 13 cases. Epithelial ulceration was noticed in 5 cases along with granulation tissue and chronic inflammatory cell infiltrates. Presence of lamina propria was documented in 16 specimens. All of those specimens showed focal chronic inflammation with plasma cells, lymphocytes that extended into the lamina propria. Noncaseating granulomata were identified in 7 of 27 (25%) of the children. In the 2 children with unremarkable histology, aphthous ulcers were noted at endoscopy.

Back to Top | Article Outline

Treatment and Outcome

The mean duration of follow-up for the 9 children with upper gastrointestinal symptoms and ECD was 3.1 years (7 months to 5 years). Our group uses acid suppression in the treatment of ECD only in children with specific upper gastrointestinal symptoms and/or endoscopic lesions. Seven of these children were treated with acid suppressant medications (omeprazole n = 5, ranitidine n = 2) along with treatment for ileo-colonic disease. Omeprazole was started at 1.4 mg/kg/d in a single morning dose and ranitidine was given 4 to 6 mg/kg/d in two divided doses. The two children who presented with nausea and vomiting and only microscopic ECD were not given any specific acid suppressive treatment.

However, one child had an ileo-colonic resection for persistent lower gastrointestinal symptoms and the other child received a 5-ASA compound and steroids. Neither child had any recurrence of symptoms. During the last visit, 3 children were receiving acid suppression and one child was still symptomatic. In the remaining 4 children, acid suppression was discontinued with resolution of gastrointestinal symptoms.

Of the 9 symptomatic ECD children, 4 children required 6-MP because of steroid dependency. Two of these children required an ileo-colonic resection with primary anastomosis and one had a stricturoplasty, small bowel resection, and right hemicolectomy.

The mean duration of follow-up of the asymptomatic ECD group was 3.0 years (2 months—11.7 years). Two children presenting with deep ulcers were treated with omeprazole. The remaining children received treatment only for ileo-colonic disease. In this group, 2 children required a right hemicolectomy and one child had a subtotal colectomy for refractory Crohn colitis following an unsuccessful trial with intravenous cyclosporine. Three children were steroid dependent and required 6-MP in their management. Four children reported heartburn and nausea once their diarrhea and anorexia were controlled by steroids and/or nutritional supplementation. Three of them received omeprazole and one had ranitidine.

None of the children with ECD developed esophageal complications such as stricture or stenosis or fistulae during the follow-up period.

Back to Top | Article Outline

Prevalence Estimates

Of the 144 children with IBD who underwent an UGI endoscopy, 16 children had endoscopic evidence of ECD and 11 children had only histologic findings. The lowest estimate of the prevalence of ECD based on endoscopic validated cases alone is therefore 16 of 210 (7.6%). Including histologic findings as evidence for ECD, the prevalence is 27 of 210 (12.8%). The actual prevalence may have been higher if all children with CD had had an upper GI endoscopy. This prevalence can be estimated first by calculating the prevalence of asymptomatic ECD in those who had undergone upper GI endoscopy (18/118 = 15.2%; see Figure 1). Then apply this ratio to the number of children with no upper gastrointestinal symptoms who did not have an upper GI endoscopy (210–144 = 66 children; see Figure 1). Calculations predict that an additional 10 children with ECD would have been detected bringing the total children with ECD to 37 of 210. Thus the estimated prevalence of ECD in our patient population is 17.6%.

Back to Top | Article Outline


This large cohort of 210 children with Crohn disease allows for the determination of a reliable estimate of the prevalence of esophageal Crohn disease and correlation of clinical course to endoscopic findings.

