Various intestinal and extraintestinal conditions may complicate IBD. Intestinal complications include infections, strictures, obstruction, fistulas, dysplasia, and malignancy. Extraintestinal complications or associations include cutaneous disease, arthropathy, ocular problems, primary sclerosing cholangitis, and other forms of liver disease. The intestinal complications whose diagnosis and assessment relies most heavily on histopathology are cytomegalovirus (CMV) infection, dysplasia, and malignancy. Reactivation of latent CMV is more likely in ulcerative colitis than Crohn’s disease, is more common in those with immunosuppression and/or severe disease, and is associated with a worse clinical outcome. CMV inclusions are often detectable on H&E examination. Immunohistochemistry for CMV is specific and sensitive. Histopathology may also play a role in recognising other infections such as amoebiasis and TB and may contribute to the assessment of fissures and fistulas. Confident recognition of dysplasia depends on histology, although endoscopic methods of detection are becoming more reliable. Colorectal carcinoma (CRC) is the most common malignancy to complicate IBD and is more likely than non-IBD CRC to present at a younger age and to be right-sided and is more likely to show signet ring cell differentiation, a peritumoral lymphocytic reaction, and a mucinous phenotype. Small bowel cancer may complicate Crohn’s disease but is rare overall. A lymphoproliferative disorder may complicate IBD, especially in the setting of thiopurine therapy.
Various intestinal and extraintestinal conditions may complicate inflammatory bowel disease (IBD). Intestinal complications include infection, strictures, obstruction, perforation, fistulas, dysplasia, and malignancy. Extraintestinal complications are numerous and include cutaneous disorders, arthropathy, ocular problems, primary sclerosing cholangitis, other forms of liver diseases, and perianal fissures and fistulas.
Histopathology assists the diagnosis of many of the foregoing, but is particularly important for the assessment of cytomegalovirus (CMV) infection, dysplasia, and malignancy in the setting of IBD. Apart from these common concerns, pathologists may play a role in recognising other complications, e.g. amoebiasis, and may contribute to the assessment of fissures and fistulas. Although not discussed in this chapter, liver biopsy contributes significantly to the diagnosis and assessment of primary sclerosing cholangitis and of other hepatic complications of IBD such as drug-induced hepatitis and steatosis.
Infection in Inflammatory Bowel Disease
The relationship between IBD and infection has several aspects (Fact Sheet 23.1).
The main differential diagnosis for a suspected new diagnosis of IBD is acute infective colitis (see Chapter 21).
Specific infective enterocolitides such as tuberculosis, lymphogranuloma venereum, and amoebiasis can mimic IBD clinically and histologically. Distinction is important because treatment and clinical course are very different.
Macroscopic and microscopic features of infection can mimic those of IBD. Examples include granulomas, segmental (discontinuous) colitis, and fulminant colitis (see Chapter 22).
Infection can occur on a background of IBD, i.e. ‘superinfection’ of IBD. Various infective agents can be present and may or may not change the clinical presentation, but some agents, particularly CMV, can modify the picture or the disease course.1
Infection may contribute to the onset of symptoms or relapse in IBD. They may also increase the risk of complications. For example, infective agents and their products may enhance the risk of fibrosis or fistula formation.
Recent IBD research includes studies of dysbiosis, i.e. alteration of the microbial composition of the gastrointestinal (GI) tract. For example, Crohn’s disease may be associated with a reduction in microbial diversity and an increase in specific groups of bacteria.2
Contribute to the initial presentation
Worsen clinical outcome
Contribute to relapse
Confound the presentation
Mimic the features of IBD
Obscure the features of IBD
Contribute to or exacerbate complications, e.g. fistulas
Acute infective colitis is the main differential diagnosis of new IBD
Specific infections may mimic IBD clinically and histologically, e.g. amoebiasis, tuberculosis, yersiniosis, lymphogranuloma venereum, syphilis
Superinfection of IBD that may be clinically significant
Infection may have a role in the pathogenesis and progression of IBD
Altered gut microbiota may be important
Diagnosis of Infection in Inflammatory Bowel Disease
There may already be a clinical suspicion of infection in the setting of IBD prior to the submission of a biopsy. Indeed, biopsy is unnecessary if there is enough evidence of infection from serological, microbiological, and molecular tests. However, histology may detect infection for the first time or may help confirm clinical suspicion (Table 23.1). Histology may also be very useful when the differential diagnosis is between IBD and a clinical mimic such as tuberculosis (TB) or amoebiasis.
|Evidence of infection||Example|
|Microorganisms or inclusions identifiable on H&E|
|Histological patterns suggestive of a specific agent||Granulomas, raising the possibility of TB or other infection|
|Histological or clinical suspicion of an infective agent that prompts the arrangement of additional tests||Ulceration, prompting immunohistochemistry for CMV|
CMV, cytomegalovirus; H&E, haematoxylin and eosin; TB, tuberculosis.
