Biopsy assessment of chronic idiopathic inflammatory bowel disease (IBD) is important for diagnosis, classification, determination of activity, documentation of anatomical extent and distribution, detection of complications, and diagnosis and grading of dysplasia. In addition, it may predict treatment response, clinical behaviour after therapy, and risk of future dysplasia. Microscopic features favouring IBD over non-IBD in biopsies include basal plasmacytosis and crypt architectural changes. Typically, UC extends from the rectum proximally and in a continuous fashion, while Crohn’s disease is discontinuous with intervening areas of sparing. Microscopic features that distinguish most reliably between Crohn’s disease and UC include granulomas (strongly favouring Crohn’s disease) and the distribution of architectural abnormalities / chronic inflammation between sites and within sites. However, discontinuity may also occur in UC. A caecal “patch”, i.e. a discontinuous area of caecal inflammation, is common in UC. Rectal sparing and other forms of discontinuity are more frequent in longstanding UC than new UC. The label IBD unclassified (IBDU) is available for difficult cases. In very early IBD (i.e., symptoms for < 6 weeks), the architectural changes are often absent. IBD is common in patients with primary sclerosing cholangitis but may have unusual features. In summary, histological assessment in the light of the clinical and endoscopic findings plays an important role in optimising management of IBD.
Inflammatory bowel disease (IBD) is becoming more common in many countries, particularly in Asia, Africa, and South America.1 Information about its pathogenesis, clinical features, and prognosis continues to accumulate. At the same time, histopathology makes an increasingly important contribution to the diagnosis and management of IBD (Table 21.1).2 Biopsy assessment may play a major role at initial presentation, helping to confirm or refute the diagnosis of IBD and assisting with subclassification as ulcerative colitis (UC) or Crohn’s disease. Other roles of biopsy include assessment of histological activity, assessment of the distribution of microscopic disease, recognition of complications (e.g. cytomegalovirus [CMV] infection), diagnosis and grading of dysplasia, and exclusion of carcinoma and other malignancies. Despite recent advances in endoscopic techniques, histology is currently the only reliable way to diagnose and assess dysplasia.
|Role of histology||Comment|
|Diagnosis of IBD||New ≫ longstanding IBD|
|Classification of IBD as UC or Crohn’s disease||New> longstanding IBD|
|Activity||No agreement on scoring|
|Distribution of disease|
|Dysplasia: presence and grade||Longstanding> new IBD|
|Malignancy||Longstanding> new IBD|
|Serrated polyps (hyperplastic polyp, sessile serrated polyp, traditional serrated adenoma)||Currently managed as for non-IBD serrated polyps|
|Serrated epithelial change|
|Prediction of response to drug therapy / assessment of response to drug therapy / determination of completeness of remission / prediction of need for further therapy after drug treatment|
|Prediction of risk of dysplasia||Not a formal aspect of clinical practice|
There is considerable interest in the contribution of histology to the assessment of drug therapy in clinical trials. Individual histological features such as crypt abscesses, mucin depletion, basal plasmacytosis, and lamina propria eosinophils may help predict the effectiveness of drug therapy, but study findings are inconsistent.3, 4 Similarly, microscopic assessment after administration of drugs may help assess the state of the mucosa. For example, categories of post-treatment biopsies may include normal (‘histologic normalisation’), quiescent disease, and active disease, or the study may use one of many scoring schemes to produce a histological grade of inflammation or grade of activity.5, 6 The aims are to determine the degree of mucosal healing and to determine the likelihood that further therapy will be necessary.7 In addition to the assessment of drug therapy, the presence and severity of histological activity may be markers for the risk of subsequent dysplasia.8
There are two types of IBD: UC and Crohn’s disease. When IBD is difficult to classify, the term ‘IBD unclassified’ (IBDU) is an option. The overused term ‘indeterminate colitis’ refers exclusively to IBD in a resection specimen that is unclassifiable despite comprehensive assessment in the light of all other investigations. The term indeterminate colitis is not applicable to biopsy reporting (or endoscopic reporting).
