Inflammatory Bowel Diseases
Diarrhea is common, and it has been estimated that even in Western countries most experience at least two episodes of infectious diarrhea annually (see chapter on infections). It is generally considered chronic when the symptoms last more than 4 weeks. However, to put this in perspective, in Western societies irritable bowel syndrome (IBS) is by far the most common cause of chronic diarrhea and the most common exit diagnosis from most gastroenterology clinics. Consequently, most biopsies from the large bowel are from patients with IBS, in whom colonoscopy has been carried out to ensure there is not a more treatable cause of diarrhea such as microscopic colitis or unsuspected chronic idiopathic inflammatory bowel disease (IBD).
Diarrhea is defined in adults by abnormal stool weight (>200 g per day), consistency (loose or liquid), and/or frequency (>3 times per day). It is a common symptom in adults and the prevalence of diarrhea is approximately 1% to 5%, making it a major cause of disability.1 The list of etiology for diarrhea is fairly long; however, the major causes include IBS, infections, systemic disorders, drug toxicity, immunemediated disease, malabsorption, microscopic colitis, and IBD. While a vast majority of diarrhea is acute and self-limited necessitating no medical attention or investigations, patients with chronic diarrhea with or without the passage of blood are likely to be further investigated, often with some form of endoscopy and biopsy. Several studies show that colonoscopy with biopsy is useful in the investigation of patients with chronic diarrhea yielding a histological diagnosis in 22% to 31% of patients despite macroscopically normal colon at colonoscopy.2, 3, 4 A study evaluating more than 800 patients with chronic diarrhea who underwent colonoscopy with biopsy found that 122 (15%) of them had abnormal histopathology,3 which still means that 85% of patients had normal biopsies. Of those with abnormal biopsies, 2% would have been missed if only a flexible sigmoidoscopy had been performed. In patients older than age 50 years, colonoscopy is the method of choice.2, 3, 5, 6, 7, 8 However in a further study, 15% to 20% of patients with diarrhea-predominant IBS had features of microscopic colitis on biopsy.9 In most patients, colonoscopy shows mucosal abnormalities and the histological diagnoses include IBD, a variety of infections, systemic disorders, and drug-induced mucosal injury, among many others. However, various forms of colitis can be present in the absence of radiological and endoscopic evidence of colitis as shown in Table 18-1. Because the patterns of tissue response are limited despite a varied range of insults, the precise histological diagnosis of colitis requires a good knowledge of the different etiologies and of the microscopic features of different types of colitis.10, 11, 12, 13 This chapter discusses the major important causes of chronic diarrhea which include microscopic colitis and IBD. Readers are referred to other sections for the discussion of other major causes like infections (Chapter 19), immunodeficiency disorder (Chapter 3), malabsorption syndromes (Chapter 17), and drug toxicity and systemic disorders (Chapter 8).
Table 18-1 Normal Colonoscopy and Abnormal Histology | ||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||
MICROSCOPIC COLITIS
Introduction
The concept of “microscopic colitis” evolved from recognition that in some patients with chronic diarrhea, despite normal colonoscopy, inflammatory changes could be seen on microscopy. The term, microscopic colitis itself was used initially in 1980.14 It was subsequently shown that the most distinctive feature of this condition was a marked increase of the number of intraepithelial lymphocytes (IEL).15 The condition was therefore renamed “lymphocytic colitis.” It appeared to be different from “collagenous colitis,” a condition first reported in 1976. In this report, rectal biopsies from a woman with chronic watery diarrhea and an endoscopically normal rectal mucosa revealed a markedly thickened subepithelial collagen layer along with an increase in the lamina propria inflammatory cells.16 In 1993, a French and an American research group suggested the use of “microscopic colitis” as an umbrella term to cover any form of colitis in which there was histological, but no endoscopic or radiological abnormality.17, 18 Collagenous colitis and lymphocytic colitis were the two major types of microscopic colitis to be recognized. Since then, various
other forms of colitis have been described,19, 20, 21, 22, 23, 24 some of which are variants of these two disorders, while others appear to be distinct nosological entities.25, 26, 27, 28, 29, 30 However, in current practice the use of the term “microscopic colitis” largely implies either collagenous or lymphocytic colitis and their variants, and the other conditions like eosinophilic colitis or possibly mastocytic enteropathy are designated by their specific names.
other forms of colitis have been described,19, 20, 21, 22, 23, 24 some of which are variants of these two disorders, while others appear to be distinct nosological entities.25, 26, 27, 28, 29, 30 However, in current practice the use of the term “microscopic colitis” largely implies either collagenous or lymphocytic colitis and their variants, and the other conditions like eosinophilic colitis or possibly mastocytic enteropathy are designated by their specific names.
As is clear, the diagnosis of microscopic colitis requires clinicopathologic correlation (normal endoscopy, abnormal histology). A typical patient has a chronic intermittent disease course with watery diarrhea and characteristic microscopic features on biopsy. The clinicopathologic feature of collagenous colitis and lymphocytic colitis are very similar and hence discussed together, but they are not identical and some patients may even show overlapping features (discussed later). Whether these two are distinct entities or represent ends of a spectrum remains unresolved at present.
Epidemiology
The vast majority of reported patients with microscopic colitis have been from countries in Europe and North America and only a few series have been reported from Africa, Asia, Australia, and South America. It is estimated that about 10% of patients evaluated for chronic nonbloody watery diarrhea with colonoscopy appear to have microscopic colitis.31, 32 Incidence studies are rare. In a Swedish study, the mean annual incidence of collagenous colitis was 1.8 per 105 inhabitants. The median age at diagnosis was 64 years and the female to male ratio was 9:1.33 In a population-based study from Calgary, Canada from 2002 to 2004, the annual incidence was 10.0 per 105 person (lymphocytic colitis, 5.4; collagenous colitis, 4.6 per 100,000). Patients older than 65 years were five times more likely to develop microscopic colitis (RR, 5.6; 95% confidence interval [CI], 4.0-7.7). Women were at higher risk of acquiring microscopic colitis for both collagenous colitis (RR, 3.44; 95% CI, 2.07-5.97) and lymphocytic colitis (RR 6.29; 95% CI, 3.21-13.74). Elderly women with a history of malignancy were also associated with a higher risk of microscopic colitis (RR, 3.59; 95% CI, 1.68-7.01), as were patients with celiac disease (RR, 7.9; 95% CI, 4.0-14.2) and hypothyroidism (RR, 6.1; 95% CI, 3.5-10.0).34 In a population-based study from Olmsted County, Minnesota from 1985 to 2001, the annual incidence was 8.6 cases per 105 persons. There was a significant secular trend, with annual incidence increasing from 1.1 per 105 early in the study to 19.6 per 105 by the end (p < 0.001). Rates increased with age (p < 0.001). By subtype, the incidence was 3.1 per 105 for collagenous colitis and 5.5 per 105 for lymphocytic colitis. In this study, collagenous colitis was associated with female gender (p < 0.001), but lymphocytic colitis was not. Prevalence (per 105 persons) calculated on the last day of the study (December 31, 2001) was 103.0 (39.3 for collagenous colitis and 63.7 for lymphocytic colitis).35 Overall, it appears that the incidence of microscopic colitis is rising; among microscopic colitis, lymphocytic colitis appears to be more common than collagenous colitis and the predilection for female gender is stronger in collagenous colitis compared to lymphocytic colitis.
Etiology and Pathogenesis
The etiology of lymphocytic colitis and collagenous colitis is largely unknown and is probably multifactorial. It is suggested that they represent disorders with an abnormal mucosal immune response in predisposed individuals to various noxious luminal agents that include infections and toxins/medications. Many medications have been associated with lymphocytic (especially) and collagenous colitis, of which common ones include nonsteroidal anti-inflammatory drugs (NSAIDs), lansoprazole and ticlodipine (Table 18-2).36 In one study, 37.5% of patients taking long-term steroids and having diarrhea had lymphocytic colitis.37 While many of the drugs listed are known to cause or exacerbate diarrhea, these are also very commonly used drugs, and case-control studies clearly establishing an etiologic relationship are still lacking. In one review of 104 patients with microscopic colitis, about 35% of patients reported NSAID use.38 There are considerably fewer cases of drug-associated collagenous colitis compared to lymphocytic colitis.36
Table 18-2 Medications Potentially Related to the Development of Lymphocytic Colitis | |
|---|---|
|
Table 18-3 Autoimmune Disorders Associated with Microscopic Colitis | ||
|---|---|---|
|
Many autoimmune or systemic inflammatory diseases are associated with microscopic colitis suggestive of a role for autoimmune or immune deregulation in the etiopathogenesis (Table 18-3).39, 40, 41, 42, 43 Collagenous colitis has also been reported to occur rarely as a paraneoplastic phenomenon in a patient with colon cancer, which resolved following tumor resection.44
Enteric infections such as Clostridium difficile, Campylobacter jejuni, and Yersinia enterocolitica have also been associated with the onset of microscopic colitis,45, 46, 47 and anecdotally, we have seen the same with collagenous colitis. A condition with histopathologic resemblance to lymphocytic colitis is “Brainerd diarrhea,” an outbreak of chronic watery diarrhea characterized by acute onset and prolonged duration, for which an infectious cause has been suggested (see later).13 Symptoms in lymphocytic colitis seems to occur more frequently in summer and fall, and this seasonal variation suggests a possible etiologic role of infections.48 Interestingly, a few cases of collagenous colitis with concomitant Helicobacter pylori gastritis have been reported that had resolution of diarrhea following treatment of H. pylori.49, 50 Immunologic response to H. pylori antigen has been postulated as a possible mechanism. Benefit from probiotic therapy in patients with collagenous colitis suggests that altered gut flora may also play a role in its pathogenesis in some patients.51 Genetically engineered mice that express HLA-B27 routinely develop lymphocytic colitis-like picture, but not in the absence of luminal bacteria.52 The postdysenteric IBS also mimics lymphocytic colitis. In this condition, the number of IEL can increase and decrease over time, in common with CD3+ lymphocytes in the lamina propria.10
Rare examples of familial cases have been described with both collagenous and lymphocytic colitis suggesting a role for genetic factors.53, 54 Increased frequency of HLA-A1 and a decreased frequency of HLA-A3 have been reported in lymphocytic colitis compared to controls and patients with collagenous colitis.55 Interestingly HLA-B27, which is commonly associated with other autoimmune disorders is not linked with microscopic colitis. NOD2/CARD15 gene polymorphisms that are associated with IBD are also not involved in the susceptibility to collagenous colitis.56
Celiac disease has also been associated with microscopic colitis, especially lymphocytic colitis. In a study by Wolber et al.,57 12 out of 39 (31%) patients with newly diagnosed celiac disease had a simultaneous diagnosis of lymphocytic colitis. Gluten-free diet in ten of these patients resulted in a clinical improvement in nine. Rectal gluten challenge in patients with celiac disease has shown a rapid increase of the number of IEL in the rectum.58 It is thus likely that gluten is somehow responsible for the histopathologic changes of lymphocytic colitis in some patients with celiac disease.
Other noxious luminal factors (chemical toxins or bacteria) may also be responsible for microscopic colitis, as in one study fecal diversion with ileostomy or sigmoidostomy led to a clinical and histologic remission in all patients (n = 9) of collagenous colitis, and closure of the ostomy resulted in disease recurrence in all.59 Bile salts could be one of the luminal factors, as bile-binding resins like cholestyramine have beneficial effects in a subset of patients.60, 61, 62 In one remarkable case of gall bladder to transverse colon fistula seen by us, collagenous colitis was seen in the colon distal to the fistula only. Resolution of diarrhea was seen immediately following closure of the fistula, suggesting a role for bile in the pathogenesis of collagenous colitis.
The mechanism of diarrhea appears to be multifactorial and could be osmotic, secretory, or both.63 Diarrhea may continue with fasting or elemental diet. Some patients have steatorrhea, likely related to smallbowel involvement. The pathophysiology has been studied mainly in collagenous colitis.64 Most attention has been focused on the subepithelial collagen layer, epithelial barrier function, and on the presence of lamina propria, eosinophils, and mast cells. Thickened subepithelial collagen layer has been suggested to result in a diffusion barrier. Thickening of the collagen layer can be caused by overproduction, reduced degradation, or a combination of both. Overproduction is supported by the immunohistochemical finding of an increased expression of the myofibroblast marker a-smooth muscle actin and the increased expression of procollagens shown by in situ hybridization.
Reduced matrix degradation is supported by the reduced expression of the collagenase matrix-metalloproteinase (MMP)-1 and increased expression of tissue inhibitor of metalloproteinase-1 (TIMP-1).65 A rare case of collagenous colitis associated with osteogenesis imperfecta, which is an inherited disorder of collagen synthesis has also been reported.66 Decreased expression of occludins and claudin-4 has been shown in collagenous colitis suggesting defects of epithelial barrier function in collagenous colitis.64 Increased expression of TGF-β1 in eosinophils, cyclooxygenase-2 (COX-2) in macrophages, connective tissue growth factor (CTGF) in myofibroblasts, mucosal basic fibroblast growth factor (bFGF), increased mucosal expression of vascular endothelial growth factor (VEGF), and increased endoluminal nitric oxide have all been reported in collagenous colitis.67, 68, 69, 70, 71 A reduced expression of CD1d has been shown in colonic epithelium of patients with microscopic colitis.72 CD1d is a major histocompatibility complex class I—like molecule, which is involved in regulating normal immunity of the gastrointestinal tract (GI) and thus disordered mucosal immunity may play a role in the pathogenesis. Active collagenous colitis is also associated with a TH1 and interferonγ mucosal cytokine profile, similar to patterns found in celiac disease.73 Number of eosinophils and mast cells are increased in microscopic colitis. Eosinophil activation may affect mucosal permeability.74, 75 Also, increased production of histamine by mast cells may promote diarrhea.76
Reduced matrix degradation is supported by the reduced expression of the collagenase matrix-metalloproteinase (MMP)-1 and increased expression of tissue inhibitor of metalloproteinase-1 (TIMP-1).65 A rare case of collagenous colitis associated with osteogenesis imperfecta, which is an inherited disorder of collagen synthesis has also been reported.66 Decreased expression of occludins and claudin-4 has been shown in collagenous colitis suggesting defects of epithelial barrier function in collagenous colitis.64 Increased expression of TGF-β1 in eosinophils, cyclooxygenase-2 (COX-2) in macrophages, connective tissue growth factor (CTGF) in myofibroblasts, mucosal basic fibroblast growth factor (bFGF), increased mucosal expression of vascular endothelial growth factor (VEGF), and increased endoluminal nitric oxide have all been reported in collagenous colitis.67, 68, 69, 70, 71 A reduced expression of CD1d has been shown in colonic epithelium of patients with microscopic colitis.72 CD1d is a major histocompatibility complex class I—like molecule, which is involved in regulating normal immunity of the gastrointestinal tract (GI) and thus disordered mucosal immunity may play a role in the pathogenesis. Active collagenous colitis is also associated with a TH1 and interferonγ mucosal cytokine profile, similar to patterns found in celiac disease.73 Number of eosinophils and mast cells are increased in microscopic colitis. Eosinophil activation may affect mucosal permeability.74, 75 Also, increased production of histamine by mast cells may promote diarrhea.76
The thickened collagen layer may contribute to disturbed water and electrolyte absorption; however, there is no correlation of the thickness of the collagen layer with the severity of diarrhea, and diarrhea may persist despite disappearance of the subepithelial collagen layer.60, 77 On the other hand, thickened subepithelial collagen layer may be still be present despite disappearance of diarrhea or never having one.60 It is also possible that increase in subepithelial collagen layer is a mere morphological marker of the disease with very little role in the pathophysiology of the diarrhea. The situation is similar with lymphocytic colitis. In one study, follow-up biopsies in 13 patients with lymphocytic colitis, who had achieved remission and were symptom free, persistent lymphocytic colitis and mild nonspecific inflammation were present in 3 patients each, while 5 showed completely normal histology and 2 had interestingly converted to collagenous colitis.60 In the same study, follow-up biopsies of 10 patients with collagenous colitis, who were in remission and symptom free, showed persistent changes in 8, normal histology in 1, and change to lymphocytic colitis in 1 patient. This either suggests lag between clinical and histologic remission or a lack of correlation between histologic changes with symptoms.
Clinical Features and Endoscopic Findings
The typical presentation of microscopic colitis is chronic intermittent watery diarrhea. Weight loss and abdominal pain are less prominent. About one fourth of the patients may have nocturnal diarrhea. Rare cases may develop protein-losing enteropathy.78 Some patients do not have diarrhea.43 The clinical features of patients with lymphocytic colitis are less homogeneous. Overall, the age at onset, clinical symptoms, laboratory data, and occurrence of associated diseases are similar to patients with collagenous colitis, but with slightly older age of onset, less stronger females predilection, and shorter disease history.60 Blood in the stool is not a typical feature, so its presence should lead the clinician into considering other differentials. Routine blood tests are not diagnostic. Collagenous colitis patients have a higher prevalence of antinuclear, antigliadin, and anti-Saccharomyces cerevisiae antibodies than patients with lymphocytic colitis and controls (patients without GI disorders); however, these are diagnostically not useful.43, 79, 80
Radiographic and endoscopic examinations are normal or show nonspecific abnormalities. Mucosal granularity and irregularity of the rectosigmoid on double-contrast barium enema have occasionally been reported in collagenous colitis.81 Although classically there is no endoscopic abnormality, in practice, erythema or edema have been reported in up to one third of cases.
Microscopic Pathology
The major features and differences between collagenous colitis and lymphocytic colitis are summarized in Table 18-4.
Collagenous colitis. Collagenous colitis is characterized by the combination of a chronic colitis (plasma cells) and thickening of the subepithelial collagen layer (Figs. 18-1 and 18-2). Architecture is normal, although a minor degree of architectural distortion, as well as Paneth cell metaplasia, is common. However, these are rarely severe enough to confuse with IBD. Variable degree of mucosal inflammation is often present, being minimal in some cases, usually in the resolving phase. The thickness of the subepithelial collagen layer consisting of basement membrane in the normal colorectal mucosa is <4 µm with minor regional variations. It normally appears to be slightly thicker in the right side of the colon compared to the left and the normal values reported in the literature vary between 2-3 and 3-7 µm.82 Most studies suggest a minimum thickness of 10 µm of the subepithelial collagen layer for a diagnosis of collagenous colitis, although it usually measures 15 to 30 µm, or more
(up to 70). One needs to be careful when interpreting a tangential section, where the basement membrane may artifactually appear thicker (Fig. 18-3A).
(up to 70). One needs to be careful when interpreting a tangential section, where the basement membrane may artifactually appear thicker (Fig. 18-3A).
Table 18-4 Comparison of Collagenous and Lymphocytic Colitis | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Another common artifact is lighter staining of the subnuclear part of the cytoplasm of the surface epithelium that mimics a thickened collagen layer; however, closer look at higher magnification makes the differentiation fairly easy (Fig. 18-3B). Comparison with the size of nuclei of small lymphocytes or plasma cells (5-7 µm) is a rough guide for estimating thickness of subepithelial collagen layer (if thicker than two of these nuclei in a well-oriented part of the biopsy, one is usually in the pathological range). Actual measurement of the subepithelial collagen layer is seldom required to make a diagnosis, and there is no evidence that quantitative analysis improves the diagnostic accuracy.83 A cut-off of 10 µm as a diagnostic criteria has been chosen arbitrarily and cases with lessthickened subepithelial collagen layer (4-9 µm) in the presence of other supporting histologic findings likely represent cases of microscopic colitis, although as the thickening becomes less conspicuous, the distinction from lymphocytic colitis becomes almost arbitrary (Fig. 18-3C).84 The thickening of the subepithelial
collagen layer may be patchy, both within a biopsy specimen as well as between specimens from different regions of the colon, usually being thickest in the transverse colon. Thickening increases toward the right side of the colon and can be absent in the sigmoid colon and rectum. The rectum is spared in approximately 30% to 40% of cases.85, 86 Thus, rectal biopsies alone may miss this disease. While initial workup of patients (with flexible sigmoidoscopy) to rule out microscopic colitis with four biopsies every 10 cm from the left colon may be appropriate, negative results should be followed up with a full colonoscopy.87 Also, initial evaluation may miss the diagnosis in a good number of cases (one fourth to one third), which may either be related to sampling or lack of fully developed histology at the time; hence, there is a role for repeat colonoscopy.31, 88
collagen layer may be patchy, both within a biopsy specimen as well as between specimens from different regions of the colon, usually being thickest in the transverse colon. Thickening increases toward the right side of the colon and can be absent in the sigmoid colon and rectum. The rectum is spared in approximately 30% to 40% of cases.85, 86 Thus, rectal biopsies alone may miss this disease. While initial workup of patients (with flexible sigmoidoscopy) to rule out microscopic colitis with four biopsies every 10 cm from the left colon may be appropriate, negative results should be followed up with a full colonoscopy.87 Also, initial evaluation may miss the diagnosis in a good number of cases (one fourth to one third), which may either be related to sampling or lack of fully developed histology at the time; hence, there is a role for repeat colonoscopy.31, 88
 Figure 18-1. Collagenous colitis. A: Low-power view of an obvious example of collagenous colitis. Overview shows crypts with a normal architecture which may not extend as close to the muscularis mucosae as usual. The subepithelial collagen band, which occupies the upper quarter of the mucosa and measures about 60 µm in thickness, is apparent. B: High power: The luminal epithelium may be normal, but is often injured and cuboidal or low columnar with intraepithelial inflammatory cells. The inflammatory cells are mainly lymphocytes and are frequently very prominent. However, eosinophils and less frequently, neutrophils can be present. Also note the capillaries entrapped in the thick collagen layer; sometimes, when the thickness of the band is equivocal this feature may support a diagnosis of collagenous colitis. |
 Figure 18-1. (Continued) C: Trichrome stain showing normal subepithelial and pericryptal collagen, as compared with markedly thickened subepithelial collagen in an example of collagenous colitis in (D). E: Sometimes, the eosinophils can be very prominent in the inflammatory infiltrate. F: Collagenous colitis with stripping of the surface epithelium (arrow). This is sufficiently characteristic of the disease that its presence should prompt immediate examination of the underlying collagen band, which here is greatly thickened. G: Immunostain for laminin shows strong positivity in the thickened subepithelial collagen, while staining for collagen type IV (basement membrane collagen) is negative (H). |
 Figure 18-2. A: Electron micrograph of the subepithelial collagen layer. One mircometer epon section showing the thick subepithelial band and characteristic trapped capillaries underneath the epithelium. B: The basement membrane region (arrow) in collagenous colitis is commonly deficient focally (asterisk). The collagen layer itself consists of a mixture of amorphous ground substance and mature collagen. C: Detail to show the interface between the collagen fibers above and the ground substance below. (B and C, Courtesy of Dr. T.K. Shnitka.) |
 Figure 18-3. Possible errors in diagnosis. A: Tangential sectioning; the underlying crypts are visible in cross section, causing an artifactual impression of thickening of the collagen band. B: Nuclei of luminal epithelium are in the center of the cell, so that an eosinophilic zone is present beneath them, which can be mistaken for a thickened collagen band. |
 Figure 18-3. (Continued) C: Thickening of the subepithelial collagen that is <10 µm. Compare with the adjacent nuclei of lymphocytes. In the presence of increased IEL, lamina propria infiltrate, and epithelial injury, it certainly represents microscopic colitis, and a diagnosis of lymphocytic colitis would be favored. D: Marked thickening of subepithelial collagen in a hyperplastic polyp. Remaining colonic mucosa in this fragment and other colonic biopsies from this patient do not show any evidence of collagenous colitis. |
Identification of the thickened subepithelial collagen layer is easy on H&E sections in most cases and can be highlighted with histochemical stains for collagen like the trichrome stain. Another simple method is to examine the H&E sections under a fluorescent microscope as the collagen is autofluorescent and the thickening of subepithelial collagen layer is immediately obvious.89 The SCL is composed of collagen type I, III, and VI in addition to the glycoprotein tenascin and is separated from the epithelial cells by the apparently normal basement membrane composed of collagen type IV (Fig. 18-1H).90 Additional evaluation using immunohistochemistry with antibodies directed against tenascin, an extracellular matrix component involved in cell differentiation, cell adhesion, and cell migration or various collagen types is not necessary for routine clinical practice, although one study suggested increased diagnostic sensitivity and specificity when using antibodies to tenascin.91
Special stains are most useful when the thickening is borderline, patchy, or need to be differentiated from edema or amyloid deposition. The lower border of the thickened subepithelial collagen layer is ragged and the collagen fibers seem to extend and merge with the collagen in the lamina propria. Also, the subepithelial collagen layer seems to entrap small, capillaries and inflammatory cells in the superficial lamina propria (Fig. 18-1B). These two features are helpful in doubtful cases, where the subepithelial collagen layer needs to be differentiated from apparent thickening of the basement membrane due to tangential sectioning. The subepithelial collagen layer thickening may persist or disappear after treatment despite improvement in diarrhea.92 In some patients, the changes may become patchy or start resembling lymphocytic colitis over time due to the disappearance of subepithelial collagen layer.
