Inflammatory bowel disease (IBD) typically involves the large bowel, small bowel (ileum / jejunum), or both. Involvement of the upper GI tract is less common, although its frequency is uncertain because common causes of inflammation such as gastro-oesophageal reflux and Helicobacter pylori infection also require exclusion. In the setting of known IBD, upper GI inflammation generally favours Crohn’s disease over ulcerative colitis (UC), while granulomas strongly suggest involvement by IBD and indicate Crohn’s disease rather than UC. New upper GI inflammation may raise the possibility of IBD, while unexplained new upper GI granulomas require exclusion of Crohn’s disease. Unfortunately, few histological patterns apart from granulomas are specific or discriminant. Lymphocytic oesophagitis is a poorly defined entity associated with IBD and/or Crohn’s disease in some studies. Focally enhanced gastritis (FEG), though initially regarded as typical of Crohn’s disease, probably has limited value apart from perhaps predicting IBD in children. Rarely, UC patients develop a duodenitis resembling the colorectal changes histologically. UC may also be associated, infrequently, with characteristic patterns of gastritis. Compared with adults, children with IBD are more likely to have upper GI involvement, to develop GI granulomas, and to undergo investigation for upper GI disease. Overall, upper GI granulomas assist with the diagnosis and classification of IBD but few other upper GI features are discriminatory between IBD and other causes or between UC and Crohn’s disease.
Small bowel disease, affecting the ileum/jejunum, is often present in patients with Crohn’s colitis. Indeed, imaging evidence of small bowel abnormalities helps to distinguish Crohn’s colitis from ulcerative colitis (UC). Similarly, isolated large bowel disease affects only a minority of Crohn’s disease patients at presentation (14%–32%).1 Crohn’s disease is also a cause of oesophageal, gastric, and duodenal inflammation but these rarely occur, and are rarely diagnosable, in the absence of jejunoileal or colorectal disease. There is limited evidence that upper gastrointestinal (GI) tract involvement by Crohn’s disease is associated with a worse outcome, e.g. higher rates of surgery.2
UC is a disease of the large bowel, typically involving the rectum at presentation and often extending proximally into the colon. Distribution is typically continuous. Distal terminal ileal involvement by UC may reflect ‘backwash’ of bowel contents rather than UC itself, although some experts question this explanation and assert that ileal disease probably represents true ileal UC.3 Other variations on the classic continuous colorectal distribution of UC include rectal sparing (though this is rare at presentation) and a periappendiceal and caecal ‘patch’ that is discontinuous with disease more distally. Therefore, UC is capable of being discontinuous in several ways. Since the 1990s there has been increasing awareness that the ability of inflammatory bowel disease (IBD) to extend into the upper GI tract is not exclusive to Crohn’s disease and that there are even patterns of duodenal and gastric inflammation in UC patients that may be a guide to genuine upper GI UC.4, 5
Awareness of the likely patterns of histological involvement of the upper GI tract by IBD will optimise the pathologist’s contribution to the investigation of unexplained upper GI symptoms. In addition, assessment of the histology of upper GI disease may help to confirm suspected IBD and to distinguish between UC and Crohn’s disease.
Macroscopic findings associated with Crohn’s disease of the upper GI tract include aphthous ulcers, other types of ulcer, erosions, and strictures.6–8 The stomach and duodenum may show erythema and oedema, and duodenal stenosis may occur.9, 10 Gastric erosions may predict histopathological involvement while ulcers, stenosis, and erythema may be associated with a higher frequency of granulomas.9
Examination of Upper Gastrointestinal Biopsies for Inflammatory Bowel Disease
There are several considerations for the pathologist when making or supporting a diagnosis of upper GI IBD (Fact Sheet 16.1).
1. Other common causes of upper GI inflammation usually require exclusion before suggesting IBD. These include reflux oesophagitis, Helicobacter pylori–associated gastritis, chemical/reactive gastritis (often secondary to nonsteroidal anti-inflammatory drug usage), and ‘peptic’ duodenitis. All may affect patients who have IBD. Furthermore, a focal or patchy gastritis may reflect H. pylori infection and does not necessarily indicate or suggest Crohn’s disease. In particular, stains for H. pylori are important if there is unexplained gastritis. However, it is worth noting that H. pylori–associated gastritis and other diseases, if superimposed on a background of IBD, might obscure the histological changes of IBD.11
2. Although many studies of upper GI IBD have attempted to identify patterns or features that suggest IBD, granulomas remain by far the most useful and most discriminant feature. If there is a known history of Crohn’s disease, granulomas strongly favour involvement by IBD over other causes.10 Indeed, many authors assert that granulomas are the only good evidence of Crohn’s disease of the upper GI tract.9 Although the diagnostic value of upper GI granulomas is accepted, the reported prevalence in established Crohn’s disease varies widely, e.g. 0%–83% for gastric granulomas.11, 12 This depends partly on study design. In a controlled paediatric study, there was a higher prevalence of gastric granulomas in Crohn’s disease than in Helicobacter-negative controls (p < 0.01).12 Upper GI granulomas in Crohn’s disease may be more common when endoscopy is abnormal than when it is normal13 and can occur in Crohn’s disease patients who have no history of colorectal granulomas.12 Even if granulomas are present in a patient with known IBD, other causes such as drugs, mycobacterial infection, Helicobacter gastritis, foreign material, and sarcoid may require consideration.