Ulcerative colitis usually presents in late childhood and young adulthood (ages 10–40 years with a peak incidence at 20–29 years). There may be a smaller second peak after age 509 but other causes of similar endoscopic and histological changes, e.g. diverticular disease, require exclusion in older patients before acceptance of a diagnosis of IBD. Presenting symptoms usually include blood in the stools (>90% patients). Other symptoms may include loose stools, reduced stool consistency, urgency, tenesmus, and crampy abdominal pain.9 Symptom severity correlates broadly with severity and extent of disease. Patient usually have symptoms for weeks or months before requesting a medical opinion, which is fortuitous for the histopathologist and for reliability of biopsy interpretation because the histological changes that are typical of IBD often take several weeks or months to develop.9 The most important differential diagnosis for new disease is infectious colitis, which also presents with diarrhoea. Symptoms of more than 6 weeks’ duration favour IBD over infection.9
Macroscopically, UC always affects the large bowel. Disease usually involves the rectum at presentation and may extend proximally into the colon for a variable distance, with a sharp cut-off between normal and abnormal. The Montreal classification recognises three broad groups of distribution of UC: rectal, left sided, and extensive (the latter including pancolitis) (Figure 21.1A).10 Involvement of the mucosa is typically continuous, but there are exceptions (see later).
(A) Classification of ulcerative colitis is as rectal, left-sided, or extensive.
(B) Classification of Crohn’s disease is as ileal, colonic, ileocolonic, or upper gastrointestinal, the latter usually coexisting with ileal and/or colonic disease.
Endoscopically UC may show erythema, deceased vascular pattern, hyperaemia, granularity, a haemorrhagic appearance, friability, erosions, and ulceration.9 The disease is mainly mucosal, but deeper layers may be involved if there is severe inflammation and/or ulceration. Fissure ulcers can occur but are less common and less deep than in Crohn’s disease and do not involve more than 50% of the depth of the muscularis propria.11 Fibrosis is less common and typically less severe than in Crohn’s disease. Fistulas, if present, are simple rather than complex and are considerably less common than in Crohn’s disease.9
Crohn’s disease affects a slightly older population than UC but is also a disease of older children and young adults with a possible later ‘peak’ after age 50. Symptoms at presentation are heterogeneous, and include diarrhoea, abdominal pain that may be crampy, and weight loss. Blood in the stool is less common than in UC but occurs in up to 50%.12 Any site from the mouth to the anus can be involved, but small bowel and/or large bowel disease are usual and anal disease is common. Involvement of the bowel is classically discontinuous, with abnormal areas or ‘skip’ lesions between which there are macroscopically normal segments.
Endoscopically, Crohn’s disease produces longitudinal ulcers, aphthoid ulcers, confluent areas of ulceration, cobblestoning of the mucosa (secondary to ulceration), and discontinuous involvement or ‘skip’ lesions. Strictures can occur. Perianal disease includes fissures and fistulas, the latter occurring in 10% at presentation. Fistulas may also develop between two parts of the bowel or between the bowel and other structures. An ‘inflammatory’ phenotype is usual in early Crohn’s disease, while a pattern characterised by fibrosis and strictures may develop later. For example, 40% of patients have significant fibrosis after 8–10 years and 70% of patients develop fibrosis, fistulas, or obstruction at some stage during the course of the disease.13
Crohn’s disease is typically transmural, affecting all layers. There are very rare exceptions where the changes are confined to the mucosa, sometimes termed ‘ulcerative colitis-like Crohn’s disease’ or ‘superficial Crohn’s colitis’.14
Classification of anatomical distribution in the bowel is less simple for Crohn’s disease than for UC. The Montreal system classifies Crohn’s disease according to age at diagnosis (<17, 17–40, or>40 years), anatomical location (ileal, colonic, ileocolonic, or upper GI tract) (Figure 21.1B), and behaviour (non-stricturing and non-penetrating, stricturing, or penetrating). ‘Isolated colonic Crohn’s disease’ or ‘isolated Crohn’s colitis’ occurs at presentation in 14%–32%, and may be clinically and pathologically distinct from classical cases that involve both the large bowel and the small bowel or affect the small bowel only.14
The term inflammatory bowel disease (IBD) should be restricted to idiopathic cases with clinical, imaging, endoscopic, and pathological features of IBD. Histological features whose absence should cast doubt on a new diagnosis of IBD include basal plasmacytosis and architectural changes (see later). Similarly, the term IBD is not appropriate for any other inflammatory condition. Furthermore, IBD is never an ‘umbrella’ term covering both IBD and other non-IBD colitides such as microscopic colitis and diverticular colitis. The term ‘chronic idiopathic inflammatory bowel disease’ (CIIBD/CIBD) is more precise and probably preferable to the term IBD but is cumbersome and less popular.