Reports on the prevalence of ECD vary widely in both adults and pediatrics. Decker et al. (2) previously reported that only 20 of 9900 (0.2%) adults with CD seen over a period of 22 years had ECD. A higher prevalence was reported by D'Haens (3) who demonstrated ECD in 14 of 124 (11%) adults with CD investigated with upper GI endoscopy over a period of 12 years. Both of these studies were retrospective and used macroscopic endoscopic evidence as the only criteria for diagnosis that likely resulted in an underestimation of the prevalence. In the only prospective study of ECD in adults a prevalence of 5.1% was seen (9). In the pediatric age group there have been few prospective and retrospective studies that provide estimates of the prevalence of ECD. In one prospective study, endoscopy identified in ECD 2 of 40 (5%) children with CD and a total of 17 of 40 (42.5%) had histologic evidence of ECD (7). Additional prospective studies detected ECD in 8 of 31 (26%) children based on either endoscopic or histologic findings (6) and in 7 of 56 (13%) based on macroscopic findings alone (5). In our study the lowest estimate of prevalence for endoscopic ECD was 7.6% and the highest estimate was 17.6% when including histology findings as diagnostic criteria. These findings suggest that the prevalence of ECD in adults and children may be similar and is at least 5%. Low estimates in some of the retrospective adult studies may be a reflection of patient selection for UGI endoscopy.

There is wide variability in the clinical manifestations of ECD and many patients do not have specific symptoms related to ECD. Moreover, isolated esophageal Crohn disease is uncommon (10) and in our study no children were identified with isolated ECD.

Esophageal Crohn disease is usually associated with advanced ileo-colonic disease (1,11). In our study group, 10 children had severe extraesophageal CD, and 7 of them were on 6-MP and 6 of these children required surgery which indirectly reflects the severity of extra esophageal CD. All children in the study had evidence of ileo-colonic involvement. Previous reports have suggested that there is an increased association of extraintestinal manifestations with ECD (1,11). However in our study 25.9% of patients had extra intestinal manifestations which was similar to the rate of extra-intestinal manifestations seen in children with CD and no esophageal involvement (22.4%).

In two different prospective studies by Cameron (5) and Mashako (6), upper GI symptoms were present in only 15% and 25% of children with CD. However the prevalence of upper GI tract abnormalities ranged from 42% to 71% respectively. Alcantara found that 17% of 41 adults had upper gastrointestinal symptoms but only 5% had endoscopic ECD (9). In our study 33% of the children with ECD had specific upper GI symptoms. According to Lenaerts (4), dysphagia and/or epigastric pain, nausea and vomiting were noticed more frequently in children with upper gastrointestinal CD but were also seen in children without upper GI disease, suggesting that these symptoms are not specific to the upper GI tract disease. In our study there was no difference in the endoscopic findings for those ECD children with symptoms and those children without symptoms. In the pediatric age group, there is often an overlap of symptoms (such as nausea, anorexia, and abdominal pain) between esophageal and small bowel CD. It is therefore difficult to identify an upper gastrointestinal site of disease based on clinical evaluation alone. Thus UGI endoscopy and biopsies are important in the evaluation of pediatric IBD, even in the asymptomatic child.

The presence of histologic abnormalities in radiologically and endoscopically normal mucosa has been well described (12). Hence, endoscopy with biopsy is generally considered the gold standard for diagnosis of early lesions that may be asymptomatic. Huchzermeyer et al. (13) have previously observed that the spectrum of esophageal injury in Crohn disease varied from mild esophagitis with small erosions to transmural involvement with esophageal perforation and fistulization to adjacent organs. In addition to aphthous ulcers, other findings well described in ECD are the presence of erythema, ulceration, erythematous nodules, and polypoid lesions (14). Pseudomembrane formations, progressive esophageal narrowing resulting in strictures and formations of multiple mucosal bridges have been described in advanced stages of ECD (15).

Epithelioid granulomatas are considered diagnostic for Crohn disease. Seven of 27 (25%) of our patients had granulomas in the esophageal biopsy specimen. In a review of the literature, granulomas were documented in 39% of patients with ECD detected at endoscopy with biopsy and on some surgically resected specimens (16). Focal infiltration with mononuclear cells and histiocytes into the lamina propria that extends to the muscular layer is often considered histologic evidence of Crohn esophagitis (3,8). These changes in lamina propria often help to differentiate ECD from other causes of epithelitis. Presence of intraepithelial lymphocytes with irregular nuclear contour is also a diagnostic feature, but not pathognomonic (17).