While thickening of subepithelial collagen layer is the defining feature of collagenous colitis, a wide variety of other histologic changes are often present that overlap with lymphocytic colitis. The surface epithelium is usually focally or sometimes diffusely abnormal (Fig. 18-1B). Flattening of the epithelial cells, mucin depletion, and vacuolization of the cytoplasm are common features. The surface epithelium is often artifactually detached from the underlying subepithelial collagen layer, often referred to as “stripping of the epithelium”—a useful diagnostic feature on low magnification (Fig. 18-1F). Crypt distortion is uncommon and when present is focal and mild, being restricted to a few branching crypts. More widespread crypt
distortion should lead to a suspicion of IBD, and in the presence of normal endoscopy, this has been referred to as “minimal change colitis” (see later), which usually behaves as ulcerative colitis (UC).12 Paneth cell metaplasia is not uncommon and is seen more frequently with collagenous colitis than lymphocytic colitis.93 Paneth cell metaplasia has been noted more frequently in patients whose symptoms like abdominal pain are ongoing or recurrent, and it was suggested that its presence may be used as a marker to predict more persistent or severe disease.93, 94 Some patients who have patchy thickening of subepithelial collagen layer may show some biopsies with typical features of collagenous colitis, while others resemble lymphocytic colitis, possibly accounting for cases reported in the literature with combined lymphocytic colitis and collagenous colitis.43 When there is evidence of both conditions (i.e., lymphocytic colitis with a thickened subepithelial collagen band), collagenous colitis “trumps” lymphocytic colitis by convention. Some patients with lymphocytic or collagenous colitis have extensive involvement of the GI tract including stomach and small bowel (pan gastroenterocolitis) (see also collagenous gastroenterocolitis).95, 96
distortion should lead to a suspicion of IBD, and in the presence of normal endoscopy, this has been referred to as “minimal change colitis” (see later), which usually behaves as ulcerative colitis (UC).12 Paneth cell metaplasia is not uncommon and is seen more frequently with collagenous colitis than lymphocytic colitis.93 Paneth cell metaplasia has been noted more frequently in patients whose symptoms like abdominal pain are ongoing or recurrent, and it was suggested that its presence may be used as a marker to predict more persistent or severe disease.93, 94 Some patients who have patchy thickening of subepithelial collagen layer may show some biopsies with typical features of collagenous colitis, while others resemble lymphocytic colitis, possibly accounting for cases reported in the literature with combined lymphocytic colitis and collagenous colitis.43 When there is evidence of both conditions (i.e., lymphocytic colitis with a thickened subepithelial collagen band), collagenous colitis “trumps” lymphocytic colitis by convention. Some patients with lymphocytic or collagenous colitis have extensive involvement of the GI tract including stomach and small bowel (pan gastroenterocolitis) (see also collagenous gastroenterocolitis).95, 96
In addition, there is an increase in mucosal inflammatory infiltrate similar to lymphocytic colitis. There is increase in the IELs, mostly CD8+ T-cells, sometimes with other intraepithelial inflammatory cells such as neutrophils and eosinophils. The infiltrate is predominantly mononuclear, mainly consisting of plasma cells and lymphocytes. Most of the lymphocytes are CD4+ T-cells. A significant decrease of inflammatory cells can be noted after treatment.92 In collagenous colitis, there is also an increase in the number of eosinophils, often in association with the subepithelial collagen layer.97 Eosinophils can be particularly prominent in some patients. The absolute number of mast cells may also be increased, particularly in the upper part of the lamina propria.98 Neutrophils and cryptitis are occasionally present. Colitis characterized by a marked eosinophilic infiltrate (≥20 per high power field) is also seen with drug reactions and food intolerance in children (Fig. 18-1E). It can be responsible for not only rectal bleeding but also for diarrhea with normal or abnormal colonoscopy. It can be distinguished from microscopic (collagenous or lymphocytic) colitis by the lack of increase of mononuclear cells.99
Complications
i. Tears, lacerations, ulceration, perforation: Longitudinal mucosal tears (fractured colon), hemorrhagic lacerations, and submucosal dissection are seen in a minority of the patients with collagenous colitis.100, 101, 102 While in most the tears are thought to be secondary to barotrauma due to insufflation of bowel during colonoscopy, rarely these may occur spontaneously. The tears likely reflect reduced distensibility or elasticity due to increased collagen deposition in the wall. These can heal to form endoscopic scars, but occasionally, these may go on to perforation.103 Submucosal dissection of air with pneumoperitoneum and rarely perforation may occur.103, 104, 105 The pattern of mucosal tears that occurs mostly in the right side of the colon has been occasionally described as “cat scratch colon.”106 Interestingly, some patients with “cat scratch colon” don’t have collagenous colitis. While the experience with confocal endomicros-copy is still limited, it may help in establishing the diagnosis in patients with collagenous colitis on colonoscopy.107, 108 Many such cases have been associated with NSAID use and the discontinuation seems to resolve the diarrhea in many patients.109 The ulcers can be seen throughout the colon, but are more common in the right side and vary from longitudinal discrete ulcers to small punched out or aphthous ulcers. The resolution of the diarrhea may occur soon after discontinuation of NSAIDs or may take months requiring additional therapy.109 It is unclear if the resolution of the diarrhea is also accompanied by disappearance of the thickened subepithelial collagen layer. In some patients the ulcers may be ischemic in nature, and in some the etiology remains unknown.
ii. Pseudomembranes: Several studies have reported pseudomembrane formation in patients with collagenous colitis, many of whom have no evidence of concomitant C. difficile infection.110, 111 Other causes of pseudomembrane formation in this setting may include ischemia or medications, especially NSAIDs.
iii. Complications from associated disease: Some patients also have concomitant IBD, so inflammatory polyps, polyposis, toxic dilatation may inevitably be seen in patients with collagenous colitis. A fatality due to collagenous colitis has been reported.112 Rarely, toxic megacolon with collagenous colitis necessitating colonic resection has been reported, and it may also be fatal.113 Whether this is a complication of evolving IBD superimposed on collagenous colitis remains unclear.114 Because collagenous colitis is rarely associated with lymphocytic and IBD, complications associated with those disorders should not be interpreted as being secondary to collagenous colitis. Nevertheless, they may be at least additive and can pose difficult management problems.
iv. Colorectal adenomas and carcinoma. Although rare cases of colon cancer associated with microscopic colitis have been reported, there are no data to suggest an increased risk, and one study suggests that the risk may be reduced.115 It may be that the chronic voluminous diarrhea dilutes luminal carcinogens.
Differential Diagnosis of Collagenous Colitis A thickened subepithelial collagen layer in the absence of a colitis has been noted in patients with diverticular disease, megacolon, colonic carcinoma, and rarely pseudomembranous colitis. It can be sometimes very prominent in serrated/hyperplastic polyps (Fig. 18-3D). Usually, the underlying lesion, limited area of subepithelial collagen layer thickening, and absent mucosal inflammation allows differentiation from collagenous colitis.90, 116, 117, 118 Collagenous colitis is a colitis, and usually an inflammatory component is present; however, in some cases, the lamina propria infiltrate may be minimal or lacking, including the injury to surface epithelium and increase in IELs.
Subtle thickening of subepithelial collagen layer is also seen in diabetic patients; however in practice, this is less likely to be a diagnostic problem as not only the thickening is less than that seen in collagenous colitis but also the inflammatory changes are lacking. In one study, the median thickness of subepithelial collagen layer in the rectum, sigmoid, and descending colon biopsies from diabetic patients irrespective of diarrhea ranged between 4-5 µm compared to 2-3 µm in controls.119 This thickening of subepithelial collagen layer results from nonenzymatic glycosylation of collagen and immunohistochemically lacks positivity for collagen type-1 unlike collagenous colitis, while retaining positivity for fibronectin similar to normal.119
Because there can be architectural distortion and atrophy (gap between the base of the crypts and the muscularis mucosa), Paneth cell metaplasia, acute and chronic inflammation with occasional crypt abscesses, there may be a problem separating collagenous colitis from IBD (usually UC) with an associated subepithelial collagen band. The easy way of resolving this is to marry the findings with those seen endoscopically. If normal it is collagenous colitis, and if there are features of IBD, then it is UC with a thickened collagen band.
Amyloid deposition in lamina propria can rarely come in the differential diagnosis of collagenous colitis, but can be easily differentiated from collagen with special stains.
Scarring in the lamina propria can be histologically confused for collagenous colitis; however, lack of diarrhea, loss of crypts, and focality of the lesion should help in its differentiation from collagenous colitis.
In addition, overlapping histologic features with lymphocytic colitis may be present in some patients. For example, increase in IEL was seen in 28% of collagenous colitis patients and mild focal thickening of subepithelial collagen layer (<10 µm) was seen in 26% of lymphocytic colitis patients.92 Patients with lymphocytic colitis having a thin, but structurally abnormal layer of subepithelial collagen layer, evident by trichrome stain or autofluorescence have been referred to as having “minimal collagenous colitis,” while one may also refer to them as having microscopic colitis with lymphocytic colitis and collagenous colitis overlap or lymphocytic colitis.26 However, traditionally the presence of a unequivocal subepithelial collagen band in these patients results in a diagnosis of collagenous colitis, although prominent intraepithelial lymphocytosis may be noted. There is no evidence that one approach is better than the other, and clinically, this is of little significance as the treatment is no different.
One should be careful in interpreting the biopsies from the right side of the colon, especially ileocecal valve area as they may show mild increase in IELs as well as increased density of lamina propria inflammatory cells.
Lymphocytic colitis. Lymphocytic colitis is “the” colitis in which a drug-associated etiology needs to be considered. These are discussed above and listed in Table 18-2. NSAIDs, PPIs, and antiplatelet medications top the list. Histologically, it is characterized by an increase in intraepithelial and lamina propria mononuclear inflammatory infiltrate. The increase in the IELs is sometimes accompanied by other intraepithelial inflammatory cells such as neutrophils and eosinophils. A diffuse increase of the number of IELs is characteristic of lymphocytic colitis (Fig. 18-4A,B). The required minimum number varies between 10 and 20 per 100 surface epithelial cells (normal number = 4), and most studies use a cut of >20. In a series of 59 patients with lymphocytic colitis and a subepithelial collagen layer <10 µm, three to five biopsy samples were examined. The number of IEL varied between 10 and 65 (median 30). There was no tendency for a more prominent IEL increase in a particular segment of the colon.92
The surface epithelium is frequently abnormal, being attenuated, cuboidal, or low columnar. The crypt epithelium may show a reduction of goblet cells (mucin depletion) (Fig. 18-4). Similar to collagenous colitis, artifactual stripping of the surface epithelium is common. The lamina propria cellularity is increased with an excess of plasma cells that varies from minimal to marked (to the point that this entity could be called plasmacytic rather than lymphocytic colitis). More subtle increase in mucosal inflammatory cells can be sometimes difficult to judge. Normally, the mononuclear cells in the lamina propria tend to be more dense in the upper half to one-third creating a gradient. This gradient is maintained despite wide variation in the density of lamina propria infiltrate in normal colons. This is best appreciated at low magnification. In microscopic colitis, there is increased cellularity in the lamina propria with presence of increased plasma cells in the lower half of lamina propria representing a loss of this normal gradient. This
should be seen in just more than one well-oriented biopsy fragment and is a useful diagnostic feature appreciated even at low magnification.24 When inflammation is marked, plasma cells may reach the muscularis mucosae thereby mimicking IBD. A variety of other cells may also be increased in numbers, including neutrophils and eosinophils (Fig. 18-4A-C). Cryptitis when seen is usually focal.93 The crypt architecture is generally well preserved and minimal crypt distortion with an occasional bifid crypt can be seen (Fig. 18-4C).93 Paneth cell metaplasia may also be seen. Usually, the disease tends to be diffuse throughout the large bowel and the rectum is spared in about 8% of cases.120 Random rectal biopsies are therefore likely to provide false-negative biopsies, while biopsies from the more proximal left colon in addition (descending and sigmoid colon) markedly increase the chances of getting the correct diagnosis.
should be seen in just more than one well-oriented biopsy fragment and is a useful diagnostic feature appreciated even at low magnification.24 When inflammation is marked, plasma cells may reach the muscularis mucosae thereby mimicking IBD. A variety of other cells may also be increased in numbers, including neutrophils and eosinophils (Fig. 18-4A-C). Cryptitis when seen is usually focal.93 The crypt architecture is generally well preserved and minimal crypt distortion with an occasional bifid crypt can be seen (Fig. 18-4C).93 Paneth cell metaplasia may also be seen. Usually, the disease tends to be diffuse throughout the large bowel and the rectum is spared in about 8% of cases.120 Random rectal biopsies are therefore likely to provide false-negative biopsies, while biopsies from the more proximal left colon in addition (descending and sigmoid colon) markedly increase the chances of getting the correct diagnosis.
 Figure 18-4. Lymphocytic colitis. A,B: The architecture is intact, with evenly spaced crypts pushed apart by an inflammatory infiltrate. Surface epithelial cells are often cuboidal but can be tall and columnar. Numerous IEL are present. The subepithelial collagen band approximates the thickness of the diameter of the adjacent lymphocyte or plasma cell nuclei, and therefore is within the normal range although it can be slightly thickened. C: Occasional crypt branching is acceptable for the diagnosis of lymphocytic colitis. D: Staining with antibodies against CD3 confirms and illustrates more clearly the presence of numerous intraepithelial T cells; however, one seldom needs the stain to make the diagnosis. |
Microscopic colitis and involvement of small intestine. Study of ileal biopsies in a systematic retrospective analysis revealed increased villous IELs (>5/100 epithelial cells) in most cases of lymphocytic colitis and collagenous colitis compared to patients of IBD and controls.121, 122 Other studies have also shown similar findings although the number of IEL counts in the ileal villous and surface epithelium has varied between studies (Fig. 18-5).121, 122 This finding can be subtle. In some patients mild villous blunting can also be seen, of which at least some are likely related to concomitant celiac disease.123, 124 In one study, duodenal abnormalities were reported in up to 70% of patients with microscopic colitis of whom only 7% had positive antiendomysial antibodies and likely had associated celiac disease (Fig. 18-6). This implies that patients not responding to conventional therapy in microscopic colitis should undergo testing for celiac disease, as some will respond to a gluten-free diet.43 Conversely, patients with celiac disease who fail to respond appropriately with gluten-free diet need to be investigated for microscopic colitis.43 In addition, increased subepithelial collagen deposition has also been reported in the terminal ileum in some patients with collagenous colitis.45, 125 This needs to be differentiated from collagenous sprue, where the changes also involve the duodenum and are invariably accompanied by marked villous blunting and other features of celiac disease. Most of these reported patients had normal proximal intestinal/duodenal histology, and terminal ileal changes were unlikely to be related to celiac disease. While the histologic changes tend to suggest a more widespread involvement of intestinal tract in microscopic colitis, functional assays suggest that small-bowel permeability is at least not altered in collagenous colitis and small-bowel involvement in microscopic colitis otherwise may not have any added clinical significance.126
 Figure 18-5. Lymphocytic colitis can be associated with lymphocytic ileitis characterized by regular flattening of villi, cuboidal surface epithelial cells, and an increase of IEL. |
Differential Diagnosis of Lymphocytic Colitis
Diagnostic issues may also arise in biopsies that show increased IELs without increase in lamina propria mononuclear inflammatory infiltrate or those that do have the lamina propria inflammatory infiltrate but lack increase in IELs. Significance of these changes is often unclear and the etiology may include druginduced mucosal injury, resolving infection or evolving IBD, as Crohn’s disease (CD) in particular may be accompanied by an intraepithelial lymphocytosis in any part of the GI tract. In practice, one should refrain from making a diagnosis of lymphocytic colitis in these situations when the histologic criteria are not satisfied. Some of these issues are discussed in more detail in the section on lymphocytic colitis variants and relationship of microscopic colitis and IBD. Those with a hint of a subepithelial collagen band likely represent collagenous colitis (see above).
Treatment and Prognosis
Microscopic colitis is usually a benign chronic disease. Most patients have a relapsing and remitting course, although long-term remission is achievable in some. Response to treatment is variable and no single therapy is universally effective.43 In some patients, withdrawal of the offending medication or treatment of associated celiac disease is sufficient to control the diarrhea, while others require medications. Spontaneous resolution also occurs making evaluation of effectiveness of various therapies sometimes difficult. Some simply respond to nonspecific antidiarrheals or antibiotics.43 A variety of antiinflammatory and immunosuppressive drugs have been tried over the years. Antidiarrheal medications, cholestyramine, and probiotics,51 sulfasalazine or other 5-ASA derivatives have all been used. Currently, budesonide, a topically acting steroid, or high-dose bismuth preparations is considered the treatment of choice for both lymphocytic and collagenous colitis.
Although patients respond to medical therapy, relapses are frequent after stopping the medication. If the initial therapy fails, treatment with steroids or even more potent immunosuppressives can be used. Some patients appear refractory to all therapies and rarely colectomy may be considered. Fecal diversion is also reserved for extremely resistant cases, but is seldom required.43
Although patients respond to medical therapy, relapses are frequent after stopping the medication. If the initial therapy fails, treatment with steroids or even more potent immunosuppressives can be used. Some patients appear refractory to all therapies and rarely colectomy may be considered. Fecal diversion is also reserved for extremely resistant cases, but is seldom required.43
 Figure 18-6. A: Variants of lymphocytic colitis characterized by increased IEL, but lacking a dense lamina propria infiltrate of lymphoplasmacytic infiltrate. B: Another example of patient with clinical profile typical for lymphocytic colitis, but the mucosal biopsy showing only increased lamina propria lymphoplasmacytic infiltrate without increase in the IEL. C: Another variant of “microscopic colitis” characterized by the presence of numerous giant cells in the subepithelial location that has been called “giant cell colitis.” D: Granulomatous variant of microscopic colitis with poorly formed granulomas/large collections of histiocytes that were diffuse throughout the superficial lamina propria, remaining features being those of microscopic colitis. E: Occasionally, biopsies show a thickening of the subepithelial collagen band in one sample as illustrated here in a low-power view, and band of inflammatory exudate. This combination may be due to a “pseudomembranous variant” of “microscopic colitis.” |
Microscopic Colitis Variants
Several form of microscopic colitis has been described, usually as case reports or small series that appear to be variants of lymphocytic colitis.19, 20, 21, 22, 23, 24
Paucicellular lymphocytic colitis: Patients with a mean of <10 IELs/100 enterocytes have been diagnosed with “paucicellular lymphocytic colitis”
(Fig. 18-6A,B).127, 128 In addition, while most of these patients have a typical lamina propria lymphoplasmacytosis, some of these may have either patchy involvement or minimal to no increase in IELs (Fig. 18-6B),127, 128 in which case the distinction from normal or Brainerdtype diarrheas can be tricky. These patients seem to have similar clinicopathologic associations as classic lymphocytic colitis patients.22, 128 However, another study showed a different result suggesting paucicellular lymphocytic colitis does not behave like lymphocytic colitis,127 but that unlike most patients with typical collagenous colitis or lymphocytic colitis who have numerous CD25+ FoxP3+ (T-reg) cells in the lamina propria, these patients appear to lack them.127 The low number of IEL can be explained in some patients by a response to symptomatic treatment.
(Fig. 18-6A,B).127, 128 In addition, while most of these patients have a typical lamina propria lymphoplasmacytosis, some of these may have either patchy involvement or minimal to no increase in IELs (Fig. 18-6B),127, 128 in which case the distinction from normal or Brainerdtype diarrheas can be tricky. These patients seem to have similar clinicopathologic associations as classic lymphocytic colitis patients.22, 128 However, another study showed a different result suggesting paucicellular lymphocytic colitis does not behave like lymphocytic colitis,127 but that unlike most patients with typical collagenous colitis or lymphocytic colitis who have numerous CD25+ FoxP3+ (T-reg) cells in the lamina propria, these patients appear to lack them.127 The low number of IEL can be explained in some patients by a response to symptomatic treatment.
The number of IELs is also increased in biopsy samples from patients obtained following acute Campylobacter enteritis and in postdysenteric IBS. In this condition, in common with CD3+ lymphocytes in the lamina propria, the number of IELs also decreases over time.10 Some of the patients with paucicellular lymphocytic colitis may in fact represent a postdysenteric IBS. A focal lymphocytic colitis-like pattern was reported in four patients who eventually evolved to CD.129 Possibly, this entity of paucicellular lymphocytic colitis represents a heterogeneous group and more studies are needed to clearly understand its clinical implications.
Occasional cases with absence of increased IELs in the surface epithelium and of epithelial injury with increased IELs in the crypts have been reported as “cryptal lymphocytic colitis.”29 Surface intraepithelial neutrophils are usually few in number, but active crypt inflammation is relatively common. In these cases endoscopy can be abnormal. Cryptal lymphocytic colitis probably also represents a variant of lymphocytic colitis.
Giant cell and granulomatous lymphocytic colitis: Accumulation of giant cells just underneath the surface epithelium has been reported in a small number of patients with a clinical and endoscopic presentation suggestive of microscopic colitis (Fig. 18-6C). It is not clear if this so-called “microscopic colitis with giant cells” is really a distinct pathological entity. Response to therapy is comparable with that observed for collagenous and lymphocytic colitis. A granulomatous inflammation, characterized by scattered non-necrotizing granulomas, usually not associated with crypt epithelium (cryptolytic granulomas) has also been reported in patients with microscopic colitis (Fig. 18-6D), sometimes.21 The pattern must not be confused with CD that is typically focal and has obvious colonoscopic abnormalities. Another variant of microscopic colitis characterized by accumulation of clear CD68+ macrophages in the lamina propria in children has been reported.130 It was reported in about 2% of children presenting with abdominal pain, diarrhea, and some with constipation. It is unclear to us if these clear macrophages are truly different from muciphages commonly seen in colonic and rectal biopsies. Whether this is a distinct clinicopathologic entity will have to await further studies.
Pseudomembranous collagenous colitis: In some cases, a pseudomembranous reaction with an ulceration can be present in collagenous colitis as discussed earlier (Fig. 18-6E).23 Majority of the patient with this pattern don’t have concomitant C. difficle infection; however, this needs to be excluded clinically. In some of the patients this is associated with NSAIDs, while in some it may be manifestation of superimposed ischemic injury. All these variants seem to be rare, and unlikely to represent specific entities.25
Brainerd Diarrhea
An outbreak of chronic diarrhea of sudden onset occurred in 122 residents of Brainerd, Minnesota between December 1983 and July 1984 after which this form of colitis has been named.131 Subsequently, it has been associated with consumption of raw milk, and in another outbreak in 1992 with drinking water in an outbreak on a cruise ship.132 The disease has a characteristic sudden onset and marked urgency with a secretory diarrhea, a lack of systemic symptoms, and no response to antibiotics.
Histologically, colonic biopsies show a surface intraepithelial lymphocytosis without involvement of the crypt epithelium.13 The crypt architecture and the subepithelial collagen plate remain normal. There is no increase in lamina propria inflammatory cells. This superficial intraepithelial lymphocytosis was demonstrated in 20 out of 22 cases. Three cases had a focal active colitis, and two were completely normal. The overall histology tends to be more like lymphocytic colitis, rather than acute infectious/self-limited colitis. All cases had normal duodenal mucosal biopsies.13 The condition is self-limiting and usually resolves in few weeks; however, sometimes it may take as long as 18 months. It is thought to be of viral etiology.132 Making a histological diagnosis requires care as one is effectively attributing symptoms to a surface or crypt intraepithelial lymphocytosis (or both) but usually with little or no increase in lamina propria infiltrate (Fig. 18-6A), and suggesting that the disease is largely untreatable but will resolve on its own. It is usually worth some form of verbal communication with the clinician. If the intraepithelial lymphocytosis is only in the right colon then celiac disease needs to be considered.
Microscopic Colitis in Children
Microscopic colitis has also been reported in children mostly older than 4 years, although it remains uncommon with isolated case reports and few small series.49
Youngest child reported with collagenous colitis has been of 9 months of age, where collagenous colitis was associated with collagenous gastritis and collagenous spruce with no response to gluten-free diet.133 In general, children have more frequently concomitant collagenous gastritis, duodenitis, and other systemic autoimmune disorders including celiac disease and immunodeficiency syndromes.134, 135 Association with autism, H. pylori, and medications have also been reported.49 Response to treatment is variable and long-term outcome of microscopic colitis in children remains unclear.
Youngest child reported with collagenous colitis has been of 9 months of age, where collagenous colitis was associated with collagenous gastritis and collagenous spruce with no response to gluten-free diet.133 In general, children have more frequently concomitant collagenous gastritis, duodenitis, and other systemic autoimmune disorders including celiac disease and immunodeficiency syndromes.134, 135 Association with autism, H. pylori, and medications have also been reported.49 Response to treatment is variable and long-term outcome of microscopic colitis in children remains unclear.