There is reclassification of a minority of patients who initially have a diagnosis of UC or Crohn’s disease, e.g., reclassification of 10%–15% of Crohn’s disease as UC within 1 year in some reports.12 Furthermore, the classification can change considerably later, particularly in children. In one paediatric study, a diagnosis of UC on the basis of resection specimen assessment was subsequently changed to Crohn’s disease in 24% of cases after a median follow-up of 7.6 years.15
Diagnosis of IBD depends on a combination of clinical, endoscopic, imaging, and microscopic findings. Biopsy helps to confirm IBD, particularly when distinction from other causes of symptoms is difficult clinically and may be useful for its classification as UC or Crohn’s disease. However, histological assessment is complementary to other investigations and is rarely diagnostic in isolation.9, 16 If clinical details are not available, the pathologist should record their absence, interpret the histological material with caution, and attempt to ask the relevant questions (Practice Points 21.1). Sampling should be adequate, particularly at initial diagnosis, and should include at least two biopsy fragments from the ileum and two fragments or more from each of at least four sites in the large bowel.9 The anatomical site of origin of each biopsy requires clear identification. Compared to single-site biopsies, multiple-site biopsies are particularly helpful for distinguishing Crohn’s disease from UC and for confirming Crohn’s disease.12
Does the patient have suspected new IBD or longstanding treated IBD?
If new, what are the symptoms and their duration?
Evidence of previous surgery to the bowel?
Raised lesions or polyps?
Endoscopist’s diagnosis or provisional diagnosis?
Distribution of Disease
Assessment of distribution of disease is very helpful for distinguishing UC from Crohn’s disease. Initial assessment of distribution depends mainly on endoscopy but also relies, particularly for small bowel disease, on imaging. The correlation between endoscopic and histological distribution is often poor. The pathologist should assess distribution between different anatomical sites, between biopsies from the same site, and within each biopsy. The first of these is the most useful.
Involvement or non-involvement of each site by chronic changes and by acute inflammation should be determined. Chronic changes include a chronic inflammatory cell infiltrate and architectural abnormalities and are a more reliable indicator than acute inflammatory changes of the involvement of a particular site by IBD (see Fact Sheet 21.1 for terminology).
Focal chronic inflammation: well-defined foci of increased chronic inflammatory cell density in a normocellular background (Figure 21.2A)
‘Continuous between sites’ or ‘discontinuous between sites’ avoids confusion with the aforementioned terms
‘Segmental’ is another term for discontinuity between sites
Endoscopists also use these terms
(A) Diagrammatic representation of diffuse, patchy, and focal distribution of chronic inflammation within and between biopsies from a single site.
(B) A biopsy from a patient with a new diagnosis of UC showing diffuse lamina propria chronic inflammation and diffuse architectural changes that include crypt distortion and crypt atrophy.
(C) A biopsy from a patient with Crohn’s disease showing non-diffuse changes, including patchy chronic inflammation (i.e. a variable density of inflammatory cells) and focal crypt branching (arrow).
New Inflammatory Bowel Disease: Diagnostic Value of the Assessment of Distribution
Determination of the distribution of changes between sites, within sites, and within biopsies is very useful for distinguishing new UC from new Crohn’s disease. It may also help distinguish IBD from non-IBD (Fact Sheet 21.2).17 IBD is classified more accurately if multiple sites are sampled than if sampling is limited to distal sites or to a few random sites, particularly if the diagnosis is Crohn’s disease.12, 17–21 Gastroenterologists should be encouraged, when assessing suspected new IBD, to do full colonoscopy, to biopsy each site adequately, and to use a reliable method for informing the pathologist of the site of origin of each biopsy.
Continuous involvement from the rectum proximally favours UC over Crohn’s disease
Discontinuous (segmental) abnormalities favour Crohn’s disease over UC
There are several exceptions (Table 21.2)
Diffuse architectural changes favour UC while focal or absent architectural changes (non-diffuse) favour Crohn’s disease
Diffuse chronic inflammation favours UC over Crohn’s disease but is less discriminant than diffuse architectural changes
Non-diffuse chronic inflammation favours Crohn’s disease over UC but is less discriminant than non-diffuse architectural changes
Abnormalities that are more severe distally than proximally tend to favour UC
Right-sided predominance tends to favour Crohn’s disease
Solely distal or left sided involvement extending from the rectum favours UC over Crohn’s disease
|Periappendiceal or caecal patch in UC (Figure 21.3A)|
|Rectal sparing in UC|
|Does discontinuity in UC suggest Crohn’s disease?|
Figure 21.3 Caecal patch in ulcerative colitis (UC). (A) Typical distribution of caecal patch, extending from the periappendiceal mucosa for a variable distance into the caecum and ascending colon. (B, C) Inflammation of the caecum (B) and rectum (C) in a patient with UC, between which there was normal mucosa. The changes in the caecal patch (B) tend to be similar to but milder than the distal changes (C).