There is a definite association between IBD and CMV superinfection. Superinfection is significantly more common in ulcerative colitis (UC) than Crohn’s disease, and there is more documentation of the association with UC than of any association with Crohn’s disease. However, the role of CMV and its relationship with the clinical behaviour of IBD are controversial.
Reactivation of latent CMV is common in severe ulcerative colitis, e.g. affecting 4.5%–16.6%. The figure may be 25% in patients who require colectomy, and even higher in those with steroid-refractory disease. Factors that probably favour CMV reactivation include the inflammatory state of the mucosa and the actions of immunosuppressive drugs. Corticosteroid use, in particular, increases the risk of CMV.3, 4 In general, CMV reactivation worsens the clinical outcome of UC. Specifically, the presence of CMV associates with a higher rate of steroid resistance.
Evidence is inconsistent, but testing for CMV reactivation in IBD (Practice Points 23.1) is probably appropriate in the following circumstances:
1. Moderate to severe colitis clinically
2. Colitis that is refractory to steroids and/or to other immunosuppressive therapy
3. Ulcerated mucosa on histology, particularly if ulceration is severe or if there are significant amounts of granulation tissue5
4. A suggestion or suspicion of viral inclusions on haematoxylin and eosin (H&E) examination
Pathologists should search routinely for CMV inclusions in IBD biopsies, and with particular care if there is ulceration or severe activity or if there is a clinically severe colitis (Figures 11.2, 18.1, and 23.1–23.6). Inclusions are easy to miss. Typically, there is an ‘owl’s eye’ appearance as a result of an eosinophilic intranuclear inclusion with a variable degree of clearing (Figure 23.1), while cytoplasmic inclusions may have a less distinctive ‘smudged’ purplish or basophilic appearance that is subtler than the classical owl’s eye (Figure 23.2). Inclusions are most often in endothelial cells but may be in smooth muscle cells, neurons or, less commonly, epithelial cells. The type of cell that they occupy is often not obvious (Figures 23.1 and 23.2).
Colitis refractory to drug therapy
Moderate or severe colitis clinically and/or histologically
Ulcerated mucosa on endoscopy and/or histology
‘Owl’s eye’ eosinophilic intranuclear inclusion
‘Smudged’ purplish cytoplasmic inclusions
May be in endothelial cells, smooth muscle cells, neurons, or (less often) epithelial cells
Ulcers, especially if deep and with abundant granulation tissue
Ischaemic-type features, e.g. endothelialitis, thrombi
Frequent crypt epithelial cell apoptoses
Degenerate endothelial, stromal, and epithelial cells
Figure 23.1 Typical eosinophilic inclusion of CMV surrounded by a clear area, giving rise to an ‘owl’s eye’ appearance.
Figure 23.2 (A–C) CMV inclusions may have a smudged purple appearance. The type of cell containing the inclusion may not be obvious.
Figure 23.3 (A) Ulceration and granulation tissue on a background of ulcerative colitis should lower the threshold for consideration of cytomegalovirus (CMV).
(B) Immunohistochemistry shows a few CMV inclusions.
Figure 23.4 Ulceration can induce cytological atypia that may mimic the appearances of CMV inclusions.
Figure 23.5 Sparse ganglion cells are present in the mucosa (arrow) and can resemble CMV inclusions.
Figure 23.6 (A, B) Immunohistochemical expression of CMV. Positive staining is homogeneous and clearly separated from the negatively staining surrounding tissue as in these examples. Artefactual staining is usually less intense and less well circumscribed and may not be in the plane of the remainder of the section.
Other changes caused by CMV include ulcers that may be deep and may have an abundant granulation tissue component (Figure 23.3A and B). There may be ischaemic-type features such as endothelialitis, small thrombi, and necrosis.6 An increase in the number of crypt epithelial cell apoptoses may occur in CMV but may also be a feature of IBD, drug-induced colitis, other infections, and, in particular, graft-versus-host disease (GvHD). Indeed, support for a diagnosis of GvHD relies partly on the detection of an increased number of crypt epithelial cell apoptosis and accordingly the exclusion of CMV is often appropriate in this setting before attribution of apoptosis to GvHD.