Microscopic Colitis and Idiopathic Inflammatory Bowel Diseases
The number of IELs in UC, Crohn’s colitis, and infectious colitis is usually not increased, although a focal increase and a lymphocytic colitis-like pattern in CD have been described.136 The lymphocytic colitis-like pattern may precede the eventual clinical pathologic diagnosis of CD. In one study, four such patients eventually were diagnosed with CD.137 In this study, the biopsies from one patient showed a collagenous colitis morphology preceding the eventual diagnosis of CD. Colonoscopic abnormalities were however found in four of these patients making a diagnosis of microscopic colitis less likely.137 In a comment on this paper, it was stressed that colonic epithelial lymphocytosis without a thickened subepithelial collagen layer can be found in a clinically heterogeneous group of patients.129, 136 In a review of patients with colonic biopsies at the Mayo Clinic between 1978 and 1998, 15 patients (9 female, 6 male) were identified who had both a histological diagnosis of classic IBD and microscopic colitis.138 The mean age of the patients at the time of the microscopic colitis was 61 years. Thirteen patients had collagenous colitis. NSAIDs use was documented in 11 patients. In 12 patients the IBD was UC, in 1 it was CD, and the remaining 2 were diagnosed with indeterminate colitis. Either IBD or microscopic colitis may precede the onset of the other. The association of microscopic colitis and IBD is found predominantly in patients with pancolonic involvement.138 Progression of collagenous colitis to UC may also be observed in patients testing positive for perinuclear antineutrophil cytoplasmic antibodies (pANCA), although it is unclear if any specific clinical, laboratory, or pathologic findings reliably predicts future progression or development of IBD.139 Also, it is unclear if a diagnosis of IBD instead of microscopic colitis at that stage would have changed the patient outcome or the clinical management significantly. In summary, it seems reasonable to assume that there is a rare association between microscopic colitis and IBD. The nature of this association is however unclear whether it is by chance that two common disorders occur in the same patient, this represents progression of one disease to another, or misdiagnosis of one disease as another.
CHRONIC IDIOPATHIC INFLAMMATORY BOWEL DISEASES
Introduction
Chronic idiopathic IBD is a group of chronic relapsing diseases characterized by chronic diarrhea that can be bloody or watery, and includes two chronic GI disorders of unknown etiology: UC and CD. Sometimes, the nature of IBD remains undetermined also known as “indeterminate colitis” or IBD-unclassified.140 The presence of overt endoscopic changes differentiates this form of chronic colitis from microscopic colitis as discussed in the preceding section. However, these are all well-defined diseases, and term such as “nonspecific inflammatory bowel disease” is discouraged. Indeed the term “nonspecific,” whether used as a clinical, endoscopic, or histological prefix, is invariably a hedge for ignorance, and the omission of this “prefix” never changes the meaning of the sentence or diagnosis.
IBD tends to present typically with chronic diarrhea that may be bloody, often with a degree of abdominal pain, especially with CD. However, some patients initially present with acute symptoms and colitis, and only on follow-up turn out to be IBD. In one study, 46 patients (52.3%) out of a series of 88 adult patients presenting with acute colitis and followed for 2 to 3 years, relapsed and were subsequently classified as IBD: 54% UC; 33% CD, and 13% IBD-type unclassified.141 The mean age at onset was significantly lower in patients with IBD (32.3 years) than in patients with acute infectious or drug-related colitis (42.6 years).141 No clinical data (diarrhea, abdominal pain, bloody stool, fever, weight loss) were predictive of the final diagnosis. Follow-up of acute colitis may thus be indicated, especially when it occurs in patients <40 years.141 Disease severity at its onset, disease extent, and patient age at the time of diagnosis, along with other patient variables, determine overall disease severity and the likelihood of subsequent morbidity and mortality. Once established, IBD patients suffer episodic acute attacks that become superimposed on a chronic disease. As a result, the patients may suffer from disabling disease for decades. CD and UC must be considered in the differential diagnosis of clinically acute colitis because of differences in treatment strategies between infections and IBD.
Epidemiology
Neither UC nor CD are new diseases. A report on a series of patients likely representing UC was published during the 19th century. A patient presented by Lesniowski in Poland (1904) and a series published by Kenneth Dalziel (pronounced dee-yell) in 1913 constitute cases of CD more than 20 years before Dr. Burrill B. Crohn’s original description of the disease was published.142, 143, 144 This in turn likely should have had the senior surgeon (Dr. A. Berg) as the first author, but anecdote has it that he felt he did not need the publication at this stage in his career, which he recognized as being different from tuberculosis. However, Dalziel’s graphic description of the intraoperative appearances: “The affected bowel gives the consistence and smoothness of an eel in a state of rigor mortis, and the glands, though enlarged, are evidently not caseous,” which he called chronic interstitial enteritis, is convincing.
IBD has a worldwide distribution. Studies have shown a worldwide increase in the incidence of both UC and CD.145 Before World War-II, IBD was largely recognized in North America and northern Europe; however, since then increase in incidence has been recognized initially in central and Western Europe, Japan and Australia, then in South America and Eastern Europe. Now in the last decade, IBD has also been emerging in the vast Asia-Pacific region,146, 147 with an increase in, and strong correlation in the occurrence of UC and CD. Areas or populations with a high incidence or mortality attributable to UC also have a high incidence or mortality due to CD and vice versa. There is also a strikingly consistent pattern with regard to temporal trends for both diseases. An increasing incidence of UC generally precedes an increase in CD with a time lag of approximately 15 to 20 years. In most instances, the incidence of UC is higher than for CD, although the difference in their incidence rates varies among different parts of the world. In nations where IBD is emerging, UC constitutes a much higher proportion than CD; however, with time the incidence of CD has been increasing. This is true of Japan, China, and Eastern European countries.135, 148, 149, 150 In the case of CD, the incidence seems to level off around 6.0 per 100,000, although incidence rates as high as 15 to 16 per 105 have been reported for Canada and New Zealand. A plateau in the incidence has been found in five studies, but a continuous increase in the incidence has been demonstrated in four other studies involving populations from Derby, Florence, Iceland, and Copenhagen.151 The increase in incidence for CD has occurred for all disease locations, but has been relatively more common for colonic CD. For UC, the incidence is approximately 15 to 20 per 105 although higher figures have been reported. The increase in incidence has involved all age-groups except early childhood (<11 years) and elderly (>80 years). From a population-based cohort from Copenhagen County, it appears that approximately 7% of the patients diagnosed with UC and 6% of the patients with CD had an onset before 15 years of age. The median age at diagnosis was 12 years, indicating a steep increase in incidence around puberty. During the 26 years of the study, the mean incidence of IBD in children remained low at 2.2 per 105 individuals. UC occurs in all ages, but the major peak is in the 15 to 25 age range, but some suggest a second smaller peak in the sixties. It is extremely uncommon above age 75. In the old age-group, the distinction from chronic ischemia can be difficult.
A global north-south variation in the incidence of IBD has been documented. Standardized incidence rates of 10.9 to 12.8 for UC and 6.0 to 7.0 for CD have been reported from northern California and Scandinavia. Rates of 2.0 to 6.3 for UC and 0.9 to 3.1 for CD have been observed in the southern hemisphere. The results from a major study in Europe including 20 different centers in the early 1990s to test this hypothesis, suggest however, that the north-south gradient could be more of a historic phenomenon. For instance, rates are as high for CD in New Zealand as in Canada.150
The prevalence of IBD depends on the incidence, the survival, and the age distribution of the population in question. In a British study, two national birth cohorts (from 1958 and 1970) were followed and the prevalence rate was calculated at age 30 years for both UC and CD. A significantly higher prevalence of CD was found in the population cohort born in 1970 (375 per 105) compared to the result obtained in the cohort born in 1958 (211 per 105) reflecting the rising incidence of CD. In UC, similar prevalence rates were found for the two cohorts: 303 per 105 individuals and 273 per 105 individuals, respectively. As the proportion of elderly in a population increases, the prevalence of IBD is likely to increase as well.
A consistent finding in early observational studies for both CD and UC has been the association between an increased risk and high socioeconomic status based on differences in diet, crowding, and hygiene.150 The incidence in many developing countries is also increasing and changing lifestyles, changing trends in infections, and better recognition of IBD may all be responsible, rather than simply changing of socioeconomic status.
The association with ethnicity is to some extent a similar problem. In Crohn’s original description, all 14 patients were Jewish.144 Ever since then, there has been a perceived association between Jewish ethnicity
and an increased risk for IBD. This association has been substantiated by the results from studies in the United States, Sweden, United Kingdom, and South Africa. The results from incidence studies in Israel however are not so different from those obtained in other high-incidence areas such as the United Kingdom and Scandinavia. In the past, IBD was thought to be more common in whites than in blacks. In Cape Town, South Africa, incidence rates between 1980 and 1984 were eight times higher for whites for UC. Changes in trend have however been observed over the years and now it appears that incidence rates in blacks may actually be closer to those in Caucasians.152, 153
and an increased risk for IBD. This association has been substantiated by the results from studies in the United States, Sweden, United Kingdom, and South Africa. The results from incidence studies in Israel however are not so different from those obtained in other high-incidence areas such as the United Kingdom and Scandinavia. In the past, IBD was thought to be more common in whites than in blacks. In Cape Town, South Africa, incidence rates between 1980 and 1984 were eight times higher for whites for UC. Changes in trend have however been observed over the years and now it appears that incidence rates in blacks may actually be closer to those in Caucasians.152, 153
The sexes are affected approximately evenly, although there may be a very slight female predominance for CD with female to male ratios being 1.89 to 1.0. Studies using somewhat different methods and definitions show that women are at 20% to 30% greater risk to develop CD than men. The sex ratio is equal in UC.154, 155, 156, 157
It is worth noting that the children who migrate from developing countries to developed countries with higher incidence of IBD acquire the higher risk that of the country of residence, suggesting an important role for the environmental factors. For example children of South-East Asian origin, but raised in Canada or United Kingdom have taken on the higher risk of developing IBD that is prevalent in the country of residence.158
Etiology and Pathogenesis
Despite years of intensive research, the precise etiopathogenesis of CD and UC remains unknown. The current understanding of the pathogenesis of IBD implies a dysfunction of the interplay between genetic factors, environmental factors (the inciting agent may be environmental such as a bacterial infection), intestinal microbiota, epithelial cell dysfunction (leaky junctions), and aberrations of innate and adaptive immune response (Fig. 18-7). However, so far no single factor can be implicated and interplay between multiple factors seems to be involved, and different factors may be at play in different patients or patient populations. These factors are discussed in more details below; however, three factors seem to be essential and need to be mentioned here. First is the key role of the immune system as suggested by amelioration of CD in patients treated with immunomodulating drugs, including stem cell transplant (autologous or allogeneic).159, 160, 161 Second, allogeneic stem cell transplant from patients carrying the susceptibility genes for CD can transfer the disease to patients who do not have IBD arguing strongly for role of immune system and genetic factors.162, 163 Finally, microbes are essential to generate IBD, as in all experimental models of IBD the disease cannot be reproduced in a germ-free environment. While so far, no single pathogen has been consistently shown to be associated with IBD, it seems clear that the interplay between microorganisms and the immune system in a susceptible host plays a central role in the etiopathogenesis of IBD.
A real problem with IBD, especially in genetically susceptible individuals such as identical twins, is why one should get IBD and the other not. There clearly needs to be an initiating factor that promotes the gut “leakiness” that is then rarely completely downregulated again, so that patients rarely are able to go into complete permanent remission. There appear to be two factors that seem to act as initiators. One is smoking. Some patients with UC will clearly date the onset of symptoms to the time they quit smoking. Further, some of these patients are best controlled by smoking a few cigarettes a day. The second is an episode of acute infectious “gastroenteritis.” Although anecdotal, many patients will also note that they developed what had all of the appearances of an acute infection—often while visiting a high-risk area (vacation). The acute episode largely resolved but not completely, and when the diarrhea returned, it was with other features of IBD. The latter needs to be interpreted cautiously, as it is estimated that most of us get several episodes of acute diarrhea annually. However, an infection that precipitates bowel leakiness, and which, by virtue of genetic predisposition and other factors, the patient is never able to completely down-regulate, is not an unreasonable proposition.
Genetic factors. There is ample evidence that CD and UC are, in part, the result of a genetic predisposition. A positive family history of IBD seems to be the single most important factor determining an individual’s risk of developing IBD, while both family studies and twin studies indicate a stronger genetic effect for CD than for UC. Of 522 patients with CD, 187 (35.2%) were found to have an affected family member, of which 87 (16.7%) were first-degree relatives.164 The equivalent frequency in the group with UC was 20 out of a total of 171 patients (11.7%).164 Other more population-based studies show that the relatives of UC and CD probands have an 8- to 10-fold greater risk of developing IBD.165 In another study, Satsangi found that 41 of 317 (13%) patients with CD had a positive family history of IBD.166 The greatest risk is in those who have a monozygotic twin with CD or those who have two affected parents. The risk of getting IBD if two parents are affected is approximately 30% to 40%.167 Kirsner reported that of 103 families with more than one case of IBD, 31 families contained both CD and UC patients.168 Two UK studies have also demonstrated similar findings.164, 169 Epidemiologic studies in monozygotic and dizygotic twins, also indicate that
genetic factors may have a more dominant role in CD than in UC. The rate of concordance for CD has been reported to be 30.3% in monozygotic twins and 3.6% in dizygotic twins.170 When identical twins with CD have both IBD, the disease appears to be remarkably similar.171 With UC, the reported concordance rates are 15.4% for monozygotic twins and 3.9% for dizygotic twins.170 Significantly higher concordance rates in monozygotic twins for CD compared to UC suggest that genetic factors probably play a larger role in CD compared to UC.
genetic factors may have a more dominant role in CD than in UC. The rate of concordance for CD has been reported to be 30.3% in monozygotic twins and 3.6% in dizygotic twins.170 When identical twins with CD have both IBD, the disease appears to be remarkably similar.171 With UC, the reported concordance rates are 15.4% for monozygotic twins and 3.9% for dizygotic twins.170 Significantly higher concordance rates in monozygotic twins for CD compared to UC suggest that genetic factors probably play a larger role in CD compared to UC.
Now there are innumerable “IBD” genes and to date more than 100 IBD susceptibility loci have been identified of which about one third are associated with both UC and CD, one third with UC, and one third with CD (Table 18-5).172, 173, 174 A diverse array of genes has been implicated that play an important role in microbial recognition, lymphocyte activation, cytokine signaling, and intestinal epithelial defense. While there is striking overlap between UC and CD as genes involved with leukocyte trafficking are associated with both, some association patterns are distinct for each disease. NOD2 is predominantly associated with CD, while genes located in the major histocompatibility (HLA) region on chromosome 6p21 are predominantly UC associated.
The search for susceptibility genes had its first success in 1996, when the first susceptibility locus for CD was identified in the pericentromeric region of chromosome 16, which was called IBD1.175 Subsequently, the gene located at this site has been identified as nucleotide oligomerization domain containing 2 (NOD2) or also known as caspase recruitment domain-containing 15 (CARD15). In 2001, NOD2 gene was shown to be mutated in 20% to 30% of patients, primarily with terminal ileal CD, establishing a proof of principle for the genetic concept.176 Multiple mutations in the NOD2 gene have since been identified, three of which have been shown to be independently associated with CD (Cins1007fs, G908R and R702W). These mutations are consistently more common in CD patients (30%-40%) as compared to healthy controls in Caucasian population (10%-20%).177, 178 The relative risk for developing CD varies widely. Homozygosity for the polymorphisms confers the greatest risk (by a factor of 11-27), while the risk is lower with heterozygosity (by a factor of 1.75-4). NOD2 protein is an intracellular sensor of bacterial peptidoglycans and has been shown to be present in the cytoplasm of intestinal epithelial cells, Paneth cells, dendritic cells, macrophages, and endothelium. Activation of NOD2 by bacterial peptidoglycans leads to activation of nuclear factor kB (NF-kB) and mitogen-activated protein (MAP) kinase pathways which lead to production of various cytokines (e.g., TNF and IL-1β) and antimicrobial peptides. Also, within CD, the polymorphisms of the NOD2 gene seem to correlate with the clinical behavior (ileal disease involvement, stricturing, and perhaps internal fistulising disease). It is less common in patients who have perianal complications. As NOD2 variants seem to account for <30% of CD and because they are rare or absent in Asians, Arabs, Africans, and African Americans, other candidate susceptibility genes are likely involved and need to be investigated.178
After NOD2, the most common variants associated with CD are those involved in IL-23 pathway. As discussed later, this pathway of lymphocyte activation is thought to play a key role in the pathogenesis of IBD. In addition to IBD, polymorphisms in IL-23R have been associated with psoriasis and ankylosing spondylitis.
The autophagy-related genes, autophagy 16-like 1 (ATG16L1) and immunity-related GTPase M protein (IRGM), have been also associated with IBD, especially CD. Autophagy is the mechanism by which cells get rid of intracellular structures that include microbes, besides various intracellular organelles and apoptotic bodies. Involvement of these genes likely leads to impaired clearance of microbes, increased apoptosis, increased cell turn over, and inflammation.
The human leucocyte antigen complex (HLA) is a highly polymorphic, gene-dense region on chromosome 6p encompassing at least 130 expressed genes, many of which have immunoregulatory roles. Consistent evidence of linkage to IBD3 (6p21.1-23), an area which encompasses the HLA complex, has been demonstrated in several independent studies of both UC and CD. The most consistent associations are with the classical class II gene, HLA-DR. In UC, the strongest associations are seen with overall disease susceptibility.
Polymorphisms within the SLC22A4/5 genes, coding for the Organic Cathionic Transporters (OCTN) 1 and 2 on chromosome 5q are also associated with CD, indicating another possible functional and pathogenic link as these transporters are expressed by intestinal epithelial cells.
A number of other genetic polymorphisms have been reported to play a role in IBD pathogenesis. Polymorphisms causing a lower expression of the human multidrug resistance 1 gene product P-glycoprotein have been associated with the risk of developing UC or IBD in general. Toll-like receptors (TLRs) are receptors of the innate immune system involved in recognition of bacterial products. The TLR4 Asp299GLY polymorphism was associated with CD and UC in various studies, but the association has not uniformly confirmed. Increased intestinal epithelial expression of TLR4 has however been noted in IBD. Thus far, it appears that there are multiple susceptibility genes, some common to both diseases and some linked
separately to one disease or the other. For example, they can involve intracellular (NOD2) or cell surface (TLR) bacterial recognition, antigen processing (HLA molecules), or a breakdown in epithelial integrity (DLG5—Drosophila Discs Large Homolog).179 New loci for UC include genes implicated in mucosal barrier function (ECM1, CDH1, HNF4α, and laminin B1); further, E-cadherin is the first genetic correlation between colorectal cancer (CRC) and UC. Impaired IL-10 signaling also seems to be a key pathway in intestinal inflammation.180
separately to one disease or the other. For example, they can involve intracellular (NOD2) or cell surface (TLR) bacterial recognition, antigen processing (HLA molecules), or a breakdown in epithelial integrity (DLG5—Drosophila Discs Large Homolog).179 New loci for UC include genes implicated in mucosal barrier function (ECM1, CDH1, HNF4α, and laminin B1); further, E-cadherin is the first genetic correlation between colorectal cancer (CRC) and UC. Impaired IL-10 signaling also seems to be a key pathway in intestinal inflammation.180
Table 18-5 Genomic Regions Significantly Associated with IBD in Genome-wide Association Studies of Patients of European Ancestry | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Environmental factors
Infections Because of the similarities between infectious colitis or enteritis and CD or UC, infections have repeatedly been proposed as the etiology for both diseases. Overall, the relationship between infection and IBD is complex. Several intestinal infections such as Y. enterocolitica, C. jejuni and coli, Salmonella, Shigella dysenteriae, S. flexneri, S. boydii and S. sonnei, and Escherichia coli can not only mimic the clinical features of IBD (diarrhea, abdominal pain, and weight loss) but they can also mimic the endoscopic features and sometimes present with identical extraintestinal manifestations (reactive arthritis, erythema nodosum).181
An infectious etiology for CD was suggested by Dalziel way back in 1913, who emphasized the resemblance between the lesions observed in his patients with those observed in Johne’s disease (regional enteritis in cattle caused by Mycobacterium paratuberculosis).143 It was well recognized that granulomatous CD resembles intestinal tuberculosis, both clinically and in its mucosal pathology. Attempts to isolate mycobacteria from IBD tissue have been fraught with difficulties. The quest for a mycobacterial etiology of CD has been ongoing for many years. Because these fastidious organisms are extremely difficult to culture, and the leakiness of the gut in IBD casts doubt on the significance of any organisms found, M. paratuberculosis has been demonstrated in CD tissue and less so in UC tissue.182 However, no consistent results have yet been found to support the notion that this bacterium is specifically involved in the pathogenesis of either form of IBD, particularly CD. The clinical data that immunosuppressive treatment of IBD is usually beneficial, even in those with profound immunosuppression,183 would suggest that mycobacterial infection is unlikely to cause IBD.