Exceptions: Discontinuous Change in Ulcerative Colitis
UC is typically continuous between anatomical sites. However, there are several exceptions (Table 21.2). Discontinuity between sites is a recognised feature of some cases of new UC and is common in longstanding UC. Longstanding UC often also shows non-diffuse changes within sites and within biopsies.
Diagnosis of Inflammatory Bowel Disease
Histological Features That May Occur in Inflammatory Bowel Disease
Some features occur commonly in IBD (Fact Sheet 21.3). A few of these, e.g. crypt distortion and basal plasmacytosis, are strong pointers to a diagnosis of IBD. Others, e.g. cryptitis and crypt abscesses, are common in IBD but do not distinguish it from other forms of colorectal mucosal inflammation. Some features, e.g. apoptosis and intraepithelial lymphocytes, are more characteristic of other conditions but can also occur in IBD. Table 21.3 lists the definitions of relevant features.
Irregular/villiform mucosal surface (Figure 21.4A)
Increase in lamina proprial lymphocytes (non-aggregated)
Lymphoid aggregates / lymphoid follicles (Figure 21.6)
Giant cells, especially basal (Figure 21.9E)
Increase in the number of crypt epithelial cell apoptoses (Figure 21.10)
Increase in the density of intraepithelgial lymphocytes
Lamina propria acute inflammation
Increase in lamina propria eosinophils
Lamina propria fibrosis
Lamina propria hypocellularity
|Villiform/irregular mucosal surface||21.4A||Separation of crypts yielding surface contour of broad villus-like projections.||9,28|
|Crypt distortion||May include branching, loss of parallelism, irregularity, tortuosity, dilatation, and variation in shape and size.||9,23,27,29–33|
|Crypt branching||21.4C||Branching between mucosal hillocks / at innominate grooves is normal||19,23,29,31,33,34|
|Basal plasmacytosis||21.5||Basal plasma cells are normal in the caecum and ascending colon.||9,23,28,29|
|Basal lymphoid aggregates||21.6||One or two transmucosal lymphoid nodules are acceptable in normal mucosa; they can extend across the muscularis mucosae. Pathological aggregates may be difficult to distinguish from normal.||9,29,31,33|
|Cryptitis||21.7A||Neutrophils in crypt epithelium.||23,27,29|
|Crypt abscess||21.7B||Neutrophils in crypt lumen. Minimum number of neutrophils is not specified.||Often located near base of crypt.||19,23,27,29|
|Granulomas||21.9||Discrete collection of at least five epithelioid macrophages.||Consider crypt rupture as a cause. Serial sections may help with this.||29,31,34|
|Mucin depletion||21.5C||Unequivocal reduction of goblet cell mucin in crypt or surface epithelium.||20,33,34,36|
|Paneth cell metaplasia||21.8||Pyramidal crypt epithelial cells with supranuclear eosinophilic granular cytoplasm.||Paneth cells normal in caecum and right colon, and sparse distal to splenic flexure.||31,33,37|
|Crypt epithelial cell apoptoses||‘Round or ovoid structures apparently bounded by a discrete membrane and enclosing aggregates of pyknotic intensely basophilic fragments of nuclear chromatin often surrounded by a thin mantle of cytoplasm’.38||May occur in normal mucosa in small numbers.|
(A) Villiform surface change in the large bowel mucosa in UC, reminiscent architecturally of small bowel mucosa.
(B) UC biopsy showing mild crypt distortion (crypt irregularity and loss of parallelism) and crypt atrophy (crypt shortening and a slight increase in intercrypt spaces).
(C) UC biopsy showing severe crypt atrophy and crypt distortion (branching and malorientation).
(A) Early inflammatory bowel disease with basal plasmacytosis and relatively mild crypt changes.
(B) A particularly dense basal infiltrate of plasma cells.
(C) ‘Crypts with their feet in pools of plasma cells’ describes some examples of basal plasmacytosis. Eosinophils are also quite numerous in this case. There is severe epithelial mucin depletion, a feature that favours ulcerative colitis over Crohn’s disease.