Various cells can show changes resembling CMV effect. Degenerate endothelial, stromal, and epithelial cells can show cytomegaly and nuclear abnormalities (Figure 23.4), often secondary to ulceration and/or severe inflammation. Ganglion cells are a potential mimic of CMV inclusions, either within ulcerated areas of the mucosa (although the mucosa rarely contains ganglion cells) (Figure 23.5) or in the underlying inflamed submucosa (where ganglion cells are usually present).
The inflammatory changes of CMV may obscure the histological features of IBD. This is important when a patient presents with intestinal CMV infection de novo and has no known existing diagnosis of IBD or even of GI disease. In this setting, diagnosing underlying IBD and distinguishing between UC and Crohn’s disease can be difficult. Features suggesting pre-existing IBD include basal plasmacytosis and crypt architectural distortion, although CMV-induced ulceration may itself induce crypt distortion and a severe inflammatory infiltrate that includes plasma cells. Therefore, assessment of the adjacent non-ulcerated mucosa may help make the distinction between those patients with and without underlying IBD. Clinically and on imaging, superimposition of CMV on UC may mimic acute exacerbation of UC or may cause changes resembling Crohn’s disease. Therefore, interpretation of the histological appearances and diagnosis of underlying disease should be cautious if CMV is present in a patient with no known past history of GI disease.
Confirmation of Cytomegalovirus
The common tests for CMV on histology samples are H&E staining and immunohistochemistry. Typical H&E inclusions are highly specific, e.g. 92%–100%, but their sensitivity is relatively low. Immunohistochemistry has a higher rate of sensitivity, e.g. 78%–93%. CMV DNA real-time polymerase chain reaction (R-T PCR) may be useful and is the most accurate way to quantify the viral load but requires a separate sample (Practice Points 23.2).4
Specific, e.g. 92%–100%
Fairly low sensitivity
Accurate for quantifying viral load
In practice, pathologists often use immunohistochemistry, as it is more sensitive than H&E. However, the interpretation of immunohistochemistry can be difficult (Practice Points 23.3). Well-demarcated homogeneous staining of a cell within the plane of section is usually a requirement for positivity (Figures 23.3B, 23.6A and B). The report should mention obvious artefactual staining and should record it as negative. CMV inclusion numbers are often low in the setting of IBD, and sparse immunopositive cells may be difficult to distinguish from artefact. In equivocal cases, a repeat stain may be useful. If there are equivocal CMV inclusions on H&E and immunohistochemistry is negative, the pathologist should avoid a diagnosis of CMV. If CMV immunohistochemistry is unequivocally positive in the absence of obvious inclusions on H&E, a diagnosis of CMV infection is advisable.
Clinical suspicion of CMV
Moderate/severe colitis with significant endoscopic or histological activity
Viral inclusions seen or suspected on H&E
Ulceration and/or granulation tissue on H&E
Other histological features that could reflect CMV, e.g. ischaemic-type changes, vascular changes
Artefactual immunohistochemical staining can occur.
To diagnose positivity, the staining should be
Within the plane of section
Within a structure that is likely to be a cell
Intense and homogeneous (ideally)
Repeat the procedure if there is doubt
Compare with H&E, even if inclusions were initially not apparent, because diagnosis is more robust if both H&E inclusions and immunopositivity are present
If immunohistochemistry is negative: record as negative unless there are classic H&E inclusions
If immunohistochemistry is unequivocally positive: record as positive regardless of H&E appearances
Report sparse immunopositive cells, even though their clinical significance is uncertain
Give an indication of the number and density of immunopositive cells
Semiquantitative immunohistochemistry entails reporting the total number of infected cells and the number of biopsy fragments that contain immunopositive cells. According to some reports this helps predict clinical outcome, although other studies suggest no correlation between immunohistochemistry and the true viral load.4 In general, an indication by the pathologist of the number and density of inclusions is advisable because it may be helpful to the clinical teams.
If CMV is present in IBD mucosa in the appropriate clinical setting, treatment with anti-viral drugs, e.g. ganciclovir, is usual.3, 5 Remission rates following treatment of CMV are as high as 71%–86%, often allowing cessation of steroids and avoidance of proctocolectomy.
If sought, Epstein–Barr virus (EBV) is detectable in the intestinal mucosa of a significant number of IBD patients. Steroid use increases the likelihood of EBV infection. Mucosal loads of EBV are higher in refractory than non-refractory IBD and may correlate with depth of inflammation, ulceration, and future need for colectomy, while EBV might play a role in the activation or maintenance of inflammation.7–9 Ganciclovir and other antiviral agents are not effective, and there is no treatment that lowers EBV levels and improves clinical outcome. According to one report, mucosal EBER-1 levels may have the potential in clinical practice to predict refractoriness of UC.9 Currently, assessment for EBV is not included in routine clinical or pathological assessment.