Epidemiologic data certainly provide arguments in favor of a transmissible agent being responsible for UC and CD. Numerous pathogens including bacteria, viruses, and fungi have been suggested for both conditions as possible specific pathogens, yet there is no definite proof for any of the candidates. Candidates proposed include aerobic bacteria, mycobacteria, cell wall-deficient bacteria, anaërobic bacteria, and the common intestinal flora. None of the aerobic bacteria (Shigellae, E. Coli strains producing Shigalike toxin—for UC—Salmonella, C. jejuni, Yersinia, and
enterovirulent E. coli—for CD and UC) have been unequivocally identified as etiologic agents, although some may definitely initiate a first attack of IBD.184, 185 A number of groups in North America and Europe have identified a novel E. coli of the B2 and D phylotypes to be associated with CD that have been termed adherent-invasive E. coli.186, 187, 188, 189 Helicobacters have been isolated from human feces and human colon tissue, and in animal models they can cause an IBD-like disease, but on the evidence available, H. pylori does not seem to have a role in the cause of IBD.190, 191, 192, 193
enterovirulent E. coli—for CD and UC) have been unequivocally identified as etiologic agents, although some may definitely initiate a first attack of IBD.184, 185 A number of groups in North America and Europe have identified a novel E. coli of the B2 and D phylotypes to be associated with CD that have been termed adherent-invasive E. coli.186, 187, 188, 189 Helicobacters have been isolated from human feces and human colon tissue, and in animal models they can cause an IBD-like disease, but on the evidence available, H. pylori does not seem to have a role in the cause of IBD.190, 191, 192, 193
A viral etiology has also received some consideration.194 Arguments in favor come from case control studies, epidemiology, and the seasonal relationship of IBD relapses. Paramyxoviruses have been demonstrated by transmission electron microscopy (TEM) in giant cells and endothelial cells in patients with Crohn’s, but most studies failed to demonstrate viral particles.195 Viruses studied also include measles virus. In a review of the maternity records of 25,000 babies, four mothers had measles during pregnancy, and three of the four children had developed severe CD.196 The three cases showed the presence of the measles virus antigen by immunogold electron microscopy. The authors concluded that exposure to measles virus in utero or in the perinatal period inferred an increased risk of CD. The measles virus has been demonstrated by means of in situ hybridization and immunohistochemistry in several cases.195 Case control and other epidemiologic data support the hypothesis, but the geographical distribution of the measles virus is different from that of CD and the hypothesis has not been confirmed. A further study based on the UK population suggested that the risk might also apply to children vaccinated against measles.197 Immunohistochemical evidence has indicated that the measles protein apparently demonstrated in the human tissues is of human, and not viral, origin: the current conclusion is that there is no specificity to the association between measles exposure and subsequent CD.198, 199 During this same period and by same research group, another association was reported between MMR vaccine (mumps, measles, and rubella) and development of regressive autism, and IBD-like enterocolitis, which was labeled “autistic enterocolitis.”200, 201, 202 Careful review of these reports suggests that despite various GI symptoms that these children have, there is no evidence to support that this represents a true entity. Rubella virus, Epstein-Barr virus, and adenovirus infections of the respiratory tract were found to be associated with acute exacerbations of IBD in children, but a causative role is unlikely.203
The current thinking strongly supports the role of normal colonic flora (intestinal microbiota) rather than a specific pathogen in the initiation of the disease and its progression to chronicity.204 The intestinal microorganisms (microbiome) is acquired at birth, changes rapidly over the first year of life, and plays an important part in the development of immune system. In adults, in each individual the unique population of the intestinal microbiota remains fairly stable over time. Alteration in the normal intestinal microbiota may occur due to increased use of antibiotics, changing life style, change in environment, changing diet, etc. Experimentally, it was shown in rats that certain broad-spectrum antibiotics prevent chronicity following colitis induced by trinitrobenzenesulfonic acid.205 Bacteria or bacterial products seem to be critical to the induction of mucosal ulcerations in the small bowel by indomethacin, as germ-free rats develop minimal lesions.206 Bacteria have a wide array of influences. They may initiate or/and amplify the immune response either directly (antigens, toxins) or indirectly. Because of molecular analogy between some bacterial and human heat shock proteins (HSP), an autoimmune reaction could also be induced. Bacterial products (peptidoglycans) can be responsible for an immune reaction, and bacterial cell wall fragments can provoke a granulomatous response. Adherent bacteria may breach the intestinal mucosal barrier.207 So, it is likely that no single (or several) specific pathogen is responsible in the pathogenesis of IBD. In contrast, alterations in the “normal” intestinal microflora composition (dysbiosis) may be responsible for (initiation and) amplification of the increased intestinal inflammation. Bacteroides vulgatus and E. coli were detected more frequently and in greater numbers in samples from patients with IBD.208 Enterobacteria seem significantly increased in active and quiescent CD compared with healthy controls and more than 30% of the dominant flora in CD would belong to undefined phylogenetic groups.209 Intestinal lesions (ulcers and fissures) are colonized by E. coli and streptococci in over 60% of the cases studied.210, 211
The hypothesis implying a role for microorganisms in the pathogenesis of IBD, and especially CD, is strengthened by the increased familial occurrence of the disease as well as by the observations obtained with genetic studies. The proteins encoded for by NOD2 are involved in the recognition of bacterial products.212 Further evidence comes from studies investigating the expression of human a-defensins, components of the innate immunity which contribute to local intestinal host defense in microbial infections and from studies examining TLRs.213 The latter are also involved in the recognition of bacteria and act as extracellular receptors. One study has shown that children with IBD were more likely to have used antibiotics in the first year of life (which may impact on the developing gut flora) than matched unaffected controls.214
The search for a microbial cause for UC has been inconclusive, but has produced fewer false trails than
similar work on CD.215, 216 Attempts to identify infective agents in the stool and mucosa or to demonstrate raised antibody titers to particular organisms in patients’ sera are problematic. The difficulty is to establish whether an abnormality, when present, is the primary defect or simply a secondary phenomenon. Patients with UC have raised antibody titers to numerous organisms including a lipopolysaccharide extract of E. coli O14, an antibody common to most Enterobacteria.217, 218 This antibody cross-reacts with goblet cell antigen in colonic epithelium.219 Similar antibodies are found in first-degree female relatives who have an undamaged mucosa.220, 221 The sharing of antigens between the large intestine and intestinal bacteria has also proved the basis of methods for inducing experimental colitis in animals.219, 222 Raised antibody titers have also been demonstrated to certain Bacteroides spp., Eubacterium, Peptococcus, and Mycobacterium kansasii antigens,223, 224 possibly suggesting a leaky mucosa for bacteria and secondary antibody generation. ASCA, ANCA, and similar antibodies may be generated on the same basis.
similar work on CD.215, 216 Attempts to identify infective agents in the stool and mucosa or to demonstrate raised antibody titers to particular organisms in patients’ sera are problematic. The difficulty is to establish whether an abnormality, when present, is the primary defect or simply a secondary phenomenon. Patients with UC have raised antibody titers to numerous organisms including a lipopolysaccharide extract of E. coli O14, an antibody common to most Enterobacteria.217, 218 This antibody cross-reacts with goblet cell antigen in colonic epithelium.219 Similar antibodies are found in first-degree female relatives who have an undamaged mucosa.220, 221 The sharing of antigens between the large intestine and intestinal bacteria has also proved the basis of methods for inducing experimental colitis in animals.219, 222 Raised antibody titers have also been demonstrated to certain Bacteroides spp., Eubacterium, Peptococcus, and Mycobacterium kansasii antigens,223, 224 possibly suggesting a leaky mucosa for bacteria and secondary antibody generation. ASCA, ANCA, and similar antibodies may be generated on the same basis.
While the role of specific infections in IBD is still unclear, infections can certainly cause exacerbation of the disease. Patients with established IBD who present with flare-ups often have superimposed infections. Viral infections, and other opportunistic infections especially cytomegalovirus (CMV), are a possible cause of exacerbation of IBD, particularly in fulminant cases.225, 226, 227 Extraluminal spread of enteric bacteria is one of the most frequent complications of IBD. Common and possible clinical presentations of this complication are intra-abdominal abscesses, fistulae, postoperative infections (wound infections), bacterial overgrowth, and bacterial translocation to mesenteric lymph nodes in patients undergoing surgery (or not) with systemic spread with sepsis.228, 229, 230, 231
Psychological Factors In the early and middle decades of the 20th century, a primary psychosomatic origin was considered for UC and to a lesser extent in CD. The theories developed out of retrospective studies showing an association between “well-marked time relationship between emotional disturbance and symptoms.” A particular psychological profile or personality type was thought to predispose to the development of UC. Psychosocial factors possibly contribute to IBD and include sexual abuse in childhood, inefficient coping strategies, depressive disorders, chronic life stress, and threatening life events.232 Patients with CD seem to have slightly higher frequencies (up to 50%) of psychological disturbances and lower quality of life compared to patients with UC and to those with other chronic diseases. While these disturbances could be a consequence of the illness rather than being etiologic, a recent population-based study reported that IBD patients were over twice as likely to have a lifetime diagnosis of mood disorders long antedating their IBD diagnosis than community controls.233 The degree of psychological distress correlates with the disease severity. However, there is evidence of an interaction between psychological factors and disease activity. Depression and perceived chronic distress seem to represent risk factors for relapse of the disease.234, 235, 236 Studies of the effect of stress on colonic mucosa reinforced this concept.237 At present, it seems that an association of psychological factors with the etiology of IBD is unproven, although there is evidence for a moderate influence on the course of the disease.
Diet and Allergy Diet has been considered to play a role in IBD for a long time. In the early part of the 20th century, nutritional deficiencies were suggested to be causes of IBD. As the rising incidence in the second half of the 20th century coincided with profound changes in dietary habits, various lifestyle aspects have been considered to play a role. Breastfeeding as well as longer duration of breastfeeding has been suggested to be protective against the development of CD and UC. An increase in carbohydrate intake, simple carbohydrates in particular, and/or fast food have been associated with a risk for CD. For UC, a positive correlation between consumption of soft drinks and chocolate was reported.238, 239 In another study, a high fat intake was associated with an increased risk for UC, whereas a negative correlation with vitamin C and fruit consumption was found.240 All of these findings may however also be an expression of modern dietary practices. However, in a recent study of UC, it was shown that dietary factors such as red and processed meat, protein, alcohol as well as sulfur and sulphate intake were positively associated with relapses. These results may be related mainly to the sulfur and sulphate contents of the food. Sulfur not recycled by an intestinal sulfur salvage pump passes into the colon where it is metabolized by sulphate reducing bacteria, which were shown to be common in individuals with UC. These bacteria could yield compounds, which in turn interact with other luminal contents and generate sulfoxides, which can be injurious to the colonic mucosa. Confirmation of these findings is however needed.241, 242 Ingestion of toothpaste has been linked to IBD and CD in particular.238 A variety of substances in toothpaste including particulate substances used as abrasives, such as tricalcium phosphate and quartz, have indeed been identified as being capable of penetrating the epithelium and creating lesions similar to CD, in animal studies. Other particles considered include talc, carrageenan, which has been used in an animal model of colitis, and other abrasive agents such as silicates and calcium pyrophosphate. This
hypothesis has not been rigorously tested. The concept has however been broadened to include a large range of microparticles ingested, as part of the diet. These particles are taken up by the specialized M cells that cover the lymphoid follicles, especially Peyer’s patches. They are undegradable and accumulate in lymphoid tissue. They may act as adjuvants, permitting the absorption of other antigens and by doing so, alter the immune system.
hypothesis has not been rigorously tested. The concept has however been broadened to include a large range of microparticles ingested, as part of the diet. These particles are taken up by the specialized M cells that cover the lymphoid follicles, especially Peyer’s patches. They are undegradable and accumulate in lymphoid tissue. They may act as adjuvants, permitting the absorption of other antigens and by doing so, alter the immune system.
Allergy, another potentially food-related process has also been considered as a possible underlying cause of IBD. An allergy to cow’s milk was particularly suggested with regard to UC.243 Specific antibodies to proteins in milk have been found to correlate with disease activity in adults, particularly in UC, but these findings are inconsistent and may reflect rather increased mucosal permeability.
Increased Antigen Load and Permeability Increased intestinal permeability has also been proposed as a possible factor in the development of IBD, especially CD.244 This possibly leads to increased exposure of the immune system to the antigens that cross the intestinal barrier. Numerous studies have documented increased intestinal permeability in patients with Crohn’s and even asymptomatic relatives, but not in those with UC. These observations have led to the “leaky gut” hypothesis.245 One of the early lesions, described in CD is the aphthoid ulcer, which occurs primarily in the region of M cells overlying lymphoid follicles. M cells are specifically modified epithelial cells, which sample intestinal organisms and antigens. In addition, there is also some morphologic evidence that intercellular tight junctions are abnormal in diseased and nondiseased areas in CD. This would explain the increased permeability in CD. In UC, basic epithelial lesions are less apparent. There is however some evidence of abnormal mucin production and for a potential role of peroxisome proliferators-activated receptors (PPARs).222, 223, 224 In UC, the expression of PPAR-γ is indeed impaired.
Miscellaneous It has also been suggested that the bowel is being exposed to either new bacteria or concentrations of bacteria which did not exist before the 20th Century, or that the balance of intestinal flora has been altered (dysbiosis) possibly because of antibiotic use; or because bacteria which can survive at low temperatures are ingested chronically, as a result of the widespread use of refrigeration for preserving food (cold chain hypothesis).214, 246, 247 Another suggestion involves the loss of helminths as a relatively recent change in the intestinal flora and as the fundamental permissive factor, enabling the emergence of IBD.248, 249 The role of oral contraceptives in the pathogenesis is uncertain and contradictory.
Depletion of goblet cell mucin is a characteristic feature of UC and mucus has an important role in preserving the integrity of the colonic mucosa against trauma and bacterial attack. Primary abnormality of colonic mucus has been found, several components of colonic mucin have been identified, and a reduction in one type has been found in UC, even in patients in remission.250 Changes in mucin pattern are not related to alteration in bacterial fecal degradation enzymes or to any differential susceptibility of mucus in UC to desialylation or desulfation.
Smoking and IBD. Another factor involved in the etiology of IBD is smoking. The relationship, first reported in 1984, is complex.251 In contrast to CD, smoking is preventative in UC. Some patients have their first attack when they stop smoking, while a small subset are best controlled by unobtrusively smoking a few cigarettes a day. Interestingly, nicotine in other forms seems not to have the same effect. This could be linked to the finding that the colonic mucosa of smokers demonstrates increased glycoprotein synthesis, compared to nonsmokers, which would help maintain the protective colonic mucosal barrier.252, 253, 254 Patients with CD are unequivocally more likely to be smokers, and smokers have a higher likelihood of disease relapse after medically or surgically induced remission.255 However, smoking is more likely to modulate the disease than cause it since the countries with the highest rates of smokers are among the countries with lowest rates of CD (mostly in the developing world). The relationship between smoking and IBD may be explained by modulation of the cellular and humoral mechanisms of immunity. Other potential mechanisms include changes in cytokine and eicosanoid levels, gut motility, permeability, colonic mucus, blood flow, and oxygen radicals.255
Appendectomy and IBD. An intriguing feature of UC is the possibility that prior appendectomy may be protective against its development. Several epidemiologic studies have shown a significant negative association between appendectomy prior to the diagnosis of UC and the development of UC. In a meta-analysis of 13 case-control studies with evidence from 2,770 patients with UC and 3,352 controls, an overall odds ratio of 0.307 (95% CI = 0.249-0.377) was found.256 The association is stronger when the patients are operated before the age of 20 years for an inflamed appendix. The data also suggest that appendectomy may also delay disease onset as well as producing a milder IBD phenotype,257, 258 although the extent of disease may well be greater.259 In one prospective study of patients with ulcerative proctitis, who underwent appendectomy in the absence of history suggestive of previous appendicitis, 90% of them improved their
clinical activity index, with 40% experiencing remission by 12 months. Appendiceal histology in 29 cases demonstrated ulcerative appendicitis, but no patient had any evidence of typical acute appendicitis.260 A randomized control trial of appendectomy as therapy for UC is clearly needed. Many studies show however severe methodologic shortcomings.261 Some suggest that it is appendicitis and not appendectomy, which is protective.262 The mechanism of this association is unclear. Immunologic mechanisms have been proposed as the basis of the relationship between UC and appendectomy.
clinical activity index, with 40% experiencing remission by 12 months. Appendiceal histology in 29 cases demonstrated ulcerative appendicitis, but no patient had any evidence of typical acute appendicitis.260 A randomized control trial of appendectomy as therapy for UC is clearly needed. Many studies show however severe methodologic shortcomings.261 Some suggest that it is appendicitis and not appendectomy, which is protective.262 The mechanism of this association is unclear. Immunologic mechanisms have been proposed as the basis of the relationship between UC and appendectomy.
The appendix differs from the colon by the presence of abundant organized lymphoid tissue in the mucosa spread around its luminal circumference and by the different composition of the surface and crypt epithelium. It has been estimated that the amount of lymphoid tissue in the appendix is equal to that in the ascending, transverse, and descending colon. In the appendix, bacterial translocation as part of the normal antigen-sampling processes starts from the second week after birth and lasts throughout the first year of life.263 The appendix may act as a reservoir or safe house for maintaining large bowel flora, serving as a reservoir for normal flora, Its removal might therefore disturb normal large bowel flora, or, if the response is immune mediated, raises the possibility of immune mimicry between appendiceal and rectal mucosa, so that inflammation at one site induces inflammation at the other. Much of the time, the activity at one site is similar to the other in UC, suggesting they behave as a single unit. However, periappendiceal and cecal patches are also found in other forms of IBD.
The immune system. Abnormalities of the innate and adaptive immune system have been proposed as both primary and contributing defects to IBD. The innate immune response provides an initial and rapid response to microbes, compared to the adaptive immune response. However, once the adaptive immune system has been exposed to antigen, it also mounts a rapid and a very robust response on subsequent exposure. The cells of the innate immune system have receptors that recognize general microbial patterns (pattern recognition receptors) rather than specific antigens. Cells in the innate immune system include neutrophils, macrophages, dendritic cells, and NK cells. The role of Paneth cells in innate immune response has been discussed in detail in Chapter 16. Goblet cells release the mucin that covers the epithelial surface, which prevent binding of the microbes to the epithelium. Mice that lack the gene encoding MUC2, a component of mucin, develop spontaneous intestinal inflammation.264
The importance of the innate immune system in IBD is supported by many observations. Patients with genetically defined syndromes of neutrophil and monocyte dysfunction, such as chronic granulomatous disease and glycogen storage disease type-Ib develop colitis that is indistinguishable from CD (see later). This is further supported by identification of genetic polymorphisms in IBD that play an important role in microbial recognition like NOD2 and autophagy genes (ATG16L1 and IRGM). Autophagy is the mechanism by which cells get rid of intracellular structures including intracellular organelles and apoptotic bodies, as well as intracellular microbes. Identification of defensin modifications in patients with CD also supports this hypothesis. An abnormal innate immune system may not be able to adequately contain intestinal bacteria, which then may lead to activation of T- and B-cells with secondary tissue injury.
Innate immune cells interact with either T- or B-cells to initiate an adaptive immune response, and aberrations of adaptive immune response are also thought to play an important role in IBD; T-cell dysregulation is a key factor in initiation and progression of chronic inflammation.265 CD4+ T helper (H) cells regulate critical aspects of the acquired immune response. Two of the major subtypes have been classified as either T-helper-1 (TH1) or TH2 on the basis of function and according to their ability to elaborate specific cytokines. Besides these two helper cell subtypes, TH17 cells have now also been shown to play a critical role in the pathogenesis of IBD. TH1 cells orchestrate cell-mediated immune responses and are characterized by the ability to secrete interleukin (IL)-2, IL-12 and interferon-γ (IFN-γ). TH2 cells, in contrast, mediate humoral responses and secrete IL-4, IL-5, IL-6, IL-10, and IL-13. TH17 cells are characterized by secretion of IL-17, but they also secrete other cytokines such as IL-21 and IL-22. IL-23 secreted by macrophages and dendritic cells leads to proliferation and/or increased survival of TH17 cells. This IL-23 pathway is central to the TH17 function, and polymorphisms of various genes involved in this pathway have now been associated with both UC and CD.265 The various T cell subsets regulate each other reciprocally. Despite the wide variety of causes, mucosal inflammation is almost always mediated by dysregulation in TH1, TH2, and/or TH17-cell responses, ultimately leading to an increase of the T cell subsetassociated cytokines within the intestinal tissues.
Intestinal TH1 activation and cytokine release is associated with the generation of activated matrix metalloproteinase, which are essential mediators of tissue destruction. Moreover, cytokines act on the local microvasculature, up-regulate adhesion molecules, and enhance recruitment of additional effector cells, including neutrophils and phagocytes, which contribute to amplification of the inflammatory response and subsequent tissue damage.
Once activated, down-regulation of the immune response at an appropriate time is equally important in maintaining the intestinal homeostasis. Activated mucosal T cells are under the control of at least two mechanisms. First, regulatory T-cells are able to control or prevent inflammation through the production of IL-10 and tumor growth factor-β (TGF-β), as well as through direct cell-cell interactions. Second, induction of apoptosis in activated T cells can end the immune response. In CD, mucosal T cells can be relatively resistant to apoptosis, which leads to accumulation of T cells and persistence of inflammation. In UC, T cells are more susceptible to FAS-mediated apoptosis.
In comparison to T-cells, the role of B-cell has not been extensively studied in IBD. UC is associated with production of various autoantibodies such as anti-neutrophil cytoplasmic antibody (pANCA) and antitropomyosin. Similarly in CD antibodies to Saccharomyces cerevisae, outer membrane porin C (OMP-C) and a murine flagellin (Cbir-1) are seen, which are somewhat specific for CD, although each is <50% sensitive. The presence of these various autoantibodies is suggestive of increased B-cell activation, although their role in the development or progression of the disease is unclear.
 Figure 18-7. Complex interplay between microbes and various components of the innate and adaptive immune system in the pathogenesis of IBD. Normally secreted IgA, antimicrobial peptides secreted by Paneth cells, and mucin (MUC2) secreted by goblet cells provide protection against microbes trying to reach the epithelium. The epithelium with its tight junctions provides an additional protective barrier. When luminal bacteria do come in contact with the epithelium or enter the tissues release of various cytokines by the epithelium, and activation of tissue macrophages and dendritic cells lead to activation of the adaptive immune response. Macrophages are in addition able to kill the organisms by phagocytosis. Under physiologic condition the innate and adaptive responses are able to clear the bacteria without setting of chronic inflammation and secondary inflammatory injury to tissues. Under pathologic situation, either luminal bacteria or other environmental toxins/antigens are able to cross the mucin and epithelial barrier (leaky epithelium), activate dendritic cells, and macrophages. The dendritic cells lead to activation of naïve CD4+ T helper cells (TH) in the Peyer’s patches and mesenteric lymph nodes. TGF-β and IL-10 modulate the differentiation of CD4+ cells T-cell subgroups (H1, TH2, TH17, and regulatory T-cells) with their characteristic cytokine profile as shown. These activated T-cells then move back to the lamina propria to carry out their effector functions. MP, macrophage; NK, natural killer cell; DC, dendritic cell; IL, interleukin; TNF, tissue necrosis factor; TGF, transforming growth factor; Th, T-helper. |
In summary, the pathogenesis of IBD is thought to be due to a dysregulated response to intestinal microbiota secondary to a combination of inappropriate innate immune and adaptive immune responses in the genetically susceptible host as shown in Figure 18-7.
Ulcerative Colitis
UC is a chronic inflammatory disease (IBD) of the large intestine, which, most typically begins in the rectum. It is unclear when was the disease first recognized, although certainly, some of the diarrheal diseases described during the Roman era may represent UC. The term UC to describe this entity was first used by Sir Samuels Wilks in 1859.266 Some years later, the
Surgeon General of the Union Army while describing the history of medicine during American civil war described UC and illustrated the histology using photomicrographs, a remarkable achievement for the time.267 Within several years, the disease was well characterized and in 1909 at a symposium of Royal Society of Medicine, no less than 300 case have been collected from various London hospitals.266 The term UC is a poor one as ulceration is not a sine qua non for the disease and the word colitis fails to recognize that rectal involvement rather than colonic disease is the prerequisite for the diagnosis. However, the term has stuck and has been well ingrained into the medical literature such that until some major etiopathologic breakthrough leads to a newer nomenclature, there is no good reason to abandon it!
Surgeon General of the Union Army while describing the history of medicine during American civil war described UC and illustrated the histology using photomicrographs, a remarkable achievement for the time.267 Within several years, the disease was well characterized and in 1909 at a symposium of Royal Society of Medicine, no less than 300 case have been collected from various London hospitals.266 The term UC is a poor one as ulceration is not a sine qua non for the disease and the word colitis fails to recognize that rectal involvement rather than colonic disease is the prerequisite for the diagnosis. However, the term has stuck and has been well ingrained into the medical literature such that until some major etiopathologic breakthrough leads to a newer nomenclature, there is no good reason to abandon it!
UC is a disease characterized by periods of exacerbation and remission. Less commonly, there is continuous low-grade activity or an initial single attack, which years later causes questions as to whether this was really an infection (hence the need for initial biopsy documentation), or presentation with severe disease, sometimes with dilatation (toxic megacolon). In many patients, disease is extensive at first presentation, and there is nothing to suggest the disease was ever anything other than extensive. Attempts have been made to explain the variability of extent of colitis amongst patients; genetics may be important in deciding the extent of disease while anatomical variations in arterial supply to the colon, we believe, is a less likely explanation.268
While it is a disease of the large bowel, however, involvement of the terminal ileum can occur in patients with pan-colitis, and more recently, histologic involvement of the upper GI tract in a minority of patients has been recognized. The disease in some patients remains limited to the rectum (ulcerative proctitis), while in others, it may extend proximally to involve a variable length of the large intestine, and sometimes the entire large bowel (pan-colitis) in a continuous or diffuse fashion, although the changes are nearly always more severe in the distal large intestine. Proctitis (9%), proctosigmoiditis (23%), and left colonic involvement (24%) are the most common forms. Pancolitis is observed in 11% to 30% of the cases.269 It has been increasingly recognized that sometimes there can be rectal-sparing endoscopically and/or histologically or patchy distribution of inflammation, but these findings may enhance the likelihood of using the term IBD-type unclassified (or indeterminate colitis—a term initiated for uncertainty on review of colectomy specimens). Nonetheless, endoscopic rectal sparing, especially initially, can still be consistent with an ultimate diagnosis of UC and patchiness is common in treated patients.270 UC with rectum having a completely normal appearance both histologically as well as endoscopically, is rare, but anecdotally does seem to exist, although the rectum ultimately does become involved. This can be clearly seen in diverticular-associated colitis which eventually evolves into UC, and possibly in some patients with colitis accompanying primary sclerosing cholangitis (PSC).
UC is primarily an inflammatory condition of the mucosa although in cases of severe disease, deeper layers of the bowel wall can be involved. The appearance of the bowel depends much on the severity and duration of the disease. While CD may also be initially mucosal, inflammation has a much greater aptitude to involve deeper layers of the bowel. It is important for pathologists to recognize that in some patients there can be reversal of both endoscopic and histologic changes, to the point that the biopsies appear absolutely normal. This is seen especially in surveillance biopsies in patients with long-standing disease. The implication of this is that while we can report “no evidence of active IBD seen,” we can never say that the patient does not have IBD. Return to normality after treatment of classical UC may result in so-called “rectal sparing” that has been occasionally reported.271, 272, 273
Clinical features. Patients with UC typically present with bloody diarrhea. Abdominal pain is much less prominent than in CD, but abdominal cramps especially just prior to bowel movements is common. Fecal incontinence and nocturnal awakening for bowel movements should raise “red flags” that patients who may have been suspected of having IBS in fact may have an inflammatory disorder. Weight loss is another “red flag.” Bloody diarrhea is colitis until proven otherwise, although some patients with bloody diarrhea may have a noncolitic diarrheal disorder with bleeding hemorrhoids. Fatigue, fever, and other extraintestinal manifestations like arthralgias or frank arthritis, and inflammatory diseases of the skin and eye may accompany the presentation of UC or occur later independently. Physical examination of the patient is often unrevealing other than finding blood on the examining finger on digital rectal exam. When patients present with possible UC, stool samples should be checked for bacterial infection, especially C. difficile and for parasites if there is reason to suspect contact with parasites that could induce colitis (i.e., after travel to endemic areas). The blood work usually obtained includes a complete blood count, serum ferritin, albumin, and C-reactive protein. The liver enzymes are also checked since inflammatory liver diseases may occur in approximately 3% of patients, but also as a baseline since many drugs used to treat the disease may occasionally cause liver injury. Serum perinuclear antineutrophilic cytoplasmic antibody
(pANCA) are positive in 70% of patients and this can distinguish patients from those having infectious colitis; however, it is less helpful in distinguishing from Crohn’s colitis since up to 50% of these patients may be pANCA positive.274 Endoscopy with biopsy though is the key diagnostic testing to be undertaken.