3. Distinction between IBD and other causes of ileitis or colitis is sometimes difficult. Multiple site GI biopsies may help make this distinction. Overall, upper GI inflammation may be more common in IBD than in other types of colitis or ileitis and may assist the distinction between IBD and other diagnoses as a cause of ileitis/colitis. According to one study, inflammatory changes in upper GI biopsies were considerably more common in those with Crohn’s ileitis than in those with non-Crohn’s ileitis.14
4. Distinction between ileocolonic Crohn’s disease and UC is also sometimes difficult or impossible and the presence, pattern, and distribution of upper GI histological changes may assist with this distinction. Granulomas, as usual, strongly favour Crohn’s disease over UC and are almost diagnostic of the former. Upper GI inflammation per se is often interpreted as a pointer towards Crohn’s disease rather than UC, and this remains broadly true.5 However, the increasing recognition that UC may be a direct cause of upper GI tract inflammation has led to a more cautious approach. Therefore, the initial classification or subsequent reclassification of IBD as Crohn’s disease rather than UC should not depend either solely or mainly on the presence of upper GI inflammation.15
5. A new diagnosis of IBD is very rarely made or suggested on the basis of upper GI inflammatory changes.16 The exception is granulomas, which raise the possibility of Crohn’s disease. Some upper GI histological patterns may be associated with or predictive of IBD. For example, ‘lymphocytic oesophagitis’ and ‘focally enhanced gastritis’ (FEG) may be predictive of IBD in children. However, reports are often inconsistent.17
6. Upper GI involvement by IBD appears to be more common in children than in adults. Part of the reason is that clinicians diagnose upper GI IBD more readily in children than in adults because inflammatory conditions such as gastro-oesophageal reflux and Helicobacter gastritis are less common. In addition, children are more likely to have upper GI endoscopy (and full ileocolonoscopy) both at initial diagnosis and at follow-up, e.g. 100% in children and 54% in adults with Crohn’s disease according to one report.14 However, many studies show a genuinely higher prevalence of upper GI histological changes in children with IBD (see Table 16.1). Specifically, upper GI granulomas are more common in children. In one study, the prevalence of upper GI histological abnormalities after an initial diagnosis of Crohn’s ileitis was the same in children as in adults, but oesophageal granulomas occurred only in children (33%) while ‘lymphocytic oesophagitis’ was much more common in children (41% vs. 2%)14 and upper GI granulomas (i.e. oesophageal, gastric, and duodenal combined) were probably more common in children.14
7. In Crohn’s disease, histological changes are more common than clinical abnormalities and symptoms, and histological abnormalities may be more common than endoscopic abnormalities.9 However, figures vary. Furthermore, distinction between upper GI IBD and other upper GI diseases is difficult endoscopically and clinically as well as histologically. For example, a study of new IBD, after exclusion of non-steroidal anti-inflammatory drugs (NSAIDs) and Helicobacter infection as causes, showed upper GI tract changes as follows10:
Symptoms in 32% Crohn’s disease
Endoscopic abnormalities in 55% Crohn’s disease and 0% UC
Focally enhanced gastritis in 54% Crohn’s disease and 23% UC
Granulomas in 28% Crohn’s disease and 0% UC
Overall, the true prevalence of involvement of the upper GI tract by IBD is very difficult to ascertain for several reasons.
Distinguishing IBD from other causes of upper GI inflammation is very difficult.
Very few upper GI histological features or morphological patterns apart from granulomas are specific for IBD, UC, or Crohn’s disease.