Management of esophageal Crohn disease is highly variable. Case studies have shown that steroids were effective in controlling inflammation (18,19). However, upper gastrointestinal symptoms may be effectively managed with acid suppression rather than steroids. Interestingly, four of our children in the asymptomatic category become aware of heartburn and epigastric pain only after their lower gastrointestinal symptoms were controlled with standard induction treatment. These symptoms were well controlled with acid suppressive medications.

Our group has adopted the approach of using acid suppression treatment for ECD only in children with specific upper gastrointestinal symptoms and/or endoscopic lesions. However, microscopic inflammatory changes per se were not considered as indication for treatment. The small number of children who were treated with acid suppression and the variation in disease involvement preclude a direct comparison of omeprazole and ranitidine for the symptomatic ECD.

Duration of acid suppressive medication is dependent upon the periodic reappraisal of symptoms during follow-up. Meihsler et al. (20), using gastroduodenal and intestinal permeability studies in patients with upper GI Crohn lesions, have shown that symptom scores consistently decreased following treatment with antisecretory drugs, despite no change in intestinal permeability.

In a study by Decker et al. (2), 11 of 20 adult patients with ECD received either azothioprine, 6-mercaptopurine, cyclosporin A alone or in combination with steroids for the management of refractory esophageal Crohn disease and coexisting extra esophageal lesions. However, in this study only 6 patients among 11 responded to immunomodulator therapy. Three of those who received immunomodulators (27.2%) showed improvement on azothioprine or 6-MP together with steroids and one patient responded to cyclosoprine A with esophageal dilations. Two of the responders received immunomodulatory therapy primarily for extraesophageal CD. Hence it is difficult to judge the true efficacy of immunomodulators in resistant ECD. Moreover, randomized controlled trials using various therapeutic agents in the management of ECD patients have been faced with the problem of recruiting adequate numbers.

The natural course of esophageal Crohn disease has previously been described by D'haens et al (3), who showed that 57% (8/14) of their patients had no recurrence of esophageal lesions or symptoms during subsequent flare-up. Three patients with flare up (21.4%) had persistent lesions despite corticosteroid treatment. The remaining three (21.4%) with a flare up had a recurrence of esophageal symptoms and associated endoscopic lesions following complete healing with the initial treatment. None of the children in our study had repeat endoscopy for recurrence of upper gastrointestinal symptoms. Thus, the natural course of ECD in children remains poorly described and the outcome of those who presented with endoscopic lesions without symptoms remain unknown.

The clinical significance of histologic ECD can only be determined by a long-term follow-up and long-term endoscopic surveillance. Nevertheless, considering the high prevalence of histologic ECD in children throughout several different centers in the world and the low rates of endoscopic ECD in adults it may be reasonable to postulate that histologic ECD is unlikely to progress to a more overt disease over time.

In conclusion, in our study all of our pediatric ECD patients were presented with ileo-colonic symptoms, but the majority of them were asymptomatic from their esophageal lesions. Currently there are no standardized treatment guidelines for patients with either specific upper gastrointestinal symptoms or without symptoms. Steroids are an effective therapy, however, recurrence of symptoms is commonly encountered during steroid withdrawal. Acid suppressive treatment has a definite role in controlling symptoms. Further controlled studies are required to follow the natural course of asymptomatic ECD in children and outcome of those managed with or without acid suppression.