Figure 21.6 Basal lymphoid aggregates are a feature of inflammatory bowel disease but also occur in normal mucosa.
(A) Neutrophil cryptitis characterised by neutrophils in the crypt epithelium. Mucin depletion is a frequent accompaniment.
(B) A crypt abscess, defined as neutrophils in the crypt lumen. Neutrophils are also present in the epithelium. This crypt abscess is large and shows crypt dilatation and epithelial attenuation, but many crypt abscesses are small and some are subtle and easy to miss.
Figure 21.8 Numerous Paneth cells in a rectal biopsy from a patient with longstanding treated UC (arrows). Paneth cells may be present in normal mucosa but are more likely to be abnormal and to represent metaplasia if distal to the splenic flexure than if proximal. In this biopsy, the crypts are atrophic and distorted and the lamina propria shows mild fibrosis but there is no basal plasmacytosis. This combination is quite common in longstanding UC.
(A) Granulomas in Crohn’s disease show no necrosis, are often small, and typically contain no giant cells.
(B) Small granulomas beneath the surface epithelium are quite common in Crohn’s disease and are easy to miss if not sought.
(C) A loose granuloma with multinucleate cells can occur close to an intestinal abscess or fistula, regardless of the underlying cause of the abscess/fistula.
(D) Larger granulomas with giant cells can occur in Crohn’s disease, but if numerous and confluent might raise the alternative possibility of tuberculosis. Another important cause of granulomas is crypt rupture, which is a consideration if a granuloma is mucosal but not if it is submucosal (as in this example).
(E) Basal giant cells favour inflammatory bowel disease over other causes but may be secondary to crypt rupture or foreign material.
Figure 21.10 Crypt epithelial cell apoptoses (arrows) are sometimes numerous in inflammatory bowel disease (IBD) but are not usually regarded as a typical feature of IBD and are also found in a variety of other conditions, e.g. drug-induced colitis, infection, and graft-versus-host disease.
Features Favouring Inflammatory Bowel Disease over Other Causes
This section and subsequent sections allocate histological features to a category of reliability, based on the evidence available.
Reliable Features Distinguishing Inflammatory Bowel Disease from Other Causes
IBD shares many histological abnormalities with other inflammatory diseases. Therefore, knowledge of the features that discriminate reliably between IBD and non-IBD causes is useful. Studies comparing IBD mucosa with normal are informative. However, distinction of IBD from other causes of inflammation is a more common challenge in practice. Most studies compare IBD with acute infective colitis (acute self-limiting colitis), while some compare it with a mixture of ‘non-IBD’ colitides.
Acute infective colitis is the most common clinical mimic of new IBD, typically presenting with diarrhoea and associated symptoms. Unlike IBD, the symptoms usually last for less than six weeks and do not recur. Common infections such as Campylobacter, Salmonella, and Shigella are often responsible.28, 31, 35, 39 Histological distinction between infective colitis and new IBD may be necessary because infection sometimes persists for weeks or months, resulting in a clinical picture that suggests IBD or is difficult to interpret.40
Basal Plasmacytosis and the Plasma Cell Gradient
The normal mucosa has a plasma cell gradient, with a higher density of plasma cells in the upper third than in the lower third (Figure 21.5). Basal plasmacytosis requires the density of plasma cells in the lower third of mucosa to be similar to or greater than the density in the upper third. In other words, there is loss of the plasma cell gradient (Practice Points 21.2). Excess plasma cells are easiest to detect beneath crypt bases but also occupy the deep part of intercrypt spaces.9, 23, 28, 29 Schumacher refers eloquently to ‘crypts with their feet in pools of plasma cells’ while acknowledging the variations on this picture.28 Plasma cells are easier than lymphocytes to identify and quantify.
Strongly supports a diagnosis of IBD
Implies loss of the normal mucosal plasma cell gradient
May be diffuse or non-diffuse
In Crohn’s disease may be focal (Figure 21.11)
In isolation is not specific for IBD and can also occur in diverticular colitis, diversion colitis, and microscopic colitis, and less often in chronic infective disorders, radiation damage, and other settings
Often subsides or disappears in longstanding/treated IBD
Is a requirement for a confident histological diagnosis of new IBD
Caution is advisable in the right colon, where basal plasma cells may be normal
Basal plasmacytosis in initial biopsies is the strongest predictor of a future diagnosis of IBD. Demonstration of its value as a discriminator between IBD and non-IBD is consistent.28–30 Basal plasmacytosis may subside in chronic/treated disease but is usually present at least focally at diagnosis. Depending partly on whether the IBD is of UC or Crohn’s type, basal plasmacytosis may be diffuse, extensive, patchy, or focal. A search to exclude focal basal plasma cells is advisable before rejection of a diagnosis of IBD.