Entamoeba histolytica is the cause of amoebiasis, a very common infection worldwide. Entamoeba dispar is a similar parasite that probably has few or no pathological effects on the bowel. Entamoeba histolytica causes intestinal amoebiasis, amoebic dysentery, and liver abscesses.10 In Europe and North America it occurs mainly in high-risk patients, e.g. those infected with HIV or those who have travelled to areas with a high prevalence of amoebiasis. The colorectal sites most commonly affected, in descending order of frequency, are the caecum, the remaining right colon, the rectum, and the sigmoid colon. The clinical presentation ranges from asymptomatic to mild symptoms to a severe colitis that can be fulminant. The latter can resemble fulminant IBD.
Histologically there may be surface exudate or debris, sometimes haemorrhagic, containing trophozoites 25–40 micrometres in diameter with a small, round, pale purple and usually eccentric nucleus that has a central karyosome resembling a nucleolus, and pale foamy cytoplasm containing ingested red cells (Figures 18.20 and 22.7). The appearance may be similar to that of a macrophage, but the nucleus tends to be smaller. Periodic acid–Schiff (PAS) and trichrome stains are positive (although this may also occur in macrophages) and there is negative immunohistochemical staining for macrophage markers such as CD68. Mucosal changes in the early stages include a neutrophil infiltrate with typical organisms on the surface. Later there are often ulcers, and these may be deep. The changes may resemble IBD, and there is crypt architectural distortion and basal plasmacytosis in some cases (Figure 22.7B–D) (see Chapter 22).
The relationship with IBD has several aspects, including amoebiasis mimicking IBD clinically and pathologically (see Chapter 22); IBD mimicking amoebiasis clinically and pathologically; and coexistence of IBD and amoebiasis (Figure 23.7 and Table 23.2).
(A) A colonic biopsy from a patient with known Crohn’s disease showing chronic inflammatory changes and a prominent surface exudate.
(B) The exudate contains many amoebic trophozoites. The inflammation in this example could be a result of IBD, amoebic infection, or both.
|Amoebiasis mimicking IBD clinically|
|Amoebiasis mimicking IBD pathologically|
|IBD mimicking amoebiasis|
|Coexistence of IBD and amoebiasis|
|Ulcerative colitis as a sequel to amoebiasis||Very rare reports exist|
The prevalence of coexistent amoebiasis in patients with established IBD is probably higher than the prevalence in the general population. However, this depends on geographic location, and is a more important consideration in areas with a high prevalence of amoebiasis than in those with a low prevalence. In a Turkish study, Entamoeba species (both Entamoeba histolytica and Entamoeba dispar) were detectable in 10% UC patients, 3.3% Crohn’s disease patients, and 1.9% of the general population.11 In areas with a high prevalence, empirical anti-amoebic therapy may be worth considering for refractory IBD.
If amoebiasis superinfects known IBD, the relative contributions of each aetiology to the histological changes may be difficult or impossible to determine (Figure 23.7). Similarly, categorisation of the histological severity and activity of the underlying IBD is often unreliable in this circumstance. If amoebiasis coexists with previously undiagnosed IBD, a histological diagnosis of IBD and accurate interpretation will be even more difficult.
There are very rare reports of the development of UC as a sequel to amoebiasis. In two patients, apparently successful treatment of amoebic dysentery was followed by an illness with the clinical, pathological, and radiological characteristics of UC.12
Giardia lamblia may cause giardiasis in both immunocompetent and immunosuppressed people. The frequency of detection of Giardia is considerably higher in duodenal biopsies than in biopsies from other sites such as stomach, ileum, or colon. However, trophozoites are occasionally found at non-duodenal GI sites in the absence of duodenal involvement.13 If infected, the small bowel mucosa is often normal or may show mild or moderate villous blunting. A reliable histological diagnosis requires typical trophozoites. These have a ‘streaming’ appearance in the lumen, with an arched / crescent moon shape in tangential section and a binucleate pear shape en face (Figure 14.1).14, 15 Immunohistochemistry for CD117 is positive.16 There is limited evidence of an increased prevalence of Giardia infection in Crohn’s disease, and the symptoms of intestinal giardiasis may mimic an acute exacerbation of known Crohn’s disease. Possibly, Giardia causes an epithelial barrier dysfunction resulting in clinical features similar to those of IBD.17, 18