(pANCA) are positive in 70% of patients and this can distinguish patients from those having infectious colitis; however, it is less helpful in distinguishing from Crohn’s colitis since up to 50% of these patients may be pANCA positive.274 Endoscopy with biopsy though is the key diagnostic testing to be undertaken.
Endoscopic features and gross pathology. The serosa is intact and retains its normal shiny surface though there is considerable congestion and dilatation of blood vessels. The exception is severe disease that has become transmural, including toxic dilatation, either of which can perforate. The regional lymph nodes are sometimes enlarged due to reactive hyperplasia, but do not contain granulomas (unless there is another cause). The length of the colon and rectum may be shortened, sometimes markedly so with obliteration of the sigmoid loop. This shortening appears to be due to contraction of both muscle layers and is most obvious in the distal colon and rectum. Fibrosis is often present in the submucosa, which may be completely obliterated. Clinically, shortening of the colon is sometimes reversible if remission is maintained.275 The contraction is also accompanied by a reduction in the transverse caliber, which is also particularly marked in the distal large bowel. In the rectum it accounts for the increase in the sacrorectal distance, which is an important sign in the radiographic diagnosis of UC. Involvement of the terminal ileum in UC (so-called backwash ileitis) is rarely seen on external examination. Severe involvement is rare, but can be associated with a tendency to dilatation, rigidity, and muscular thickening of the bowel wall. When these changes are present, the possibility that this is CD should always be considered.
Surgical resection is usually performed for chronic extensive colitis resistant to, or dependent on, unacceptable levels of therapy, for severe disease, and in patients with other complications including dysplasia or carcinoma. Thus study of surgical specimens therefore shows only a limited spectrum of the disease. The spectrum of mucosal appearance is wide and depends on the duration of the disease, disease activity, and prior therapy. On opening a fresh surgical specimen of UC, the first notable feature in active disease is the amount of dark fluid blood present in the lumen, justifying the use of the French terminology rectocolite hémorrhagique. The mucosa is typically red, granular or velvety, and is extremely friable with bleeding on contact (Fig. 18-8). The earliest form of macroscopically recognizable mucosal damage is redness with a prominent vascular pattern and erosions with purulent foci (Fig. 18-8). The various gross features are shown in Figures 18-9 and 18-10. In advanced cases, the mucosal fold may be lost and the mucosa may appear flat. In some, the mucosa tends to have a beefy red appearance without any discrete ulcers. Redness is probably due to a combination of capillary dilatation, mucin depletion, and interstitial hemorrhage. The mucosa may show numerous petechial hemorrhages. Haustrations tend to disappear. Various sized ulcers can be seen. They can be small, rounded and superficial, or more irregular, somewhat geographic in configuration. Fissuring ulcers are not seen, except in some cases of toxic megacolon. Ulcers can become more extensive and undermine the mucosa so that mucosal bridges (Fig. 18-11) with an underlying inflammatory infiltrate develop. In the radiology literature, these ulcers have been called “collar button ulcers” (Fig. 18-12). Full-thickness ulceration of the mucosa is usually patchy, but any intact intervening mucosa is always diseased. Multiple confluent ulcerations result in complete or extensive denudation of the mucosa (Fig. 18-12). In severe disease the ulceration may have a linear distribution, especially in the colon where it is related to the line of attachment of the taeniae coli. When ulcers are deep, the inner circular muscle coats may become visible from surface and full thickness penetration may occur in fulminant UC. Extensive longitudinal ulcers, especially if connected by transverse ulcers, are not a feature of UC but rather of CD. Islands of mucosa may become entrapped in the submucosa and can eventually become cystically dilated presenting as “colitis cystica profunda.” In the more advanced stages, the entire bowel becomes fibrotic, narrowed, and shortened. Conversely, in resections with prior steroid therapy or when performed for dysplasia, the disease may be completely quiescent. Even in quiescent disease, loss of mucosal folds and shortening of the bowel may be evident, despite lack of ulcers or lack of mucosal inflammatory changes.
In general, the severity of the mucosal changes in surgical specimens is usually greatest in the distal large bowel and tends to diminish proximally. Even in pancolitis, the disease is usually more severe in the left colon and rectum. This pattern is very useful when seen in sequential biopsies. Macroscopically, the proximal limit of the disease most frequently shows an abrupt transition from disease to normal mucosa, but a gradual change is more usually seen histologically.
One of the significant features of UC is the relative lack of fibrosis in the muscularis propria or lamina propria, although thickening of the muscularis mucosae is often accompanied by a degree of adjacent submucosal fibrosis, marking the sites of prior ulceration. There may be an increase in the amount of collagen in the superficial submucosa, particularly of the rectum in some patients, but this is quantitatively very small even when there is a long history of
severe disease. Rarely, strictures occur on the basis of hyperplasia of the muscularis mucosae with submucosal fibrosis (Fig. 18-13).276 If true fibrous strictures are present in a colon showing diffuse inflammation, the diagnosis of CD, rather than UC, should be considered. Alternatively, stricture formation can be the result of coexistent diverticular disease or of malignant change.
severe disease. Rarely, strictures occur on the basis of hyperplasia of the muscularis mucosae with submucosal fibrosis (Fig. 18-13).276 If true fibrous strictures are present in a colon showing diffuse inflammation, the diagnosis of CD, rather than UC, should be considered. Alternatively, stricture formation can be the result of coexistent diverticular disease or of malignant change.
 Figure 18-8. A: Endoscopic images showing active UC characterized by friable mucosa (bleeding on contact with the scope) and extensive purulent exudates. B: Prominent granularity and edema of the mucosa covered by mucoid material is seen here. C: Another example showing more chronic changes with loss of vascular pattern and mucosal folds, but no active lesions. |
 Figure 18-9. Ulcerative colitis. Gross appearances. A: Shortened, haustraless bowel with narrowing in the left colon. An inflammatory polyp is present in the transverse colon. B: Total UC extending throughout the large bowel, with extension in continuity for a considerable length of the terminal ileum. |
 Figure 18-9. (Continued) C: UC affecting the large bowel in continuity from the rectum (bottom left) to the ascending colon (top left). Although the junction between the affected and unaffected bowel appears sharp, careful inspection of the remaining ascending colon reveals numerous well-defined areas of ulceration similar to a cecal patch. This could pose a diagnostic problem in multiple biopsy specimens, which would demonstrate markedly focal disease in this area, perhaps raising the question of CD. Fortunately, total colonoscopy is not usually carried out in patients with severe disease, but focality at either end in severe disease does occur. D: Another example of left-sided colitis with hemorrhagic friable mucosa that bleeds on touch. E: UC in a patient involving almost the entire colon with fulminant disease. The involvement of the right side shows many areas, which create an impression of patchiness; however, these merely represent areas of more severe inflammation. Histologically, the entire bowel showed mucosal inflammation typical of UC. F: Colitis involving almost the entire colon, sparing only the right side with the involved mucosa looking edematous red and granular. G: The close-up showing sharp demarcation between the involved and uninvolved segments. |
 Figure 18-10. Closer look of mucosa in UC showing various types of ulceration and mucosal abnormalities. A: Extensive ulcerations represented by red friable mucosa with intervening pale areas of residual but inflamed mucosa. B: Broad and irregular ulcers with raised intervening areas of involved mucosa. C: Extensive linear tram track-like ulcers with intervening edematous mucosa. D: Irregular and geographic ulcers separating markedly boggy and inflamed mucosa. E: Mucosa with lumpy bumpy inflamed mucosa without deep ulcers. F: Extremely red and friable inflamed mucosa that has completely lost normal mucosal folds that is likely to bleed on touch. |
 Figure 18-11. Regrowth of the epithelium over undermining ulcers can create mucosal bridges such as seen here. |
Microscopic pathology
Variation in Time and Disease Evolution The chronic and intermittent nature of UC with periods of exacerbation and remissions makes it convenient to divide the microscopic appearances into active disease, resolving disease, and disease in remission (quiescent colitis).
ACTIVE COLITIS. In active colitis, the most striking features are that the two major features required for a diagnosis of IBD are present, namely:
a. Crypt architectural distortion in the involved segments including any accepted additional lesions when present (Fig. 18-14) (e.g., cecal patch or periappendiceal inflammation). In addition, Paneth cell, and rarely pyloric metaplasia may be present.
 Figure 18-12. A: UC is characterized by the presence of mucosal ulcerations which extend into the submucosa (arrows). They may show a “collar-button appearance” with undermined edges. B: Another example at higher magnification. |
 Figure 18-13. Benign strictures are rare in UC but can occur as shown here (arrows). The submucosal fibrosis leads to rigidity and luminal narrowing. There is no malignancy or involvement of the muscularis propria. |
b. A diffuse chronic inflammatory cell infiltrate extending to the level of muscularis mucosae, the most reliable feature being the presence of dense plasma cells in the deeper part of the lamina propria (Fig. 18-14). However, around the ileocecal valve, this feature may be present normally in adults.
These two features establish that the mucosa has been previously damaged resulting in architectural distortion, and that long-standing chronic inflammation is present. While these two features can also be seen in chronic infections (e.g., tuberculosis, amebiasis), in Western societies, or where these diseases are rare, IBD is by far the most common cause.
Once this has been established, the next feature is the type of underlying IBD. This is established by the distribution of disease in serial biopsies, ideally
terminal ileum to rectum from the major anatomical sites, (terminal ileum, ascending, transverse, descending, sigmoid colon, and rectum) in separate containers labeled as to site. UC involves the rectum and extends for a variable distance proximally, and this applies both to the architectural distortion and inflammatory infiltrate. Conversely, CD frequently has a focal distribution and often spares the rectum (see subsequent section).
terminal ileum to rectum from the major anatomical sites, (terminal ileum, ascending, transverse, descending, sigmoid colon, and rectum) in separate containers labeled as to site. UC involves the rectum and extends for a variable distance proximally, and this applies both to the architectural distortion and inflammatory infiltrate. Conversely, CD frequently has a focal distribution and often spares the rectum (see subsequent section).
Activity in UC is gauged by the presence of:
c. Neutrophils that tend to infiltrate into the crypt epithelium (cryptitis) and into the lumen (crypt abscesses), with concomitant mucin depletion of the crypts. In UC, while lamina propria neutrophils can often be found, characteristically, they immediately appear to pass into crypt epithelium, often toward the lower or midpart of the crypts. This is in contrast to CD where neutrophils remain in the lamina propria, but if epithelium is infiltrated, it tends to be quite focal, with one part of the biopsy being severely affected, while the remaining can be relatively normal.
Other features that may be present include:
d. Congestion and dilatation of the capillary blood vessels. The vascular changes and mucosal friability account for much of the bleeding tendency experienced during endoscopic examination.
In practice, the diagnosis of active UC in biopsies is highly characteristic, and virtually a low-power diagnosis. Note that in the above description, the words “nonspecific” are not found either as a descriptor of the inflammation (no “nonspecific acute and chronic inflammation”) or in the diagnosis. The features are those of active IBD and highly suggestive of active UC. The features are not “consistent with nonspecific ulcerative colitis.” Indeed, there is nothing “nonspecific” about these features.
The mucosa appears thickened and the surface epithelium may take on an undulating or low villiform appearance (Fig. 18-14A,B). The inflammation may extend into the superficial submucosa, but the muscularis propria and serosa remain free of inflammation except in severe or fulminant colitis (Fig. 18-14B). The presence of inflammation extending into the superficial submucosa on biopsies is not a reliable way of differentiating UC from CD, although higher density of the inflammatory infiltrate in the submucosa is more common with CD.
In more detail, features found are:
Architectural changes: One of the most characteristic features of IBD, which differentiates it from other forms of acute injury to colon is distortion of the mucosal architecture as a result of chronic and persistent inflammatory damage to the epithelium and attempts at regeneration. The damage to the crypts produces abnormal-shaped crypts with branching or budding, and sometimes shortening (Fig. 18-14). The perpendicular orientation of crypts to each other may be lost. The distortion may result from reepithelialization of irregular crypt abscesses. Physiologic crypt division is from the crypt bases up, so they remain straight. Crypt distortion arises following erosions or ulcers when epithelium grows in from the edge of the erosion. New crypts are formed when the surface mucosa dips irregularly into the ulcer to form new crypts, but in a much more irregular manner than usual. This epithelium then develops a lamina propria and a neomuscularis mucosae superficial to the original, so that the muscularis mucosae may appear thickened (duplicated) (Fig. 18-15A,B). Persistent mucosal architectural abnormalities are therefore characteristic of chronic UC, but transient irregularities may be observed in regenerating mucosa from many causes, even previous biopsy sites or a healing ischemic ulcer. Crypt loss may be evident as areas of mucosa devoid of any crypts. Mucosal atrophy (synonym: crypt atrophy) is a combination of crypt dropout and shortening of crypts. While ideally the crypt architecture is best assessed on well-oriented mucosal biopsies, some of the features can be appreciated even when the sections are tangential. While assessing crypt architecture one must also be aware of slight irregularity of crypts that is normally present in the rectum or in the region of the small bowel and colonic junction at the ileocecal valve.
Cryptitis: An early feature of the histology of active UC is the formation of crypt abscesses in the mucosa (Fig. 18-14D). It must be emphasized that these reflect active disease and are not specific to UC, for they occur in a great variety of other intestinal inflammatory conditions including CD and infections. They are, however, particularly conspicuous in active UC.277 The small microabscess thus created expands and either bursts into the lumen of the bowel, elaborating pus into the feces, or spreads into the lamina propria, or, if the inflammation is severe, into the submucosa. It is significant that neutrophils are preponderant within the lumen of the crypts in UC and, unlike infective colitis, comparatively smaller numbers are seen migrating between the epithelial cells. In acute infections, they also tend to be more superficial, but are invariably plentiful in the lamina propria. This can be a helpful feature in the differentiation of UC from infective proctocolitis. Indeed, while neutrophils can be found in the lamina propria without crypt infiltration in UC, it is sufficiently uncommon that other causes, such as CD, or an infectious colitis superimposed on UC, should be considered in the differential diagnosis. Crypt abscesses play an important role in the mechanism of mucosal ulceration and in the formation of
inflammatory polyps. In severe disease, they burst into the loose submucosal tissues; there is a tendency to spread longitudinally beneath the mucosa, which sloughs off leaving an ulcer (Fig. 18-16). The mucosal margins of these ulcerated areas are further undermined and are relatively raised up to form polypoid tags of mucosa projecting into the lumen. These mucosal tags or inflammatory polyps can be short or extremely long and filiform.278
inflammatory polyps. In severe disease, they burst into the loose submucosal tissues; there is a tendency to spread longitudinally beneath the mucosa, which sloughs off leaving an ulcer (Fig. 18-16). The mucosal margins of these ulcerated areas are further undermined and are relatively raised up to form polypoid tags of mucosa projecting into the lumen. These mucosal tags or inflammatory polyps can be short or extremely long and filiform.278
 Figure 18-14. Ulcerative colitis. Low-power views (A,B) showing extensive shortening of crypts and loss of crypts. The surface can take a villiform appearance. In both pictures, mucosal inflammation is diffuse, transmucosal with basal accumulation of inflammatory cells (beneath the crypt base and the muscularis mucosae). Extension of the inflammation into the superficial submucosa as seen in this example (B) is not an evidence against UC. C: Classic biopsy findings in UC showing crypt branching and distorted crypts. The lamina propria shows dense inflammatory infiltrate composed of lymphocytes and plasma cells. Scattered eosinophils are visible even at this magnification. Note that crypts fail to reach the muscularis mucosae creating a subcryptal space occupied by dense plasma cell infiltrate. This basal plasmacytosis can be either focal or diffuse, and is a very useful diagnostic feature. It can however disappear following medical treatment. D: Neutrophils can infiltrate in between crypt epithelial cells (cryptitis) and accumulate in the crypt lumen (crypt abscess) which can lead to crypt destruction. |
 Figure 18-15. Ulcerative colitis. A,B: diffuse thickening of the muscularis mucosae can be very prominent with long-standing disease. It can be associated with an almost normal mucosa (A) or with persisting abnormality of the mucosal architecture in quiescent disease (B). |
 Figure 18-16. Crypt abscesses rupture into the submucosa (arrow) spreading laterally and eventually lead to ulceration as seen on the left side of the figure. |
Epithelial changes: The inflammatory damage to the crypts produces a variety of degenerative and regenerative changes in the crypt epithelium. There is loss of goblet cells resulting in a mucin-depleted appearance of the crypt and surface epithelium. Crypts are usually uniformly lined with many goblet cells, especially in the upper two thirds. Mucin loss may follow any acute mucosal injury, including sometimes even bowel preparation with various bowel cleansing solutions (e.g., fleet’s phophosoda) before colonoscopy. The loss of mucin following purgation or in association with inflammation can be explained by damage and loss of surface epithelial cells, and replacement of these cells by young relatively undifferentiated regenerating cells migrating toward the surface. An increased rate of epithelial cell proliferation is also reflected in the increased numbers of mitotic figures, which correlates with epithelial cell damage and regeneration. It is commonly seen in active disease, but has no diagnostic value. This should not be confused with dysplasia, where mucin depletion may also be seen due to lack of normal differentiation of neoplastic cells. In the presence of attenuated or restituting epithelium anywhere in the superficial epithelium, the changes in the crypts are almost certainly reactive. Some degree of mucin depletion is therefore not unusual in any type of colorectal inflammation although it is less common in CD compared to UC. Severe almost total mucin depletion is a characteristic and fairly consistent finding in active UC and tends to affect all crypts in the biopsy being maximal in the distal colon. The amount of mucus may reappear as disease activity subsides.
Paneth cells metaplasia is also a feature of IBD (Fig. 18-17A).279 Paneth cells, with the exception of the right colon and rarely in the transverse colon, are absent from the rest of the large bowel. These can be numerous in patients with long-standing colitis; however, Paneth cell metaplasia can be seen with other chronic inflammatory conditions of the bowel and are diagnostically not specific and unhelpful (Fig. 18-17D). Increased numbers of endocrine cells may also be seen in the base of the crypts.261 Pyloric gland metaplasia may also occur, but it is distinctly uncommon in UC compared to CD, especially in the small bowel (Fig. 18-17B,C).
Mucin granulomas (crypt rupture or cryptolytic granulomas): A particular problem in the differential diagnosis of IBD is the giant cell reaction and poorly formed granulomas in relation to crypt damage and liberated mucin (Fig. 18-18). These granulomas usually contain mucin and/or neutrophils, histiocytes, and occasionally giant cells.258 However, sometimes the granuloma may also be very well formed and sarcoid like. The mucin is not always readily visible, and often the best clue is that their location is precisely in the location in which a crypt should have been, even given the irregularity that occurs in UC. Further clue is that crypt remnants may be visible although it may take levels, mucin stain and/or epithelial immunostain to demonstrate them. Such granulomas may be seen in other forms of colitis, including CD, infectious colitis, and colitis associated with diverticular disease,280 as well as in ileal pouches.266, 267, 281, 282 Care needs to be taken not to say that these are “consistent with CD” on this basis alone, for once that diagnosis is suggested, it has a habit of sticking, and the patient is labeled as having “histologically proven CD.” At some point, this may deprive the patient of a pouch. In diverted UC, several types of granuloma may be seen at all sites in the bowel wall and in draining lymph nodes.283
Lesions of the enteric nervous system are uncommon in UC and may be a good predictor of CD.284 They are however difficult to assess on endoscopic biopsies.285 Nevertheless in CD, involvement of the myenteric plexus by inflammation predicts earlier recurrence (see subsequent section).
Chronic inflammation: In active colitis, accompanying the changes in the crypts and surface epithelium, there is a heavy diffuse infiltrate of inflammatory cells in the lamina propria. These include neutrophils, lymphocytes, plasma cells, eosinophils, and mast cells. Studies of immunoglobulin-containing
plasma cells in UC demonstrate an increase in the major forms of IgA, IgG, and IgM.286 The increase correlates with disease activity and the rise of IgG- and IgM-containing cells is proportionally more than for IgA-containing cells.287
plasma cells in UC demonstrate an increase in the major forms of IgA, IgG, and IgM.286 The increase correlates with disease activity and the rise of IgG- and IgM-containing cells is proportionally more than for IgA-containing cells.287
 Figure 18-17. Metaplasia in IBD. A: Paneth cell metaplasia is the most frequent and can range from an occasional cell to many in each crypt. Endocrine cells may also be increased in the crypts and need to be differentiated from Paneth cells. The endocrine cells have the cytoplasmic granules toward the basement membrane and nucleus toward the lumen (arrow), while in Paneth cells the nucleus rests on the basement membrane and the granules are toward the crypt lumen. B,C: The appearance of the ulcer-associated cell lineage (UACL) or pyloric metaplasia is rare in the large bowel in UC and CD and more common in the ileum in CD. It can appear in the deeper part of the mucosa (arrow; B = low power; C = high power) but also on the surface (D), where the cells take the form of mucin secreting cells with a small amount of apical mucin (arrow). The phenomenon indicates the occurrence of an ulcer previously and is not specific for IBD. E: Increased number of endocrine cells in the crypts (arrows). |
 Figure 18-18. Crypt rupture may result in mucin granulomas that often have many neutrophils, few eosinophils, and loose aggregates of histiocytes and should not be confused with granulomas of CD (A). Occasionally, these can be more compact and have more histiocytes and even giant cells (B). |
Eosinophils and Mast cells: In some biopsies from UC, large numbers of eosinophils may be seen in the lamina propria, and may completely dominate the inflammatory infiltrate. This infiltrate has been the subject of much study in attempts to correlate it with clinical outcome, but there have been no consistent results.287 It likely reflects the extreme end of the Th2 response that is said to characterize UC. Further, in developed countries quiescent UC is by far the most common cause of “eosinophilic colitis.” This is further discussed in the section on eosinophilic colitis. The same conclusion has to be made about mast cell numbers,287, 288, 289 although difficulties in techniques for demonstrating degranulated mast cells complicate the issue if histochemistry rather than immunohistochemistry is used. Basal lymphoid follicles/aggregates: In long-standing disease, hyperplasia of mucosal basal lymphoid follicles becomes a prominent feature, particularly in the rectum (follicular proctitis), but in resections, transmural inflammation in the form of lymphoid aggregates and hyperplasia, characteristic of CD, is never seen in UC, although it is a feature of the diverted rectum in UC.283 Vasculitic lesions of polyarteritis type in submucosal vessels are only seen rarely in UC.290 Inflammation within blood vessels may be seen close to ulcerated areas, but this is usually a secondary feature. Granulomatous vasculitis, which is sometimes seen with CD,291 has only been described in UC in association with diversion.292 Assessment of disease activity in IBD and “Deep Remission”: Assessing disease activity is an evolving area with wide variations in current practice. The primary goal of treatment in IBD has always been control of symptoms. However, with newer therapies, endoscopic remission has also become an additional end point of therapy, and histologic remission is gaining importance, especially in UC. The combination of all three parameters of remission is suggested to be “deep remission” and is the ideal aim of therapy. Although most patients do show overall histologic improvement following clinical remission, histologic correlation with symptoms and endoscopic findings is modest.293 Some patients in clinical remission continue to show dense lamina propria lymphoplasmacytosis and sometimes persistent acute inflammation, while others continue to have symptoms despite a decrease in inflammation. In some patients, persistent symptoms relate to concomitant IBS or functional, noninflammatory-related complaints.