IBD patients, especially children, may have a higher rate of upper GI inflammatory disorders than the population, regardless of whether or not these changes reflect true involvement by IBD.7
The rate of endoscopic examination of the upper GI tract and the proportion of patients who have biopsies varies between centres and between age groups and is influenced by clinical features.9
In one study, the most specific features for confirming involvement of the upper GI tract in known Crohn’s disease, and also those with the best predictive value, were focal acute inflammation of stomach or duodenum, surface intraepithelial neutrophils in the duodenum, deep acute inflammation of the duodenum, and granulomas.11
A diagnosis of upper GI IBD requires prior exclusion of common causes of upper GI inflammation such as GORD, Helicobacter pylori gastritis, and reactive gastritis.
Upper GI granulomas in the setting of Crohn’s disease are likely to indicate upper GI Crohn’s disease.
Unexplained new upper GI granulomas merit exclusion of Crohn’s disease.
Other possible causes of upper GI granulomas include H. pylori, mycobacteria, foreign material, crypt rupture, and sarcoidosis.
Apart from granulomas, upper GI inflammatory changes alone are rarely sufficient to make or suggest a new diagnosis of IBD.
Upper GI inflammation may help to favour a diagnosis of IBD over non-IBD.
Upper GI inflammatory changes may assist with subclassification of known IBD as ulcerative colitis (UC) or Crohn’s disease.
Granulomas strongly favour Crohn’s disease over UC.
Upper GI involvement is probably more common overall in Crohn’s disease than in UC.
The frequency of demonstrable involvement by IBD differs between upper GI sites, e.g. IBD-related granulomas are more common in the stomach than in the oesophagus or duodenum.
Upper GI involvement by IBD appears to be more common in children than in adults.
IBD-related upper GI granulomas are more common in children than in adults.
IBD-related oesophageal granulomas probably occur only in children.
|CD All||CD child||CD adult||UC All||UC child||UC adult||Control|
|UGI sites combined|
|Focally enhanced gastritis||12–54||44||14||12–23||21–30||19|
Common causes of oesophageal inflammation require exclusion before making or suggesting a diagnosis of oesophageal IBD (Fact Sheet 16.2). Gastro-oesophageal reflux (GORD) is particularly widespread in the general population, and also affects IBD patients.5, 12 Oesophagitis with a variety of aetiologies was observed in 72% of children with Crohn’s disease and 50% with UC in one report.12
Oesophageal granulomas related to Crohn’s disease may occur in children with Crohn’s disease but occur rarely or never in adults with Crohn’s disease.
Oesophageal Crohn’s disease can be diagnosed confidently only if granulomas are present.
The pattern of ‘lymphocytic oesophagitis’ may be associated with Crohn’s disease, especially in children, but is inconsistently and poorly defined.
Granulomas of the oesophagus may represent new Crohn’s disease but other causes, particularly infection and foreign material, always require exclusion.
Oesophagitis appears to be less common in UC than in controls.
Oesophageal UC is difficult to confirm and is probably very rare.
Oesophageal Granulomas in Inflammatory Bowel Disease
Oesophageal granulomas occur in some patients with Crohn’s disease. They are often located in the lamina propria beneath the epithelium (Figure 16.1A).18 In a paediatric study the rate was 20%12 and in a study of children presenting with Crohn’s ileitis the rate was 33%.14 In adults, the rate appears to be much lower and may be zero. In practice, granulomas strongly suggest oesophageal Crohn’s disease if there is a known history of Crohn’s disease or of IBD. Conversely, involvement of the oesophagus by Crohn’s disease cannot be diagnosed with confidence unless there are granulomas.18 In patients with no history of IBD, oesophageal granulomas in a child would raise the possibility of new IBD. Other causes of granulomas such as infection and foreign material require exclusion.18
A pattern of ‘lymphocytic oesophagitis’ (LO) may be related to Crohn’s disease, but the topic is controversial.19 All definitions of LO require prominent lymphocytosis, usually centred on the papillae and accompanied by relatively few neutrophils or eosinophils (Figure 16.1B and see Figure 11.6).19–21 However, definitions of LO vary between authors (see Chapter 11). Furthermore, the associations with IBD are not consistent. For example, LO was predictive of IBD in 40% of children in one study, was more common in children with Crohn’s disease than those with UC or controls in another study, and had no significant association with IBD in a further report.19–21 In patients presenting with Crohn’s ileitis, there was LO in 44% of children and only 2% of adults.14 In another study, there was LO in 4% UC and 14% controls.5 The evidence suggests that an LO pattern is more common in Crohn’s disease than in UC or in other settings, and that LO in Crohn’s disease is more common in children than in adults, but that LO may not be reliable for predicting IBD or Crohn’s disease. Anecdotally, lymphocytic inflammation of the oesophagus appears to be common in children with Crohn’s disease, but its significance is difficult to determine in an individual case given that reflux may also be responsible for lymphocytic infiltrates of the oesophageal mucosa (Figure 16.1C).