There are some caveats. Basal plasma cells may occur in the normal right colon, particularly the caecum, and may be sufficiently dense to cause loss of the plasma cell gradient (Figure 17.9). In addition, basal plasmacytosis is not unique to IBD and is not diagnostic in isolation. It can occur in diverticular colitis, diversion proctocolitis, and less often in other chronic mucosal inflammatory conditions such as chronic infection, drug-induced colitis, radiation colitis. Lamina propria plasmacytosis is common in microscopic colitis and may lead to loss of the plasma cell gradient, but the architectural changes and other features of IBD are not usually present.
Abnormal Architecture/Crypt Distortion/Crypt Branching
Crypt distortion is a broad term for abnormally shaped crypts and includes branching, loss of ‘parallelism’, irregularity of outlines, tortuosity, dilatation, and inter-crypt variation in shape and size (Figure 21.4). One definition of crypt branching requires the presence of two or more branched crypts in a well oriented biopsy (without mention of the size of the biopsy). Branching between mucosal hillocks/at innominate grooves is normal.19, 23, 29, 31, 33, 34 Branching may be vertical or horizontal, but this distinction probably has little value in routine diagnostic practice (Figure 21.12).
Figure 21.12 Crypt branching is sometimes classified as vertical (as in this biopsy) or horizontal. The distinction has limited diagnostic value.
Assessment of crypt architecture is unreliable in some circumstances. Adjacent to crypt abscesses and lymphoid aggregates/lymphoid follicles there may be distortion of the crypts. The anal transition zone/columnar cuff is unsuitable for assessment, because crypt distortion is quite common in this location.9, 23, 29–33, 41
Demonstration of the discriminant value of crypt architectural abnormalities for IBD is highly consistent between studies.28–32, 35, 39, 42 However, crypt changes are more likely than basal plasmacytosis to be present in non-IBD colitides. They are usually less extensive and less severe in non-IBD than in IBD. Perhaps surprisingly, interobserver variability for the assessment of crypt architectural abnormalities is higher than for the assessment of basal plasmacytosis.2
Diagnosis of crypt atrophy usually requires crypt shortening, i.e. an increase in the gap between the base of the crypt and the muscularis mucosae (Figure 21.4B and C). Normally, this gap is small or non-existent. Wider spacing of crypts is further evidence of atrophy, and definitions include ‘more than one crypt diameter between crypts’. Caution is necessary in some areas of the mucosa, e.g. near lymphoid follicles and near aggregates of muciphages, and there may be a larger space beneath or between crypts in the anal transition zone/columnar cuff, the rectum and the caecum (Figure 17.7C).9, 20, 23, 29, 31, 32, 34, 35, 39 Crypt atrophy is a reliable marker of IBD. In one report, a count of less than five crypts per mm was unique to IBD. However, moderate or severe atrophy may occasionally occur in other settings. There is moderate interobserver variability for the assessment of crypt atrophy.
Irregular/Villous Mucosal Surface
This refers to an undulating or broadly villiform surface that is partly a consequence of widening of crypt mouths (Figure 21.4A). Separation of crypts yields a surface consisting of broad villous projections. It is a reliable marker of IBD with moderate interobserver variability.9, 28, 29, 31, 32, 35, 42 It is generally more common and better developed in chronic disease than in new IBD.
Fairly Reliable Features Distinguishing Inflammatory Bowel Disease from Other Causes
A granuloma is a discrete collection of macrophages. In the setting of IBD, one definition refers to a discrete collection of at least five epithelioid macrophages (Figure 21.9).29, 31 Granulomas are strongly discriminant between Crohn’s disease and UC. They are also a pointer towards IBD, largely because they discriminate between Crohn’s disease and many other causes of colorectal inflammation. Therefore, a GI mucosal granuloma should always raise the possibility of Crohn’s disease unless there is another suspected or established cause of granulomas. Many other causes of granulomas require exclusion, e.g. infection, drugs, sarcoid, foreign material, and rarer conditions (Fact Sheet 21.4).43 Overall, granulomas favour IBD over infection but are of no value in distinguishing UC from infective colitis.28, 29, 31, 32, 42 ‘Cryptolytic granulomas’ are a result of rupture of a crypt abscess or cryptitis into the adjacent lamina propria, and can occur in any setting (see later) (Figure 21.13A).