Studies suggest that active inflammation (neutrophilic infiltrate) not only correlates with clinical disease and endoscopic severity but, along with basal plasmacytosis, predicts disease relapse as well.293 Thus, patients with basal plasmacytic infiltrates (as in Fig. 18-14B,C) are more likely to relapse than those without it (as in Fig. 18-19). Indeed, if chronic inflammation can be kept within normal limits, relapse is unlikely. The real purpose of therapy is therefore not just to get the patients into remission but to keep them there. This implies that it would be wise to continue maintenance therapy (e.g., mesalamine) in patients who fail to achieve “histologic remission.” Conversely if the patient is being treated with steroids, one might be reticent to prolong or increase therapy simply to produce histologic remission (i.e., a normal lamina propria), and the notion of upgrading therapy, for example, by adding steroids, cannot currently be justified. Well-designed clinical trials in appropriate patient populations are required for this to become
the standard of practice and care. However, the paradigm shift toward getting a patient into a remission that is clinical, endoscopic, and histologic (no acute inflammation and a normal chronic inflammatory infiltrate in the lamina propria) is increasingly being considered as ideal.293
the standard of practice and care. However, the paradigm shift toward getting a patient into a remission that is clinical, endoscopic, and histologic (no acute inflammation and a normal chronic inflammatory infiltrate in the lamina propria) is increasingly being considered as ideal.293
Histologic grading and reporting of UC/IBD biopsies: The issue then becomes how one should report these biopsies, what features should be mentioned (active vs. chronic changes), and what grading system to use. Most pathologists tend to grade the acute (active) inflammation subjectively into none (quiescent), mild, moderate, and severely active ulcerative colitis, with or without specific mention of the chronic inflammation. While this system is simple and practical, there are no specific reference criteria, so this is entirely subjective. Although more objective histologic scoring systems of disease activity in IBD have been developed, their utility has been largely limited to clinical trials and research. Truelove, in Oxford, devised the first recorded method relating symptoms to the amount of neutrophilic infiltrate and epithelial damage. Subsequently, the Riley system, which is similarly based, was designed and showed that a mild activity score (occasional small groups of neutrophils in the lamina propria) correlated with early relapse after cessation of treatment.293 The Geboes system is easy to use and has been tested for its reproducibility293 (Table 18-6).
However, for clinical practice, these systems are a bit complex and difficult to apply. Many pathologists simply report these biopsies as “mild, moderate, or severe ulcerative colitis”, with no reference to chronic features with the argument that this is already implicit in the diagnosis of ulcerative colitis. There is certainly merit in grading acute/active inflammation in symptomatic or endoscopically active patients, as these provide a good indication of disease severity. However, as the concept of “deep remission” becomes increasingly important, grading of chronic inflammatory features in addition, especially if basal plasmacytosis is present, also becomes important, especially in patients with few or no symptoms and lesser endoscopic activity, as it predicts an increased risk of relapse.
A simplified but comprehensive grading system is shown in Table 18-7A, in which both the acute and chronic components are graded separately. The system can be used for any form of IBD, which is helpful as often the nature of underlying chronic colitis remains unclear for a variety of reasons. Ulcerative colitis can be reported as, for example, “moderate chronic ulcerative colitis, with mild activity” or simply “mildly active moderate chronic ulcerative colitis,” or “mild chronic active ulcerative colitis.” One could replace ulcerative colitis with simply “chronic colitis,” and the system can be used for any form of IBD. The system is particularly useful as it can be used at the time of initial diagnosis as well as during any follow-up biopsies and can be used for any form of IBD. If there is no active or chronic inflammation but yet there are changes of architectural distortion or Paneth cell metaplasia, this can be reported as “quiescent
colitis.” Features like architectural distortion and Paneth cell/pyloric metaplasia have poor correlation with clinical disease, and their only value in reporting is to establish the presence of the underlying disease. The absence of any chronic changes or presence of minimal changes could imply complete histologic remission or even raise a doubt regarding the diagnosis of IBD. Sometimes during a follow-up biopsy, the clinicians may want to know if in fact ulcerative colitis is truly the correct underlying diagnosis, so “features of active colitis are absent, but crypt architecture changes as seen in inactive/quiescent ulcerative colitis” can be helpful. Alternatively, a sign out of “there are no features of active colitis or chronic changes” or that that the “biopsy is normal” can also be helpful. However, it must always be remembered that architecture can return completely to normal, so the absence of architectural changes never excludes the diagnosis of ulcerative colitis. Thus, “there is no evidence of IBD in these biopsies” works better than “normal biopsies” for that reason. One should recognize that the importance of routine, repeated biopsy examination in IBD is not just to assess disease activity, but sometimes to confirm the diagnosis or to raise the question of other forms of IBD, to exclude concurrent infection (especially those in which organisms may be visible), and to assess for the presence of dysplasia or carcinoma. Simple grading system for ulcerative colitis: This is based on the assumption that there is virtually never any acute inflammation (at least caused by IBD) without chronic inflammation also being present, so the paradigm of “no chronic inflammation—no acute inflammation” holds in IBD. No acute inflammation obviously implies that there is no cryptitis, crypt abscesses, erosions, or ulcers. Then logically if there is acute inflammation, it can be graded subjectively, but that in the presence of acute inflammation, grading chronic inflammation is unnecessary as it is always present to some degree, and is of no clinical significance as the presence of any acute inflammation implies an increased risk of relapse.
colitis.” Features like architectural distortion and Paneth cell/pyloric metaplasia have poor correlation with clinical disease, and their only value in reporting is to establish the presence of the underlying disease. The absence of any chronic changes or presence of minimal changes could imply complete histologic remission or even raise a doubt regarding the diagnosis of IBD. Sometimes during a follow-up biopsy, the clinicians may want to know if in fact ulcerative colitis is truly the correct underlying diagnosis, so “features of active colitis are absent, but crypt architecture changes as seen in inactive/quiescent ulcerative colitis” can be helpful. Alternatively, a sign out of “there are no features of active colitis or chronic changes” or that that the “biopsy is normal” can also be helpful. However, it must always be remembered that architecture can return completely to normal, so the absence of architectural changes never excludes the diagnosis of ulcerative colitis. Thus, “there is no evidence of IBD in these biopsies” works better than “normal biopsies” for that reason. One should recognize that the importance of routine, repeated biopsy examination in IBD is not just to assess disease activity, but sometimes to confirm the diagnosis or to raise the question of other forms of IBD, to exclude concurrent infection (especially those in which organisms may be visible), and to assess for the presence of dysplasia or carcinoma. Simple grading system for ulcerative colitis: This is based on the assumption that there is virtually never any acute inflammation (at least caused by IBD) without chronic inflammation also being present, so the paradigm of “no chronic inflammation—no acute inflammation” holds in IBD. No acute inflammation obviously implies that there is no cryptitis, crypt abscesses, erosions, or ulcers. Then logically if there is acute inflammation, it can be graded subjectively, but that in the presence of acute inflammation, grading chronic inflammation is unnecessary as it is always present to some degree, and is of no clinical significance as the presence of any acute inflammation implies an increased risk of relapse.
Table 18-6 Scoring System for the Assessment of Severity in Ulcerative Colitis | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Table 18-7A Simplified Grading System for Chronicity and Activity in IBD (Option A)* | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||
Table 18-7B Simplified Grading System for Ulcerative Colitis (Option B) | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||
If there is no acute inflammation, it is only worth grading chronic inflammation as normal (rarely decreased post therapy) or increased and if increased with or without basal plasmacytosis, the latter predisposing to relapse in patients with quiescent disease.293
A system for both grading biopsies and predicting relapse based on both the Geboes and Bessissow criteria is shown in Table 18-7B. However, one needs to be careful while dealing with biopsies from the right colon (cecum, ascending colon) that can have basal plasma cells normally, especially in adults. Biopsies from this region should therefore always be considered normal if based on chronic inflammation alone. As this tends to be an age-associated change, basal plasma cells may be more important in children in this location.
A system for both grading biopsies and predicting relapse based on both the Geboes and Bessissow criteria is shown in Table 18-7B. However, one needs to be careful while dealing with biopsies from the right colon (cecum, ascending colon) that can have basal plasma cells normally, especially in adults. Biopsies from this region should therefore always be considered normal if based on chronic inflammation alone. As this tends to be an age-associated change, basal plasma cells may be more important in children in this location.
Various examples of biopsies with UC Biopsies from the cecum and ascending colon—colonic mucosa without abnormality Biopsies of large bowel from transverse colon—architectural distortion as seen in quiescent ulcerative colitis Biopsies of the descending colon—ulcerative colitis with an excess of chronic inflammatory cells but without basal plasmacytosis Biopsies from the sigmoid colon—chronic ulcerative colitis with moderate basal plasmacytosis* Biopsies from the rectum with mildly active ulcerative colitis* |
* In patients in clinical remission, the presence of these features has been reported to increase the likelihood of clinical relapse. |
Bessissow T, et al. Prognostic value of serologic and histologic markers on clinical relapse in ulcerative colitis patients with mucosal healing. Am J Gastroenterol. 2012;107(11):1684-1692. |
This is far from perfect and does not answer questions such as how many basal plasma cells are required (more than the odd one), or how many neutrophils are required to change from one grade to another, or the significance of numerous eosinophils. However, it does provide a relatively easy-to-use system that does have clinical relevance and is semiquantitative.
RESOLVING COLITIS. The relapses and remissions of UC imply that inflammation may resolve spontaneously. Evidence that this is the case comes from numerous placebo-controlled studies when up to about a third of patients in any placebo group go into remission spontaneously. Hence in clinical trials, drugs need to demonstrate a correspondingly higher response rate to be considered beneficial. Furthermore, resolution can occur at different rates in different anatomical areas of the colon. This can give a false impression of segmental disease, not only macroscopically but also microscopically.270, 272, 295 With resolution of disease, the numbers of inflammatory cells of all types begin to diminish and their distribution becomes uneven. The goblet cell population returns toward normal and, depending on the severity of the attack, the crypt architectural changes may persist. Some crypts may appear short; others branched, the changes invariably being more marked distally.
 Figure 18-19. During quiescent phase, the lamina propria inflammation subsides and the only evidence, of prior colitis may be mild crypt distortion or Paneth cell metaplasia (A). In some instances the mucosa returns to normal architecture and only evidence of prior disease may be thickened or duplicated muscularis mucosae (B). |
QUIESCENT COLITIS. Varying degrees of crypt atrophy and distortion are the hallmarks of quiescent disease. Shortfall of the crypt with regard to the muscularis mucosae and crypt loss are convenient ways of assessing this atrophy when present, and the lamina propria may show an unusual “empty” appearance (Fig. 18-19). On the other hand, there may be lymphoid hyperplasia (follicular proctitis) in the rectum (Fig. 18-20). Crypts are normally present in the large bowel mucosa at a frequency of six per millimeter in a biopsy, which includes muscularis mucosae. In a biopsy without muscularis mucosae, the mucosa may be stretched out and appear atrophic when it is not. Edema may also mimic crypt atrophy due to separation of crypts; this may follow administration of bowel preparation fluids, especially older hyperosmolar solutions.296 Inflammatory polyps can be seen in a background of quiescent colitis. Sometimes, thickening of the muscularis mucosae is also seen, and may
be the only evidence of past inflammation or ulceration (Fig. 18-19B).297
be the only evidence of past inflammation or ulceration (Fig. 18-19B).297
 Figure 18-20. Follicular proctitis. A: Resected specimen of the rectum with a reddened granular mucosa that is otherwise indistinguishable from that of other resections for UC. B,C: Biopsies from another example that show marked lymphoid hyperplasia with germinal centres in some of them. The mucosa in between also shows a dense lymphoplasmacytic infiltrate; the crypts appear somewhat reduced in numbers with mild distortion of architecture. D: Higher magnification shows intervening areas with cryptitis and crypt abscess formation. |
While most patients have some residual changes of previous damage (crypt distortion, atrophy, Paneth cell metaplasia), it has become increasingly recognized that a group of UC patients may show complete resolution with no evidence of previous disease.271, 272, 295 This always raises the question as to whether the original diagnosis was really UC or an infective (self-limited) colitis and a careful review of all clinical data and histologic material may be warranted. After such review, there remains a patient group with genuine UC, with collateral evidence for such a diagnosis, including coexistent PSC, who undoubtedly show evidence of complete mucosal recovery, although the colitis of PSC may tend to be more right-sided than usually seen in UC.298
The condition of the patient group with histologic evidence of UC with crypt architectural distortion but with normal radiology and colonoscopy has been termed “minimal change colitis.”12 This is not a great term, and “quiescent ulcerative colitis” is much less ambiguous. It should not be confused with microscopic colitis, since there is crypt architectural distortion, which is not a feature of microscopic colitis. Minimal change colitis may be seen after treatment or at presentation, and indeed may never develop endoscopically recognizable changes, although some authors claim that minor vascular abnormalities may be seen at colonoscopy. It is not clear why some patients show quite florid atrophy, yet in others the mucosa returns to a state close to normality.
POLYPS IN ULCERATIVE COLITIS. Polyps in UC are common, being present in 12% to 20% of cases and more commonly associated with bouts of previous severe disease.299, 300 These represent both mucosal excrescences
and sometimes reepithelialized granulation tissue. These inflammatory polyps or mucosal tags may be present in large numbers and adopt bizarre shapes (Fig. 18-21A-D). The pathogenesis involves previous severe ulceration: islands of preserved inflamed and edematous mucosa that eventually take the shape of polyps and over time reepithelialization of the intervening ulcers (Fig. 18-22A-D).278, 301 The term “pseudopolyp” has been used to describe these postinflammatory polyps, but this appears to be a poor term. First, there is certainly nothing “pseudo” about these polyps, while “inflammatory polyp” is a better expression as it indicates the way in which the polyps are formed, and it satisfies the definition of polyp handsomely. However, they have to be distinguished from residual mucosal islands—which may be genuine “pseudopolyps.” These are the result of localized ulceration of the mucosa and usually submucosa, with undermining of adjacent intact mucosa. The surviving mucosa appears raised, but only relative to the adjacent ulcerated mucosa, so that it appears to form a polyp, which is really a mucosal island, and therefore appears to project into the lumen. The term mucosal island is better than pseudopolyp, removing the ambiguity inherent in that term.
and sometimes reepithelialized granulation tissue. These inflammatory polyps or mucosal tags may be present in large numbers and adopt bizarre shapes (Fig. 18-21A-D). The pathogenesis involves previous severe ulceration: islands of preserved inflamed and edematous mucosa that eventually take the shape of polyps and over time reepithelialization of the intervening ulcers (Fig. 18-22A-D).278, 301 The term “pseudopolyp” has been used to describe these postinflammatory polyps, but this appears to be a poor term. First, there is certainly nothing “pseudo” about these polyps, while “inflammatory polyp” is a better expression as it indicates the way in which the polyps are formed, and it satisfies the definition of polyp handsomely. However, they have to be distinguished from residual mucosal islands—which may be genuine “pseudopolyps.” These are the result of localized ulceration of the mucosa and usually submucosa, with undermining of adjacent intact mucosa. The surviving mucosa appears raised, but only relative to the adjacent ulcerated mucosa, so that it appears to form a polyp, which is really a mucosal island, and therefore appears to project into the lumen. The term mucosal island is better than pseudopolyp, removing the ambiguity inherent in that term.
 Figure 18-21. Inflammatory polyps in ulcerative colitis. A: Numerous varying sized polyps many of which are pedunculated and the gross appearance can closely mimic adenomatous polyposis syndrome. B: Another example with a large pedunculated solitary polyp. C: Carpet of numerous small polyps that sharply separate the involved segment from the normal mucosa in this example. D: Colitis with myriad inflammatory polyps. Note the impossibility of detecting neoplasms within such forests. |
Inflammatory polyps do not correlate with the disease activity, but indicate severe disease, usually in the past. The polyps can be found anywhere in the colon, or elsewhere in the bowel were there has been ulceration. Inflammatory polyps can adhere to one another to form mucosal bridges across the lumen (Fig. 18-11). If healing takes place, forests of polyps may remain as evidence of past disease, so-called colitis polyposa (Figs. 18-21A,C,D and 18-22D). The polyposis of UC tends to be more prominent in the colon than the rectum and may be seen proximal to the area of active disease. Even when few or isolated, polyps are less frequent in the rectum, and more common in the sigmoid and descending colon. Inflammatory polyps can occur in both major types of IBD, but are
much commoner in UC.301 They can also be found guarding the entrances and exits (and within) fistula tracts, so are sometimes called sentinel polyps. While confusion can occur grossly with adenomatous polyposis syndromes on gross examination, the appropriate histologic diagnosis should be sought by examining the background mucosa.
much commoner in UC.301 They can also be found guarding the entrances and exits (and within) fistula tracts, so are sometimes called sentinel polyps. While confusion can occur grossly with adenomatous polyposis syndromes on gross examination, the appropriate histologic diagnosis should be sought by examining the background mucosa.
 Figure 18-22. A: Early stages of inflammatory polyp formation showing extensive ulceration resulting in islands of residual inflamed and edematous mucosa that take the shape of a polyp. B: With time, the epithelium grows over ulcerated surface between multiple islands of polypoid mucosa. C: Some of the polyps can grow large with a mushroom shape, and can mimic adenomatous polyps grossly or on low power. D: Once the epithelium has completely grown over the ulcers and numerous such polyps are formed, the entire mucosa is covered with such polyps taking a lumpy bumpy shape (colitis polyposa). |
The distribution of polyps depends on the extent of the primary disorder. They are typically small (measuring <1.5 cm in height), often multiple and may resemble adenomatous polyps grossly. Inflammatory polyps (especially large ones) may produce acute obstruction or intussusception, or may mimic carcinomas. Arborization is a striking feature in some. Feces tend to become trapped within the maze of mucosal fronds and this can cause local inflammation.
Rarely, the inflammatory polyps may have a filiform configuration characterized by thin and long polyps that resemble worms or spaghetti (Fig. 18-23A,B).302, 303 Filiform polyps may be as long as 2 to 3 cm. The filiform polyposis is often segmental or focal, and frequently spares the rectum. It may sometimes completely fill the colonic lumen producing obstructive symptoms. Ulceration is conspicuous by its absence on macroscopic inspection and adjacent flat mucosa can appear quite normal grossly. Filiform polyposis, usually coexists with chronic IBD, although some patients apparently have no evidence of IBD.287, 288, 304, 305 However,
unless these are a form of hamartomatous polyposis, they must have been preceded by severe ulcerative inflammation, even if its etiology is obscure.
unless these are a form of hamartomatous polyposis, they must have been preceded by severe ulcerative inflammation, even if its etiology is obscure.
 Figure 18-23. Inflammatory polyps in ulcerative colitis. A: Rarely, the polyps can have a finger-like or worm-like shape and referred to as filiform polyps. B: Microscopy show virtually entire mucosa is converted into filiform polypoid mucosa. |
Benign inflammatory polyps rarely become dysplastic, but this does occur. The main problem with these polyps when numerous is that it is virtually impossible to carry out surveillance so that their presence inevitably leads to serious consideration of colectomy. Adenomas can also occur in UC as in the rest of the population, as do IBD-associated dysplasia and are discussed later in the section on dysplasia in IBD.
Focality/patchiness in ulcerative colitis. UC is classically diffuse in its distribution. However, not all the diseased bowel need be in a similar state of activity and this can create a false impression of segmental involvement. Activity is usually maximal in the rectum unless the patient is receiving local therapy, and tends to decrease proximally, although occasionally the colitis associated with primary sclerosing cholangitis may have an appearance of right sided ulcerative colitis (see subsequently). An isolated patch of inflamed mucosa can be seen in the cecum (cecal patch—Fig. 18-24A) or around the appendiceal ostia (periappendiceal patch—Fig. 18-24C) as a discontinuous lesion in patients with left-sided or distal UC (Fig. 18-24A,B).306, 307, 308, 309, 310, 311, 312 The histologic changes in these discontinuous cecal or periappendiceal patches are no different from typical IBD. Periappendiceal or cecal patches should not occur in patients with pancolitis, being part of the background disease, but while this can occur they are often found as isolated changes in patients with distal colitis—usually but nor always ulcerative colitis. The reported incidence of appendiceal involvement (21%-86%) and cecal patch (10%-75%) is highly variable, and occasionally, similar
skip lesions in the ascending colon have also been described in a small subset (4%). One study found that periappendiceal disease had a male predominance, and none had a prior appendectomy.313 The other situation of segmental or patchy inflammation is presentation as diverticular-associated colitis, described subsequently.
skip lesions in the ascending colon have also been described in a small subset (4%). One study found that periappendiceal disease had a male predominance, and none had a prior appendectomy.313 The other situation of segmental or patchy inflammation is presentation as diverticular-associated colitis, described subsequently.
 Figure 18-24. A: Subtotal colectomy in a patient with UC showing a distal colitis and inflammation in the cecum (cecal patch) (arrow) with mucosal erythema and thickened folds which extend to the appendiceal orifice as well. B: Mucosal inflammation around appendiceal orifice in a patient with left sided ulcerative colitis. |
Focality of inflammatory activity can also be induced by treatment, especially local steroid enema in the rectum or occasionally following spontaneous remission295 (see later). During remission the mucosa may become normal again. Healing often occurs in an irregular discontinuous way, but often occurs first in the rectum.314 Treatment may increase the heterogeneous nature of the lesions. Topical treatment of the rectum with steroid or 5-ASA enemas may induce complete rectal healing while the lesions in the other segments of the colon may persist.270
Appendiceal involvement. The role of the appendix in the possible etiology of UC is discussed previously. The appendix is involved in about 75% of total colectomy specimens performed for UC. This may be continuous or as a skip lesion.309, 310, 311, 312 Such a “skip lesion,” although apparently suggestive of CD, should not be regarded as a contraindication to pouch surgery. The appendicitis may extend into the contiguous large bowel as a periappendiceal patch, and does not lead to acute suppurative appendicitis because the inflammation remains confined to the mucosa.307 When symptoms occur, they are more likely to be the central abdominal pain, and very likely to be interpreted as the mild pain associated with the active colitis rather than due to acute appendicitis. The pathologic features are identical to the mucosal changes seen in the rest of the colon (Fig. 18-25A,B). The practical question is what should one do when one sees changes suggestive of UC in an appendectomy specimen in a patient with no preceding history of IBD. It is recognized that some patients with IBD (more often CD) will initially present with appendicitis and some will develop it years later. Most often, the changes of IBD are either masked by features of acute appendicitis or absent, while on occasions, the mucosal changes overwhelmingly mimic IBD with significant crypt distortion and minimal transmural acute inflammation. Our approach is to note these changes in the report and ask the clinician to exclude any concomitant evidence of IBD by colonoscopy and keep an eye on for future development of IBD.
 Figure 18-25. A: Involvement of the appendix with UC at low magnification showing the inflammation is largely mucosal with crypt architectural distortion. B: Higher magnification showing mucosal changes that are identical to those seen elsewhere in the colon. |
Rectal sparing. UC with an endoscopically and histologically normal rectum probably occurs only rarely.315 Occasionally, the rectal mucosa may look endoscopically normal, but show inflammatory changes on histology, hence the importance of biopsying the rectum during the initial evaluation of IBD cannot be overemphasized. Some of the older literature on rectal sparing in UC was entirely based on endoscopic evaluation only, and hence cannot be relied upon. The rectal sparing can be produced by healing in response to local steroid enemas, sometimes leading to endoscopic, but not histologic healing. In most patients architectural distortion is present, but, as discussed previously, this can also completely resolve. Even when the mucosal histology returns to normal, thickening and duplication of the muscularis mucosae and a degree of submucosal fibrosis are invariably present, but these can only be evaluated properly on resection specimens.316 It is best termed relative rectal sparing rather than true rectal sparing. On occasion, relative rectal sparing may occur in some patients with UC in the absence of rectal instillation of anti-inflammatory drugs. However, in two controlled studies of UC with apparently normal
rectal mucosa, associated with PSC, while one study showed a distinct right-sided tendency of the persistent inflammation,298 the other failed to show any difference from controls.317 There is no way of knowing whether the rectum was ever involved in these patients or this represented spontaneous resolution of inflammation.318 Relative rectal sparing has also been documented in children more frequently (see later).
rectal mucosa, associated with PSC, while one study showed a distinct right-sided tendency of the persistent inflammation,298 the other failed to show any difference from controls.317 There is no way of knowing whether the rectum was ever involved in these patients or this represented spontaneous resolution of inflammation.318 Relative rectal sparing has also been documented in children more frequently (see later).
Ileal involvement in ulcerative colitis. When the terminal ileum is involved, the mucosal changes are similar to those seen in the colon, and are always in continuity with disease in the large bowel, being associated with an open dilated and incompetent ileocecal valve. Although the expression “backwash ileitis” is not necessarily accurate, because there is, as yet, no evidence that ileal disease is the result of such a mechanism, it is in common usage. However, the term is clinically useful as it differentiates ileitis associated with UC from other causes. Ileitis is found in about 17% of colectomy specimens for UC, the extent of involvement varying from 5 to 25 cm, but rarely is considerably longer (up to 50 cm; see also Fig. 18-29A), although in most resections, the true length is difficult to evaluate as routinely very small cuff of terminal ileum is removed close to the ileocecal valve.319, 320, 321 The histologic changes are generally mild with mild crypt architectural distortion, mild villous blunting, and increase in lamina propria inflammatory infiltrate (Fig. 18-26A,B). Superficial erosions, cryptitis, and crypt abscess can also be seen. Although pseudopyloric metaplasia can be occasionally seen, the presence of granulomas or transmural inflammation should raise a concern for CD. While most of the cases are associated with pan-colitis, in one study, few patients had only extensive or left-sided colitis. The degree of inflammation in the ileum generally parallels the degree of inflammation in the cecum. In biopsies, the key is that the disease has to be at least as severe in the biopsy/biopsies immediately distal to the ileo-cecal valve. If that is not the case, then it is reasonable to exclude backwash ileitis and consider the other major causes of ileal inflammation, NSAIDs/ASA or CD. It is unclear if a cecal patch can have an associated
ileitis, but there are no good data on which to make that assessment. However we have seen ileal biopsies from patients with ulcerative colitis with occasional foci of acute inflammation in the lamina propria, but are unsure of its significance.
ileitis, but there are no good data on which to make that assessment. However we have seen ileal biopsies from patients with ulcerative colitis with occasional foci of acute inflammation in the lamina propria, but are unsure of its significance.