Back to Top | Article Outline


1. Geboes K, Janssens J, Rutgeerts P, Vantrapp G. Crohn' s disease of the esophagus. J Clin Gastroenterol 1986; 8:31–7.
2. Decker GAG, Loftus EV, Pasha TM, Tremain WJ, Sandborn WJ. Crohn's disease of the esophagus: Clinical features and outcomes. Inflammatory Bowel Disease 2001; 7:113–9.
3. D'Haens G, Rutgeerts P, Geboes K, Vantrappen G. The natural history of esophageal Crohn's disease: Three pattern of evolution. Gastrointestinal Endoscopy 1994; 40:296–300.
4. Lenaerts C, Roy CC, Vaillancourt M, Weber AM, Morin CL, Seidman E. High incidence of upper gastrointestinal tract involvement in children with Crohn's disease. Pediatrics 1989; 83:777–81.
5. Cameron DJS. Upper and lower gastrointestinal endoscopy in children and adolescents with Crohn's disease: a prospective study. J Gastroenterol Hepatol 1991; 6:355–8.
6. Mashako MNL, Cezard JP, Navarro J, Mougenot JF, Sonsino E, Gargouri A, Maherzi A. Crohn's disease lesions in the upper gastrointestinal tract: Correlation between clinical, radiological, endoscopic, and histological features in adolescents and children. J Pediatr Gastroenterol Nutr 1989; 8:2–6.
7. Ruuska T, Vaajalahti P, Arajarvi P, Maki M. Prospective evaluation of upper gastrointestinal mucosal lesions in children with ulcerative colitis and Crohn's disease. J Pediatr Gastroenterol Nutr 1994; 19:181–6.
8. Schmidt-Sommerfeld E, Kirschner BS, Stephen JK. Endoscopic and histological findings in the upper gastrointestinal tract of children with Crohn's disease. J Pediatr Gastroenterol Nutr 1990; 11:448–54.
9. Alcantara M, Rodriguez R, Potenciano LM, Carrobles JL, Munoz C, Gomez R. Endoscopic and bioptic findings in the upper gastrointestinal tract in patients with Crohn's disease. Endoscopy 1993; 25:282–6.
10. Livolsi VA, Jaretzki A. Granulomatous esophagitis. a case of Crohn's disease limited to esophagus. Gastroenterology 1974; 64:313–9.
11. Degryse HRM, De schepper AMAP. Aphthoid esophageals ulcers in CD of the small bowel and colon. Gastrointest Radiol 1984; 9:197–201.
12. Korelitz BI, Waye JD, Kreuning J, Sommers SC, Fein HD, Beeber J, Gelberg BJ. Crohn's disease in endoscopic biopsies of the gastric antrum and duodenum. Am J Gastroenterol 1981; 76:103–9.
13. Huchzermeyer H, Paul F, Seifert, Frolich H, Rasmussen ChW. Endoscopic results in five patients with Crohn's disease of the esophagus. Gastroenterology 1976; 8:75–81.
14. Walker RS, Breuer R, Victor T, Gore RM. Crohn's esophagitis: a unique cause of esophageal polyposis. Gastrointestinal Endoscopy 1996; 43:511–4.
15. Hanai H, Honda H, Sugimato K, Kageoka et al. Endoscopic therapy for multiple mucosal bridges in the esophagus of a patient with Crohn's disease. Gastrointestinal Endoscopy 1999; 50:715–7.
16. Rudolph I, Goldstein F, Di marino AJ. Crohn s disease of the esophagus: Three cases and a literature review. Can J Gastroenterol 2001; 15:117–122
17. Mangano M, Wang H, Schnitt S, Antonioli D. Nature and significance of cells with irregular nuclear contours (CINC) in esophageal mucosal biopsies (abstract). Lab Invest 1991; 64:38A.
18. Treem WR, Ragsdale BD. Crohn's disease of the esophagus. A case report and review of the literature. J Pediatr Gastroenterol Nutr 1988; 7:451–5.7.
19. Ohta M, Konno H, Kamiya K, Suwa D, Baba M, Tanaka T, et al. Crohn's disease of the esophagus: Report of the case. Surg Today 2000; 30:262–7.
20. Miehsler W, Puspok A, Oberhuber G, Vogelsang H. Impact of different therapeutic regimens on the outcome of patients with Crohn's disease of the upper digestive tract. Inflamm Bowel Dis 2001; 7:99–10.

Crohn disease; Esophagus; Upper GI endoscopy; Prevalence

© 2003 Lippincott Williams & Wilkins, Inc.