Parasites, e.g. Schistosoma spp.
(A) A cryptolytic granuloma adjacent to a ruptured crypt in a patient with known ulcerative colitis (UC). Cryptolytic granulomas are less discriminant than non-cryptolytic for Crohn’s disease versus UC but are probably more likely to occur in Crohn’s disease (image courtesy of Dr V Simanskaite, London, UK).
(B) A biopsy from a patient with established UC containing a granuloma (arrow). This lies close to a crypt and could be a consequence of crypt rupture. However, because it is not demonstrably cryptolytic it raises the alternative possibility of Crohn’s disease.
Basal Giant Cells
Basal multinucleate giant cells (Figure 21.9E) are more common in IBD than in infective colitis, but many studies do not assess them as a separate feature. Therefore, there is sparse information about giant cells.31, 32, 42 In isolation, they probably have little discriminant value because they could alternatively reflect crypt rupture or a reaction to foreign material.
Basal Lymphoid Aggregates
Lymphoid aggregates are collections of lymphocytes, usually nodular (Figure 21.6). If they have a germinal centre, the term ‘lymphoid follicle’ is preferable. They usually occupy the deep part of the lamina propria and may lie between the muscularis mucosae and crypts. In normal mucosa, there may be a small number of transmucosal lymphoid nodules. They can also extend across the muscularis mucosae (Figure 17.10A).9, 29, 31, 33. Lymphoid aggregates are more numerous in IBD than in normal mucosa, but in the absence of other histological abnormalities they have limited discriminant value because pathological aggregates are difficult to distinguish from normal aggregates. There is no definition of the acceptable maximum number in normal mucosa.29, 31, 32, 39, 42
Less Reliable Features Distinguishing Inflammatory Bowel Disease from Other Causes
Lamina Propria Chronic Inflammation/Lamina Propria Hypercellularity
There are various terms to describe an increase in lamina propria chronic inflammatory cells, but they are less precise or less well defined than the term basal plasmacytosis. In addition, plasma cells are easier than lymphocytes to identify, meaning that loss of the usual plasma cell gradient is a more objective sign of chronic inflammation than an increase in lymphocytes or other inflammatory cells. ‘Transmucosal chronic inflammation’ suggests chronic inflammation occupying the full thickness of the mucosa and correlates broadly with basal plasmacytosis. An increase in lamina propria lymphocytes is difficult to confirm objectively, but in many cases there is an obvious increase in the number and density of lymphocytes. Regardless of definitions, such an increase is less specific than basal plasmacytosis for IBD. In some conditions, unequivocal lamina propria chronic inflammation occurs with no basal plasmacytosis and no loss of the plasma cell gradient. Interobserver reproducibility for diagnosing this feature is low. Bearing these many caveats in mind, an increase in lamina propria chronic inflammatory cells, with or without identifiable basal plasmacytosis, is more likely in IBD than in other colitides.32, 39, 42
Paneth Cell Metaplasia
Paneth cells are common in the caecum and proximal colon and diminish in frequency with more distal anatomical location. They are more significant if present distal to the splenic flexure.31, 33 Several studies have challenged the assertion that they are largely or wholly absent in the left colon and rectum. In fact, their presence at distal locations, albeit in low numbers, has been recognised for some time.44 Not surprisingly, the likelihood of detection increases if sampling is thorough. For example, there were Paneth cells in the rectal mucosa in 17% of young patients in one report.37 Despite these observations, Paneth cells distal to the splenic flexure are worth noting and are more likely in this location to be significant (Figure 21.8). They probably favour a diagnosis of IBD over other causes, although the association is weak in some studies.20, 28, 42, 44, 45 Paneth cells also occur in GvHD, radiation colitis, and other chronic conditions, and were present in 44% of collagenous colitis biopsies in one report.46 They may be more common in chronic IBD than new IBD but the evidence is again inconsistent.45
Surprisingly, there is moderate interobserver variability for the recognition of Paneth cells. Distinction from endocrine cells is important. Paneth cells are pyramidal crypt epithelial cells with supranuclear brightly eosinophilic granular cytoplasm. Compared to Paneth cell granules, endocrine cell granules are smaller and less brightly eosinophilic and are subnuclear or perinuclear rather than supranuclear (Figure 9.23B). In addition, endocrine cells are usually ovoid or rectangular rather than pyramidal.37
Features Distinguishing Inflammatory Bowel Disease from Other Causes: Limited Data
The term ‘activity’ includes neutrophil cryptitis, neutrophil crypt abscesses, neutrophil infiltration of the surface epithelium, erosion, and ulceration.7 It does not usually include lamina propria acute inflammation. Activity is certainly not specific for IBD. There are a few studies of the relationship between severity or pattern of activity and the likelihood of IBD. For example, deeply located (rather than superficially located) crypt abscesses may favour IBD over other aetiologies31 while high counts of crypt intraepithelial neutrophils/crypt luminal neutrophils (defined as>10 per crypt or per crypt lumen when assessed in 10 crypts) favoured IBD in one study.39 Other studies also identify widespread activity as a feature favouring UC over Crohn’s disease (see later). A misconception that persists among some clinicians is that crypt abscesses are specific for UC or that their presence favours UC over Crohn’s disease.