 Figure 18-26. A: Ileal lesions associated with UC (so-called backwash ileitis) often show architectural distortion, regular shortening of the villi, cuboidal surface epithelial cells, and increased lamina propria inflammation. B: Another example of backwash ileitis with occasional branching crypt. The inflammation is mild. C,D: Backwash ileitis showing crypt and villous architectural distortion and areas of cryptitis. The inflammation is limited to the mucosa. A small focus of pseudopyloric metaplasia can also be seen (arrow). In this case the terminal 4 cm of the ileum was involved. |
Pan enteritis post colectomy for ulcerative colitis. This is a rare condition in which patients who have either undergone subtotal colectomy for UC, or had an ileo-anal pouch formed in addition, present with severe inflammation that involves the entirety of the small bowel (Fig. ), pouch and can also involve the stomach (it may not be in continuity). The presence of an ileal pouch seems often to be key. The features present in the initial resection are characteristic of UC. The subsequent presentation of panenteritis with histologic features of IBD, coupled with the clinical unresponsiveness to antibiotic treatment and histologic absence of CMV, suggests the diagnosis of UC-associated enteritis.
In the small bowel, the more common histologic finding is diffuse superficial chronic active enteritis, essentially similar to UC in the large bowel. Indeed the initial impression is of a flat mucosa with all of the features of UC except that the biopsies are from the small bowel. In addition, Lin et al. noted that of 4 cases of diffuse duodenitis occurring in UC, all occurred in patients with pouchitis.321a Valdez et al. also noted the presence of duodenitis in UC: 3 of 4 had duodenitis following a pouch procedure, and 1 had duodenitis preceding colectomy.321b Similarly, Wang et al. noted that while duodenal inflammation could occur in UC in the absence of pouchitis, crypt architectural changes only occurred in patients with pouchitis.321c This association of duodenitis with pouchitis may represent proximal and distal sampling, respectively, of a diffuse panenteritis following colectomy. More recently, small intestinal lesions consisting of ulcers and edema were observed by capsule endoscopy in patients with UC following proctocolectomy, though interestingly these features were also observed in active UC prior to surgery.321d Treatment is that for IBD with high-doses of corticosteroids with or without azathioprine, and once in remission patients don’t seem to relapse.321e
Prestomal ileitis. This can occur following colectomy and ileostomy, and is probably a complication of the ileostomy operation.322 In this rare complication, ulcers are scattered throughout the ileum and jejunum. The intervening mucosa may be abnormal and in continuity with the ileostomy, when possible ischemia may be considered. When there is normal intervening mucosa the possibility of CD inevitably gets raised, but there are no good data regarding this. If the lesions are ever resected then usual criteria for possible Crohn’s, and other diseases can be applied. The ulcers can perforate, resulting in peritonitis and fecal fistulae. Inflammation of the ileum also occurs in pelvic ileal reservoirs (pouchitis) with an adaptive change in the mucosa (“colonic metaplasia”) to produce a picture similar to the original colitis, and in the ileum proximal to the pouch (prepouch ileitis). Prepouch ileitis is uncommon; however, it may extend for long distance (upto 50 cm) in some patients.319 In one study of 571 patients of UC who underwent restorative proctocolectomy, 19 (3%) developed prepouch ileitis, of which 3 were reclassified as CD.319 Only 7(47%) of these had associated pouchitis and only 2 had preoperative backwash ileitis. The histologic changes tend to be similar to pouchitis.319 These changes are discussed further in the section on pouchitis.
Ulcerative colitis in children. Colonic mucosal biopsies from children presenting with new-onset UC can show significantly less histologic abnormalities across all pathologic parameters routinely used in the evaluation of chronic colitis. Rectal sparing has been well documented at the initial onset.315 In samples from children between 1 and 10 years of age, significantly less crypt branching, plasma cells in the lamina propria, cryptitis, crypt abscesses, and epithelial injury are present when compared to adults (Fig. 18-27A,B). Focal colitis and/or absence of chronic crypt abnormalities are observed in initial rectosigmoid specimens in approximately 33% of patients. In 4% to 8% of cases the initial biopsy samples are completely normal. In children older than 10 years, the histologic changes tend to be similar to adults.323 These differences can probably be explained by the duration of the disease, which is likely to be longer in older children.
Proctitis/ulcerative proctitis. Localized inflammation of the rectal mucosa can be caused by a great variety of circumstances such as mucosal prolapse, trauma, suppositories, radiation, antibiotic therapy, piles, persistent diarrhea, CD, UC, and, in more recent years, increasing numbers of infectious agents, the latter having particular importance in the homosexual population, as well as in women engaging in anal sex (Table 18-8). The old viewpoint was that when these causes have been excluded, one is left with an idiopathic pattern of distal disease324 for which there were many synonyms, including proctosigmoiditis, idiopathic proctitis, nonspecific proctitis, lymphoid follicular proctitis, and ulcerative proctitis. This implied that “nonspecific ulcerative proctitis” is a disease of exclusion. Up to a point that is correct if only clinical features are used and infections are excluded. However, on biopsies, the changes of UC are really quite characteristic and, like its more proximal
counterpart ulcerative proctocolitis, the changes are usually readily discernable on biopsy. The “nonspecific” part of this designation is therefore an historical title borne out of ignorance when all biopsies were called “nonspecific chronic active colitis/proctitis” (regrettably still all too common). There was no appreciation that the chronic infections of this region whether venereal, tuberculous, or even amebic could be diagnosed and treated. Now we recognize that a vast majority of chronic proctitis encountered in clinical practice in Westernized societies represents IBD (UC/proctitis) and crypt architectural distortion, deep lamina propria plasma cells, cryptitis, and crypt abscesses are characteristic histologic changes. The other main form of IBD, CD, rarely presents as proctitis alone. However, it should also be recognized that the changes are not entirely specific, and as listed in the table other causes need to be considered in the differential diagnosis. Some of the chronic infections at times very closely mimic IBD and require a detailed clinicopathologic workup including demonstration of the pathogen in the tissues. Rarely, the changes of ulcerative proctitis can be those of a follicular proctitis.
counterpart ulcerative proctocolitis, the changes are usually readily discernable on biopsy. The “nonspecific” part of this designation is therefore an historical title borne out of ignorance when all biopsies were called “nonspecific chronic active colitis/proctitis” (regrettably still all too common). There was no appreciation that the chronic infections of this region whether venereal, tuberculous, or even amebic could be diagnosed and treated. Now we recognize that a vast majority of chronic proctitis encountered in clinical practice in Westernized societies represents IBD (UC/proctitis) and crypt architectural distortion, deep lamina propria plasma cells, cryptitis, and crypt abscesses are characteristic histologic changes. The other main form of IBD, CD, rarely presents as proctitis alone. However, it should also be recognized that the changes are not entirely specific, and as listed in the table other causes need to be considered in the differential diagnosis. Some of the chronic infections at times very closely mimic IBD and require a detailed clinicopathologic workup including demonstration of the pathogen in the tissues. Rarely, the changes of ulcerative proctitis can be those of a follicular proctitis.
 Figure 18-27. Ulcerative colitis in children. A: Diagnostic microscopic features such as diffuse mucosal inflammation and widespread architectural abnormalities may be absent from biopsy samples obtained in young children. The picture is from a first set of biopsies obtained in a 2-year-old boy with bloody diarrhea, which subsequently turned out to be UC. The photograph shows focal basal accumulation of inflammatory cells. High power illustrates disease activity (B). |
Table 18-8 Causes of Proctitis | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The precise extent of the disease should be determined by flexible sigmoidoscopy and biopsy, for it is found that endoscopically normal mucosa may be histologically inflamed.325 Biopsies above the proximal limit of disease may reveal that the disease is much more extensive than grossly evident. This is important to document as this finding may mark the patient for surveillance colonoscopies in the future. Hence, a case should not be signed out as ulcerative proctitis if biopsies are only available from the rectum. This is also the case when biopsies are available from the rectum and an area much more proximal such as the transverse colon, but excluding sigmoid and left colon. Such cases should be signed out as having changes limited to the rectum, but the differentiation between ulcerative proctitis and UC cannot be made in the absence of sigmoid and other more proximal biopsies.
In addition, approximately 10% of patients with ulcerative proctitis will develop more extensive ulcerative proctocolitis,326 whereas 15% will have recurrent bouts of active disease and 75% will enter permanent remission. Extension of disease usually occurs within 2 years and seldom after 5 years.324 There is some evidence that distal disease may predominate in the elderly population.327 The symptoms from severe proctitis may be disabling as a result of defecatory
frequency and urgency with blood but little or no diarrhea. In addition, chronic ulcerative proctitis may sometimes be remarkably resistant to medical therapy, including 5-aminosalicylic acid, local and even systemic steroids, or immunosuppressive agents. Some patients with only rectal involvement benefit greatly from restorative proctocolectomy.328 Morphologically, it has the features of UC, although the chronic inflammatory component may be more severe than those with more extensive disease even in remission.277
frequency and urgency with blood but little or no diarrhea. In addition, chronic ulcerative proctitis may sometimes be remarkably resistant to medical therapy, including 5-aminosalicylic acid, local and even systemic steroids, or immunosuppressive agents. Some patients with only rectal involvement benefit greatly from restorative proctocolectomy.328 Morphologically, it has the features of UC, although the chronic inflammatory component may be more severe than those with more extensive disease even in remission.277
Fulminant colitis and toxic megacolon. Fulminant colitis is a severe disease and can complicate any form of colitis, including IBD and infections, and in a proportion, the colon may have to be resected as an emergency measure (Figs. 18-28 and 18-29). The clinical criteria are listed in Table 18-9. In some, the etiology remains uncertain forever. About 5% to 13% of all patients with UC have an episode of fulminant colitis330, 331, either as a first attack or during an acute relapse. Sometimes a segment, of the bowel, most commonly the transverse colon becomes acutely dilated resulting in so-called toxic megacolon, and the bowel wall is consequently markedly thinned (Fig. 18-29C,D). Single or multiple perforations of the thinned bowel, either spontaneous or produced at the time of operation, were at one time common, but these days no colitic should be allowed to reach this state, and perforation is rare. Perhaps, the most important feature of these diseases is to appreciate that fulminant colitis of any cause can be accompanied by transmural inflammation and by fissuring ulcers. The presence of these does not indicate that the underlying disease is CD.
 Figure 18-28. Fulminant UC (A-F) and CD (G,H). A: Toxic dilatation, in UC, in which the bowel wall tears even on the most gentle handling (top right), and perforation (lower left of center). B: Fulminant total UC in the absence of dilatation, with deep ulceration going into and through the muscularis propria. C: Toxic dilatation with maximal disease in the vicinity of the hepatic flexure and a sharp margin of demarcation proximally in the ascending colon. D: Toxic dilatation with maximal disease in the descending and sigmoid colon (arrows), but with a very indistinct proximal margin in the ascending colon. |
Gross Appearances There is frequently a fibrinous or fibrinopurulent exudate on the peritoneal surface, and rarely overt perforation may be seen—not necessarily in the presence of toxic dilatation, but just severe disease. The cecum and ascending colon are
not invariably involved.332 There may be proximal or distal sparing, or both, even in cases of UC and can be misleading.316, 333 Further, the transition may be in the form of aphthoid ulcers. The inflammation can be focal or diffuse, the mucosa can be either diffusely hemorrhagic or there may be linear ulcers. The intestine may have the consistency of wet paper, tearing readily with subsequent peritonitis. Extensive loss of mucosa can occur and any surviving mucosa shows intense vascular congestion and edema, although sometimes with a relatively mild inflammatory cell response. In areas of mucosal ulceration, the submucosal tissues largely disappear laying bare the deep muscle coats, which may be covered by only a thin layer of very vascular granulation tissue, and may also be visible grossly.
not invariably involved.332 There may be proximal or distal sparing, or both, even in cases of UC and can be misleading.316, 333 Further, the transition may be in the form of aphthoid ulcers. The inflammation can be focal or diffuse, the mucosa can be either diffusely hemorrhagic or there may be linear ulcers. The intestine may have the consistency of wet paper, tearing readily with subsequent peritonitis. Extensive loss of mucosa can occur and any surviving mucosa shows intense vascular congestion and edema, although sometimes with a relatively mild inflammatory cell response. In areas of mucosal ulceration, the submucosal tissues largely disappear laying bare the deep muscle coats, which may be covered by only a thin layer of very vascular granulation tissue, and may also be visible grossly.
 Figure 18-28. (Continued) E: Fulminant colitis which had perforated proximally, but with train-track ulcers distally. In fulminant disease, this does not necessarily indicate CD. F: Fixed specimen showing residual mucosal islands, with ulceration that tends to be longitudinal. Multiple perforations are present, and light is visible through them (arrows). G,H: Fulminant colitis in CD. In both illustrations, there is a suggestion of residual mucosal islands reminiscent of cobblestoning, but other features were also present to support that diagnosis. |
 Figure 18-29. IBD presenting with severe hemorrhage (A,B) in UC. Note the diffusely hemorrhagic mucosa extending well back into the terminal ileum. |
 Figure 18-29. (Continued) C: Colectomy specimen in a case of fulminant UC with toxic megacolon. The dilatation as demonstrated by filling the colon with formalin is maximal in the transverse colon. D: The same specimen is opened up to reveal markedly inflamed and ulcerated mucosa with areas of the bowel wall that are paper thin. |
Depth and Type of Inflammation In fulminant UC, there is inflammation beyond the mucosa, often into, and sometimes through the muscularis propria, unlike that usually seen in UC (Fig. 18-30). However, the pattern of this inflammation is important. In fulminant UC, the active inflammation extends into the muscularis propria as a polymorphous infiltrate. This is quite different from the focal lymphoid hyperplasia with lymphoid aggregates and follicles, which are seen in a transmural distribution in CD. Granulomas are rare, but care has to be taken with mucosal granulomas that are much more frequently secondary to ruptured crypts or foreign material. The myenteric plexus may be incidentally involved, but the colonic dilatation might be due to a primary toxic atrophy of muscle cells,331, 334 although sometimes, dilated capillaries seem to be the cause of the myocytolysis (Fig. 18-30).
Table 18-9 Clinical Criteria for the Diagnosis of Fulminant Colitis | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||
In toxic megacolon, the inflammation extends deeper and the bowel wall is usually very thin. Frequently, one sees peritonitis as well as fibrinous or fibrinopurulent exudates on the peritoneal surfaces. Edema widely separates the fibers of the muscularis propria, leading to incipient perforation. The frequency of perforation in megacolon historically was seen up to 40% patients, but now is rare, but can also be seen occasionally in patients without megacolon. However, the term megacolon is dependent on radiological measurements of bowel diameter, which is unreliable as there is usually no control and normal varies hugely.
Fissuring ulcers are part of severe or fulminant disease irrespective of the underlying etiology, and do not imply CD (Fig. 18-30E). These may go into and sometimes through the muscularis propria, but perforation is rare. However, transmural inflammation is polymorphous and plasma cells, eosinophils, macrophages, mast cells, and lymphocytes are all invariably present. The diagnosis of CD requires as a minimum transmural lymphoid aggregates (not just one—see below), or granulomas away from areas of active inflammation so that foreign material including mucin can be excluded.
Residual mucosa is worthy of careful evaluation. While all have an excess of chronic inflammation with deep plasma cells to indicate the underlying chronic
colitis, some have overt diffuse architectural changes indicative of prior UC (Fig. 18-30B). These can be signed out as severe active UC. Where it is focal, the possibility of CD needs to be considered, but still requires the features described above.
colitis, some have overt diffuse architectural changes indicative of prior UC (Fig. 18-30B). These can be signed out as severe active UC. Where it is focal, the possibility of CD needs to be considered, but still requires the features described above.
 Figure 18-30. Histology of fulminant colitis. A: Deep undermining ulceration that is frequently hemorrhagic. The ulcer is seen reaching almost upto the muscularis propria in this area. Few lymphoid aggregates can be seen in the area of the ulceration. B: There is marked architectural distortion in the pseudopolyp present, implying chronic inflammation and possibly resolved ulceration. The distribution of these changes is useful as in this example the underlying disease is very likely to be UC. Also note that the inflammation in the area of ulceration extends into the muscularis propria (transmural); however, this is different from the typical lymphoid aggregates of CD. C,D: Extensive ulceration as seen here can be a feature in some cases with only scant islands of residual mucosa. The ulcer is seen resting upon the muscularis propria. E: Fissuring ulcers. These can be seen in fulminant IBD of any etiology, and can go into or through the muscularis propria and occasionally even perforate in the absence of toxic dilatation. |
However a third group of patients have virtually no architectural abnormality. Most of these present with a first attack of severe disease that is unresponsive to most therapies, may be held in check briefly with TNF-a antagonists, but ultimately undergo colectomy. The lack of architectural changes may cause the diagnosis of UC to (reasonably) be questioned, although chronic inflammation with a basal plasmacytosis may be present, representing disease that has been present for more than a couple of weeks (the time it takes to mount a good B-cell response). The key features here are that while polymorphous transmural inflammation and fissuring ulcers may be present, there are no other mucosal features to indicate that the underlying disease is UC. Neither are there features to indicate CD, whether transmural lymphoid hyperplasia or granulomas not attributable to mucin or foreign material and ideally away from fissures or ulceration and along the subserosa. They may represent IBD or severe undiagnosed infection, but their etiology remains unknown. The problem is in knowing what to call them. While indeterminate colitis/colitis indeterminate is one way out, this term has been used in so many contexts that it is highly ambiguous (see subsequent discussion regarding indeterminate colitis). A third possibility is to call them severe colitis of uncertain pathology. Whichever term is used, it is useful to state that there are no features to indicate CD, and therefore to preclude ileoanal anastomosis, but may be at the usual risk of pouchitis.
It should also be appreciated that fulminant colitis can occur during the course of many other causes of colitis and the intensity and distribution of ulceration is similar in all, often making it difficult to make a macroscopic diagnosis of the underlying disease.326 In those fulminant colitides where the histologic features do not allow the nature of the colitis to be determined, all other clinical, radiologic, endoscopic, and histologic material should be reviewed. It is important to obtain pretreatment as well as post-treatment biopsies, since post-treatment changes may be misleadingly patchy. The length of history may be helpful: if long, the colitis is unlikely to be due to infection. Stool culture may also be helpful. The fulminant stage is one in which a diagnosis of indeterminate colitis may be appropriate if the histologic features do not allow a confidant distinction between UC and CD.335
If there has been a colectomy, with the rectum left in situ, and an ileostomy formed, the ensuing histologic changes in the diverted rectum are only helpful until the features of diversion disease supervene, usually a couple of months, after which the features of diversion disease dominate the picture and do not allow an insight into the underlying disease.
One helpful feature in this situation is the difference in response to diversion shown by CD and ulcerative proctocolitis patients. Diversion of the fecal stream in UC usually results in severe changes of diversion proctitis superimposed on UC, whereas in diversion of Crohn’s proctocolitis, there is often remission of the inflammatory changes.336, 337 Diversion-related changes are usually well established within 3 months.338
Rectum in ulcerative proctocolitis following ileorectal anastomosis. The macroscopic and histologic changes in the rectum following ileorectal anastomosis are no different from those in the rectum in continuity with an inflamed colon. Biopsies from just above the anastomosis may be misinterpreted as small-bowel metaplasia. It is crucial that the pathologist is made aware of the previous operation of ileorectal anastomosis, when presented with a biopsy from this area. The dysplasia and carcinoma risk in this circumstance relates to the original disease extent: in total or extensive colitis, it will be as high as if the whole colon were still present. The pathologic changes of inflammation are quite different from those in the diverted rectum in UC.
Effect of medical therapy on the pathology of ulcerative colitis. It is possible that many of the difficulties concerning the classification of IBD on mucosal biopsies arise because of the effects of drug therapy, or of spontaneous remissions which can also be very focal. Furthermore, the difficulties are compounded by variations in the distribution of disease when modified by drug therapy, underpinning the importance of assessing multiple colonoscopic biopsies, from multiple sites, in the accurate diagnosis of IBD at the onset.272, 273
Immunosuppressives have been used effectively in severe UC otherwise unresponsive to conventional medical therapy, and sometimes at the very outset.339, 340 It has been recognized to have important consequences for the histopathologist and may lead to misdiagnosis of dysplasia.341 The changes induced by cyclosporin include villiform mucosal regeneration and epithelial regenerative changes which are severe with marked nuclear enlargement, sometimes extending to the mucosal surface, but usually with plentiful eosinophilic cytoplasm. Perhaps, the most helpful feature is that the “pseudodysplasia” induced by cyclosporin is strikingly diffuse, with many and sometimes all crypts showing similar changes, in a way not usually associated with UC-associated dysplasia. Despite this, it is very important that the clinician alerts the pathologist to the fact that the patient has been on cyclosporine and that the pathologist is cautious not to overdiagnose dysplasia in this circumstance.
 Figure 18-31. Histologic changes post TNF-α therapy showing marked reduction in inflammation and empty looking lamina propria (A), although grossly ulcer were present and patient was still very symptomatic (B). Also note the marked submucosal fibrosis. |
However, increasingly, TNF-α antagonists are the drug of choice in steroid-resistant severe UC. Whether these have any distinct effect on the histology of the UC is still unclear and needs further study. These can have a dramatic effect on the amount of inflammation present, so that when resection takes place, the disease often appears relatively quiescent although linear ulcers may still be present (Fig. 18-31).342
Where the features in individual biopsies, and especially in a series of biopsies, are being considered, marked variation in the quantity of acute and chronic inflammation within a biopsy or in a series of biopsies needs careful consideration. In intact (no erosion or ulcers) mucosa, it is possible in UC to get (a) relative rectal sparing of inflammation relative to proximal large bowel, and there may be no increase in chronic inflammation or (b) marked variation in both acute and chronic inflammation to the point that one part of the biopsy is heavily inflamed and the other is not. While this does occur in UC with treatment, one may need to ask is the question “Are there any other features present in this patient that might suggest the underlying disease is Crohn’s disease?”
Conversely, the typical biopsy of an aphthoid ulcer in CD with an erosion/ulcer on one side of the biopsy and the other side appearing normal is so typical of CD that even under therapy it needs to be considered. Such biopsies are only seen in UC in the severe/fulminant stage, when the clinical correlation should make the distinction easy.
Ileal pouches and their complications including pouchitis
Introduction Total proctocolectomy is an established surgical procedure for patients with UC and familial polyposis coli. While this may be followed with an ileostomy, various reservoir procedures have been proposed in order to create a more continent ileostomy. Initially, the double-folded ileal reservoir (Koch pouch) that has a large capacity had been used successfully for the construction of continent ileostomies, but used an artificial ileal valve and still required an ileostomy. It was almost completely replaced by the ileal pouch-anal anastomosis (IPAA) after colectomy and mucosal proctectomy, as elimination is more physiologic and an ostomy is avoided. The term “pouchitis” was introduced by Koch to describe inflammation of a continent ileostomy.343 It also refers to active inflammation of IPAA pouch mucosa. The condition is believed to be a “nonspecific, idiopathic inflammation of the neorectal ileal mucosa,”344, 345 which really says nothing but sounds good.
Increasingly, IPAA has become the treatment of choice for younger patients with UC or familial polyposis, but as age increases, especially over about 60, results tend to be less satisfactory, possibly because blood supply and healing are poor. IPAA is not usually carried out for CD as disease recurrence in the pouch can be a disaster. However, under exceptional circumstances (e.g., young patients who might be psychologically devastated going through their developmental years with an ileostomy), it may be considered. The conditions necessary for this, apart from the understanding of the risk of pouch recurrence of the disease are (a) no terminal ileal or small-bowel disease, (b) no anal/perianal disease, and (c) disease limited only to the large bowel. Some patients appear to have results comparable to UC, but even a decade of usable function at this age may be worth the risk.