Variation in crypt diameter favoured IBD over other causes in one study.28
|1. Highly reliable||Basal plasmacytosis||21.5|
|Crypt distortion / crypt branching / abnormal crypt architecture||21.4|
|Irregular/villous mucosal surface||21.4||Some interobserver variability|
|2. Fairly reliable||Granulomas|
|Basal giant cells||21.9||Sparse data|
|Basal lymphoid aggregates||21.6||Can be difficult to distinguish from normal lymphoid aggregates|
|3. Less reliable||Lamina propria chronic inflammation / hypercellularity|
|Paneth cell metaplasia (NOS / distal to the splenic flexure)||21.8|
|4. Limited data||Deep (rather than superficial) crypt abscesses|
|Variation in crypt diameter|
|High counts of crypt intraepithelial neutrophils / crypt luminal neutrophils||>10 per crypt or per crypt lumen, assessed in 10 crypts.|
The features that point most reliably to a diagnosis of IBD rather than another colitis in colonic mucosal biopsies:
Irregular mucosal surface
Positive histological features of acute infective colitis are difficult to identify. When distinguishing infective colitis from IBD, absence of features of IBD is the most useful pointer. In particular, basal plasmacytosis is unusual in infective colitis. Crypt changes are mild or absent, crypt atrophy is rare, and mucin depletion is not severe. Granulomas are unusual in acute infective colitis, although they are typical of Mycobacterium tuberculosis infection. Lamina propria neutrophils are more likely and chronic inflammatory cells less likely in infective colitis than in IBD. Changes may be patchy. Some reports describe a characteristic pattern of upper lamina propria hypercellularity with intraepithelial neutrophils, particularly in the resolving phase (Figure 21.14)2, 39 When this pattern occurs in colorectal biopsies, the pathologist should mention the possibility of infection. The picture may resemble or overlap with lymphocytic colitis.
Figure 21.14 Infective colitis showing upper lamina propria hypercellularity, neutrophil cryptitis affecting the upper half of the crypts more than the lower half, and neutrophils in the surface epithelium. There is mild distortion of architecture, but there is no crypt atrophy or basal plasmacytosis.
IBD versus Infective Colitis: Exceptions
Histological features are useful for distinguishing IBD from infective colitis / non-IBD, but no feature is 100% specific. Crypt distortion and other architectural changes are partly a reflection of chronicity. Crypt architectural distortion, in particular, can sometimes occur in chronic infective disorders or other chronic colitides (see also Chapters 18 and 23 and Figure 18.8B).29, 32, 42, 47 In infective colitis, architectural changes are more likely to occur if there is severe or prolonged disease, and they may develop in the healing phase.28, 34, 47 However, crypt changes that are extensive and severe are unlikely to reflect infection. Basal plasmacytosis is less likely than architectural changes to occur in infective colitis. However, basal plasmacytosis is not 100% specific for IBD, whether the comparison is with infection or with other causes of inflammation.28, 29, 42 In practice, assessment of multiple features in conjunction with the clinical history helps make the distinction.
Very early IBD may lack crypt distortion, crypt atrophy, and villiform surface change. Indeed, these features are largely absent if symptom duration is less than a few weeks (see later). Similarly, basal plasmacytosis is not universal in the early days of IBD, although it is usually the earliest change in biopsies and is present in the great majority of biopsies once several weeks have elapsed. 9, 28, 29, 42 Longstanding treated disease often lacks characteristic histological features (see later).