Pouchitis is the inflammation in an ileal pouch following a continent ileal pouch and is the most common long-term complication after pouch surgery for UC. The true incidence is difficult to determine, as it depends on diagnostic criteria used to define the
syndrome, the accuracy of the evaluation, and the duration of follow-up. Reported incidence rates vary between <10% in patients undergoing pouches for familial polyposis and up to around 60% in patients who had severe colitis with backwash ileitis. The risk for development of pouchitis is highest during the first 2 years postoperatively. Three to twenty percent of patients with pouchitis develop persistent or recurrent episodes of pouchitis that frequently require immunosuppressive treatment. It may or may not be accompanied by clinical symptoms. Clinically, pouchitis typically results in increased stool frequency, liquidity, urgency, abdominal and/or pelvic discomfort, and often fecal incontinence. Therapy is with antibiotics, usually ciprofloxacin and metronidazole. Other adverse sequelae include mechanical, inflammatory, functional, neoplastic, and metabolic conditions. Common causes of failure include pouchitis, pelvic sepsis, and poor function. Occasionally, patients with chronic pouchitis develop Crohn’s-like GI and systemic complications including enteric stenoses or fistulas in the small bowel immediately proximal to the pouch, perianal fistulas, pouch stenoses or fistulas, arthritis, and pyoderma gangrenosum. Various subtypes of pouchitis have therefore been distinguished. The most common form has been called “usual pouchitis.” Persistent or recurrent disease is termed “chronic or refractory pouchitis.”345
syndrome, the accuracy of the evaluation, and the duration of follow-up. Reported incidence rates vary between <10% in patients undergoing pouches for familial polyposis and up to around 60% in patients who had severe colitis with backwash ileitis. The risk for development of pouchitis is highest during the first 2 years postoperatively. Three to twenty percent of patients with pouchitis develop persistent or recurrent episodes of pouchitis that frequently require immunosuppressive treatment. It may or may not be accompanied by clinical symptoms. Clinically, pouchitis typically results in increased stool frequency, liquidity, urgency, abdominal and/or pelvic discomfort, and often fecal incontinence. Therapy is with antibiotics, usually ciprofloxacin and metronidazole. Other adverse sequelae include mechanical, inflammatory, functional, neoplastic, and metabolic conditions. Common causes of failure include pouchitis, pelvic sepsis, and poor function. Occasionally, patients with chronic pouchitis develop Crohn’s-like GI and systemic complications including enteric stenoses or fistulas in the small bowel immediately proximal to the pouch, perianal fistulas, pouch stenoses or fistulas, arthritis, and pyoderma gangrenosum. Various subtypes of pouchitis have therefore been distinguished. The most common form has been called “usual pouchitis.” Persistent or recurrent disease is termed “chronic or refractory pouchitis.”345
Because pathologists may receive specimens at a variety of stages, it is necessary to understand a little about how they are done to understand the type of specimen that is received. Most pouches are now constructed in more than one step, as it has fewer complications overall as follows:
1. One stage, that is, colectomy, pouch, no ileostomy— this is rarely done because the leak rate tends to be high. Rarely, it may be done in familial adenomatous polyposis (FAP) patients: but virtually never in UC. When it is done, the pathologist receives an extensive pan-proctocolectomy that includes a mesorectal excision for the rectum, but without the anus. The pathologist may be unaware that a pouch has been carried out.
2. Two-stage procedures are done in two ways:
a. Subtotal colectomy and ileostomy first, and at a later date removing the rectum and making a pouch without an ileostomy. These are done mainly in patients with acute disease or high doses of steroids or where the diagnosis is uncertain. So, the pathologist initially receives a subtotal colectomy and subsequently both the rectum, and an ileostomy. This tends to be “standard” in patients with colitis. In experienced hands, the leak rate appears not to be increased in patients in whom a subtotal colectomy is carried out first, and then the pouch constructed without an ileostomy.
b. In patients with quiescent disease, a proctocolectomy may be carried out with construction of the pouch and an ileostomy. Later the ileostomy is taken down, which the pathologist receives, and continuity established between ileum, pouch, and anus.
3. Three-stage procedure: Subtotal colectomy is carried out initially, and this is followed by proctectomy and pouch construction, with a “covering” ileostomy. Finally, the ileostomy is taken down and continuity established. This is rarely done unless at surgery, there are problems or concerns with the ileoanal anastomosis.
There are several options for the pouch itself, but a J-shaped pouch (small-bowel loop folded back on itself) or less commonly an S-shaped pouch (an additional loop of small bowel goes into the pouch) are the most common. After the pouch has been constructed, the anastomosis lines are given about weeks or months to heal, during which time, the patient has a temporary ileostomy.
At one time, the residual rectal mucosa was removed till the anal canal, as residual mucosa in familial polyposis can still develop adenomas, and in IBD is at risk for dysplasia and carcinoma. However, this tends to cause more issues with the anal sphincter than it solves, so there is usually a small residual “cuff” of distal large-bowel mucosa that still has the original disease. In patients with UC, this can become inflamed as part of the underlying disease (“cuffitis”).
Pouch complications and inflammation in the postoperative period is often related to anastomosis leaks or breakdown, and some may have an ischemic basis.
Etiology and Pathogenesis It seems likely that pouchitis represents a reaction to small-bowel contents with stasis. However, the fact that pouchitis is rare in familial polyposis, but more common (up to 60% of patients) with IBD suggests that there is much more to it than this. Indeed, many have wondered if it might represent recurrent IBD in the pouch, particularly when it develops a large-bowel phenotype (colonic metaplasia).
Predisposing factors for pouchitis include the extent of colitis, the presence of backwash ileitis, polymorphisms in genes like the interleukin-1-receptor antagonist, the NOD2/CARD, and noncarrier status of TNF allele. Other factors include preproctocolectomy thrombocytosis, preoperative corticosteroid use, extraintestinal manifestations, especially PSC, the presence of p-ANCA, being a nonsmoker, and the use of NSAIDs.346
Chronic pouchitis patients may, in their pouches, have increased levels of interleukins-1 (beta), IL-6 and IL-8, and TNF-a in the pouch mucosa similar to mucosa of patients with UC. Several studies have found that antineutrophil cytoplasmic antibodies occur more frequently in patients with chronic pouchitis than in those without pouchitis. The development of Crohn’s-like GI complications in patients with chronic pouchitis may engender concern as to whether the patient was operated for UC or CD. However, the diagnosis should only be changed if review of the original resected specimen shows features of CD (see subsequent discussion for CD in the pouch).
Diagnosis The diagnosis of pouchitis is based on clinical, endoscopic, and histologic criteria. While most pouches show some degree of inflammation on histology, pouchitis is diagnosed only when it is accompanied by symptoms and endoscopic abnormalities. The most frequent symptoms include increased stool frequency and fluidity, rectal bleeding, abdominal cramping, urgency and tenesmus, and in the most severe cases, incontinence and fever. The most frequent endoscopic features include mucosal erythema, edema, friability, petechiae, granularity, and loss of vascular pattern, mucus exudates, erosions, and small superficial ulcerations (Fig. 18-32). Because of the great variability in the results of reports on the incidence of pouchitis, Sandborn and colleagues developed a Pouchitis Disease Activity Index (PDAI).347 This index consists of 18 points, calculated from three separate 6-point scores: symptoms, endoscopy, and histology. A score of 7 or more indicates active inflammation. The histology scale is composed of “polymorphonuclear infiltration,” subdivided into mild, moderate (crypt abscess) and severe, and mean “ulceration” per low-power field.347 Overall, clinical diagnosis based on symptoms is accurate in only 55% of the patients.348 Before treatment is started, it is important therefore to establish a correct diagnosis and to consider other causes of pouch dysfunction and inflammation, such as infections, irritable pouch syndrome (an IBS-like disorder affecting the pouch), and cuffitis. It has been suggested that irritable pouch syndrome may have an excess of serotonin-producing enterochromaffin cells, when compared to normal (not inflamed) pouches, but the overlap between groups was not sufficient to be of practical value.346
 Figure 18-32. Endoscopic appearance of pouchitis showing marked mucosal inflammation and exudates. |
Histology Usual pouchitis is characterized by the presence of a typical polymorphonuclear (neutrophil) infiltrate in the lamina propria and the epithelium in association with an increase in chronic inflammatory cells in the lamina propria (Fig. 18-33). Most commonly, neutrophils remain in the lamina propria with relatively little epithelial infiltration. In more severe cases, patchy intraepithelial neutrophils become more numerous and induce cryptitis, crypt abscesses, pyloric gland metaplasia, erosions, and ulcers (Fig. 18-33). The more severe grades of inflammation correlate well with the frequency of defecation and endoscopic appearance of inflammation. The inflammatory changes often have a patchy distribution with very focal erosions and ulcers (Fig. 18-33B), to the point that elsewhere in the GI tract, CD would immediately enter the differential diagnosis. However, this diagnosis is elusive in pouches as there can be marked overlap in morphology between CD and pouchitis. Two regions of the pouch, the lower and posterior, are the most frequent sites of characteristic findings. Consequently, a single random biopsy has limited diagnostic value.349 Histologic examination of pouch biopsies shows following features.
a. Acute inflammation and its sequelae: Neutrophils are rarely present in normal-appearing pouches.350 However, normal pouches are seldom biopsied, except when symptoms suggestive of pouchitis that are refractory to usual therapies are present.
b. Architectural changes: These vary from reduction in villous height often with crypt hyperplasia, to marked changes seen following ulceration, analogous to the changes seen in UC. Villous blunting and crypt hyperplasia may be a normal feature of healthy pelvic ileal pouch.
c. Colonic metaplasia: The concept of “colonic metaplasia” implies not only an almost flat mucosa resembling colonic mucosa but also a change in the type of mucin. If an alcian blue pH 2.5-high iron diamine stain is carried out, the pouch mucosa may show metaplasia to a large-bowel phenotype that,
unlike usual small bowel, does not produce the sulfated mucins. It is also supported by experimental data showing a different pattern of the expression of human tropomyosin isoform 5 (hTM5) in the ileal pouch. In genuine ileal samples, hTM5 is not expressed or focally expressed only in goblet cells. Pouch biopsies obtained at 6 months after surgery in one study showed a diffuse hTM5 staining in the goblet cells and in the nongoblet cells lining the crypts and the lumen. These changes are associated with shortening and reduced number of the villi.351 It has been suggested that these changes represent an adaptive response. “Colonic metaplasia” occurs more frequently in cases with severe pouchitis, suggesting that it may be a “reparative” rather than an “adaptive” response, although this can be interpreted either way.352
unlike usual small bowel, does not produce the sulfated mucins. It is also supported by experimental data showing a different pattern of the expression of human tropomyosin isoform 5 (hTM5) in the ileal pouch. In genuine ileal samples, hTM5 is not expressed or focally expressed only in goblet cells. Pouch biopsies obtained at 6 months after surgery in one study showed a diffuse hTM5 staining in the goblet cells and in the nongoblet cells lining the crypts and the lumen. These changes are associated with shortening and reduced number of the villi.351 It has been suggested that these changes represent an adaptive response. “Colonic metaplasia” occurs more frequently in cases with severe pouchitis, suggesting that it may be a “reparative” rather than an “adaptive” response, although this can be interpreted either way.352
 Figure 18-33. A: Pouch biopsies often show small intestinal mucosa with mild architectural distortion with shortening of the villi and infiltration of the lamina propria by a variety of cells such as lymphocytes, plasma cells, eosinophils and histiocytes. B: Active pouchitis can show intense mixed inflammation of the mucosa, which can be patchy, associated with neutrophils in the lamina propria (C), and subsequently into the crypts. D: With time more prominent architectural abnormalities and occasionally pseudopyloric metaplasia can be seen. E: The pouch mucosa sometimes may start looking more like colonic mucosa (colon metaplasia). F: In severe cases, extensive mucosal ulceration can be seen. |
 Figure 18-33. (Continued) G: Crypt rupture granulomas can be seen like many other inflammatory conditions, and should not lead to the diagnosis of CD. H: Sections from a resected pouch shows transmural lymphoid aggregates which can further mimic CD. |
d. Pyloric metaplasia: This is seen readily as a response of the small bowel to ulceration, so it is not surprising to find that it is seen particularly in pouches that have had severe ulceration (Fig. 18-33D).353
e. Muscle changes: As expected, following ulceration there is often duplication of the muscularis mucosae, as seen elsewhere in the GI tract, and submucosal fibrosis. There may also be fusion between the muscularis mucosae and muscularis propria.
f. Chronic inflammatory changes: These are found in most (up to 87%) of the pouches.354 The findings consist of infiltration of the lamina propria by a variety of cells such as lymphocytes, plasma cells, eosinophils, and histiocytes. Increase in IEL in the pouch seems not to be part of pouchitis, and if marked should raise the possibility of associated celiac disease.
g. Granulomas: These can also be part of pouchitis, but largely represent mucin granulomas (Fig. 18-33G). Great care should be taken not to interpret these biopsies as being “consistent with Crohn’s disease,” as this becomes translated clinically into the patient having “histologically proven CD of the pouch.”
In patients presenting with chronic refractory pouchitis, possible causes such as infections (particularly CMV) must be considered.
Other causes of pouch problems. These include a variety of noninflammatory conditions of the pouch, but biopsy may be carried out to exclude pouchitis. These include symptoms resulting from decreased pouch compliance, irritable pouch syndrome, strictures (inlet, outlet, anastomosis lines) efferent limb syndrome, decreased pouch emptying, and pelvic floor dysfunction. Endoscopy and biopsy are critical in excluding pouchitis to distinguish these—unless it coexists. Usually, when pouchitis causes symptoms, it is quite severe with erosions and ulcers. Minor chronic inflammation appears not to be associated with symptoms.
When are pouch biopsies done? Pouchoscopy with biopsies may be carried out for pouch dysfunction symptoms—usually when not responding to antibiotics, to confirm or refute pouchitis, or delineate other potential causes of diarrhea, and if pouchitis is present, the question of whether a treatable cause is present such as C. difficile or CMV infection. The question of whether CD is present may be asked, but this diagnosis is fraught with hazard (see subsequent
discussion). The presence of “cuffitis” (inflammation in the residual couple of cms of rectal cuff—most distal rectum) can cause symptoms similar to ulcerative proctitis, but can be surprisingly difficult to treat. An occasional problem is that of the presence of “prepouch ileitis” (see subsequent discussion).
discussion). The presence of “cuffitis” (inflammation in the residual couple of cms of rectal cuff—most distal rectum) can cause symptoms similar to ulcerative proctitis, but can be surprisingly difficult to treat. An occasional problem is that of the presence of “prepouch ileitis” (see subsequent discussion).
Pouchoscopy may also be carried out in patients with FAP, as there may be recurrence of adenomas, and rarely carcinomas, in both the pouch and the rectal cuff mucosa. Similarly, inflammatory polyps may form especially at anastomosis lines, but also as a result of pouchitis. In patients with IBD, while dysplasia and carcinoma are rare they are well described (see subsequent discussion).
Standard Biopsy Run in Pouches Besides examining and biopsing any lesions (severe inflammation, ulcers, raised lesions, or polyps) in the pouch and rectal cuff, the prepouch ileum should also be examined and biopsied (proximal and distal prepouch ileum), especially if any lesions are present.
Reporting Pouch Biopsies If biopsies are without significant abnormality—say so. However, most pouches are inflamed and the inflammation can be localized, patchy, or diffuse with or without erosions and ulcers.
The quantity of acute and chronic inflammation can be subjectively graded as mild, moderate, and severe, and associated features also described (erosions, ulcers, etc.). However, this represents tradition rather than being of much clinical value unless marked, and is likely not reproducible unless severe. The presence of marked villous blunting, erosions, and ulcers or other features indicative of ulceration (pseudopyloric metaplasia) are noteworthy as they can be associated with severe pouchitis. Granulomas, whether related to ruptured crypts (mucin) or foreign body are reported, but care should be taken not to report these as being “consistent with Crohn’s disease.” An intraepithelial lymphocytosis is not part of pouchitis and should raise the question of the usual causes of an intraepithelial lymphocytosis, especially celiac disease. There is a single case report of a patient with collagenous colitis who underwent proctocolectomy and IPAA for refractory disease.355 The changes did not extend into the afferent limb, neither was the cuff of terminal ileum in the proctocolectomy specimen involved. Both symptoms and collagen band in the pouch resolved following antibiotic therapy,355 implying once again the potential role of luminal contents, especially bacteria in its genesis.
The presence of dysplasia or neoplasia clearly requires a comment. The presence of dysplasia or carcinoma in pouches is rare, but well described and those with dysplasia or carcinoma in the resected colectomy specimen are at a higher risk. Some patients with FAP regularly seem to produce adenomas in the pouch.
Changes in Resected Pouches About 10% of pouches are ultimately removed. This is usually the result of failure of medical treatment, the presence of fistulas, and is usually carried out sometime after a defunctioning ileostomy has been carried out. The features found may therefore be those of either pouchitis or “diversion disease” or both, and can closely resemble CD histologically.
Spectrum of changes seen include redness, erosions, aphthous ulcers, or deep ulcers with fissuring and/or fistulas. The original anastomosis lines may still be identifiable grossly. Histologically, deep and even transmural lymphoid aggregates that mimic CD may be present. Granulomas are usually associated with foreign material or ruptured crypts, and we have seen a severe granulomatous venulitis that seemed not to be indicative of underlying CD, and was not present in the prior colectomy specimen. Some resected pouches have overt ischemic changes, some of which were clinically suspected of being CD. Ischemia therefore always needs to be borne in mind when examining resected pouches. Similar changes, including granulomas, have been observed in defunctioned rectal stumps left in situ after urgent total colectomy for UC. Thus a diagnosis of CD after IPAA surgery should be made extremely cautiously.283
In some patients, the more striking changes may be found in the afferent limb going into the pouch (prepouch ileitis). However, it is unclear if this really indicates CD, even though clinically this is disconcerting. NSAIDs-associated ulcers can also be seen in patients with IPAA.356
Dysplasia and cancer in pouches. Primary adenocarcinoma in pouches is rare. Dysplasia in the form of recurrent adenomas is not uncommon in patients with FAP, and is usually treated with surveillance and polypectomy; however, pouch carcinoma in FAP is rare.
In patients with IBD, dysplasia is rare in prospective studies; nevertheless, these patients can develop carcinoma within the pouch. The cancer may occur in the rectal cuff with extension into the lumen of the pouch or as a result of a pelvic recurrence infiltrating from the outside. When the carcinoma extends from the rectal cuff into the pouch, it may be difficult to decide its primary site. Under these circumstances, the presence of dysplasia or carcinoma in the prior colectomy favors that the primary was in the residual large bowel rather than the pouch. While common sense would suggest that one should look for dysplasia in the adjacent mucosa, it is known that carcinomas at either site, and even that coming in from the outside, can use the normal crypt framework to grow around, thereby mimicking dysplasia. Managing dysplasia in the pouch or cuff is usually by surveillance until there are features that suggest
invasion may be present. However, occasionally carcinomas just appear “out of the blue” and rarely with metastasis. Because dysplasia is so uncommon in pouches there are no good guidelines, so each patient has to be managed by weighing the pros and cons of all options for therapy, including local excision if a lesion is visible.
invasion may be present. However, occasionally carcinomas just appear “out of the blue” and rarely with metastasis. Because dysplasia is so uncommon in pouches there are no good guidelines, so each patient has to be managed by weighing the pros and cons of all options for therapy, including local excision if a lesion is visible.
Refractory Pouchitis The incidence of pouchitis increases with length of follow-up rising to about 50% at 5 years with a prevalence of clinically significant recurrent or chronic refractory pouchitis in 5% to 10% of patients. Because most pouchitis is treated effectively with antibiotics, the question is whether refractory disease represents a change in bacterial flora to a resistant strain. There are data suggesting that this may be the case, and extended-spectrum beta-lactamases (ESBL)-producing organisms are capable of hydrolyzing penicillins, broad-spectrum cephalosporins, and monobactams. ESBL-producing bacteria are inherently resistant to many classes of antibiotic because other determinants of resistance are often linked on the same plasmid. They appear linked to refractory pouchitis, and are also associated with prepouch ileitis, but have no implications for pathology.
Crohn’s Disease (CD) of the Pouch Clinically, CD of the pouch is suspected if there is perianal disease, a fistula that occurs more than 1 year after the IPAA was constructed, and especially if this is unrelated to a suture line. Endoscopically, the presence of features that look like Crohn’s elsewhere, especially markedly focal ulcers, and the presence of prepouch ileitis, all raise the index of clinical suspicion. There is a modest literature, primarily from the Cleveland Clinic, regarding CD in the pouch.357 While we have little doubt that this occurs, it is an elusive histologic diagnosis as the features of pouchitis, and prepouch ileitis, can completely mimic those of CD, similar in some ways to diversion disease or SCAD. There are therefore no good histologic criteria to diagnose CD of pouch and it is a good idea to state this specifically in the report. Features commonly seen in pouch include aphthoid ulcers, focal inflammation, mucosal granulomas, and transmural lymphoid aggregates that can be quite marked but are not diagnostic of CD (Fig. 18-33H). Specifically, it should be noted that mucosal granulomas can be related to foreign material including mucin. While submucosal granulomas are far more suggestive, however, these are more likely to be seen in resections rather than mucosal biopsies. We have even seen a severe granulomatous perivenulitis that immediately prompted review of the colectomy and clinical correlation, but with no other features to suggest that this was CD.
Our current take on this is that there is far too much overlap in both biopsies and resections between CD and severe pouchitis to allow a diagnosis of CD in a pouch (or prepouch ileum), with any degree of confidence or reproducibility. It is one of those situations where the clinician needs to understand that taking biopsies of the pouch with the question “rule out Crohn’s disease” is one that cannot be answered, so should not be asked.
Upper GI Changes in Pouchitis While it is tempting to believe that the involvement of the upper GI tract is indicative of CD, we know that this is not the case in UC, and the same may be true with pouchitis. The fact is that a variety of changes can be found endoscopically and on biopsy. In one study, inflammatory changes were found in about 12% of gastroduodenal biopsies in such patients.358 The issue is how these are interpreted. In this paper, they were interpreted as being indicative of CD; however, as this tends to be correlated with severity of pouchitis, it could be also interpreted as “sympathetic gastroduodenitis.” Currently, we reserve our judgment on how these should be interpreted, but prefer not to label the patient as having “Crohn’s disease” solely based on this evidence.
Other diseases affecting pouches. It should be recalled that any diseases or tumors that affect the large or small bowel can occur in the pouch or rectal cuff. Mesenchymal tumors such as inflammatory fibroid polyps have been reported; however, lymphoma seems to be the most common nonepithelial tumor in pouches.359, 360 Surprisingly, we were unable to find examples of carcinoid or gastrointestinal stromal tumors (GISTs), although they undoubtedly can occur and one is described in an ileal neobladder.361 Microcarcinoids in the rectal cuff have been reported.362
Prepouch Ileitis This is a puzzling condition that tends to be seen in patients with severe pouchitis, and has several variants. About 13% of patients with pouchitis will develop prepouch ileitis; this is a histologic entity distinct from CD and can be defined as inflammation proximal to the pouch in a patient with pouchitis and with histologic features, essentially similar to those found in pouchitis, including colonic metaplasia.319, 363 These include
a. Inflammation in the afferent limb of the pouch that is in continuity with inflammation in the pouch— a sort of backwash prepouchitis. However in some patients, the inflammation is more severe in the afferent loop than in the pouch, and in a minority, the pouch is uninflamed but the afferent limb
overtly inflamed. Occasionally, there may be some form of obstruction or stricture at the inlet, but usually this is quite patent.
overtly inflamed. Occasionally, there may be some form of obstruction or stricture at the inlet, but usually this is quite patent.
b. Inflammation in the afferent loop, usually with erosions and ulcers, that is discontinuous with the pouch, so is a skip lesion. That is, there is normal mucosa between inflamed pouch and the lesion in the afferent limb. When this occurs, it is wise to biopsy the normal intervening mucosa to document that this really is the case.
The mention of “skip” lesions immediately raises the question of CD, and some of these patients do respond to therapy for CD, including steroids and TNF-α antagonists. Indeed, they are usually only detected because the pouch is scoped for refractory symptoms, so by definition, they are resistant to the usual antibiotics. However, so many inflammatory diseases of the bowel respond to these medications, including UC, that this is not a reason to call it CD. There is rarely evidence of CD more proximally in the small bowel and this often appears to be an isolated finding of unknown etiology. It is best to sign out these biopsies descriptively without stating that the changes are “consistent with Crohn’s disease” as, like pouchitis, there are no good criteria to separate the two. The presence of numerous wellformed noncaseating granulomas having no other etiology could be used as a criterion, but this seems extremely rare.
Cuffitis Cuffitis, or inflammation of the most distal rectal “cuff” of mucosa not removed at the time of proctectomy represents recurrent/persistent UC with identical histology in the residual rectal remnant in patients with a stapled anastomosis. Biopsies show features of active UC. However, sometimes the endoscopist may not be aware how low down they are and the biopsy may be submitted as “distal/lower pouch.” Conversely, the endoscopist may submit biopsies of “rectal cuff” that is clearly small bowel in origin and therefore represents distal pouch and not the cuff. Looking at the biopsies “blind” without looking at the described site of origin is the easiest way to overcome this expectation bias. Sometimes, dysplasia or carcinoma can be found in biopsies from the rectal cuff.
Specific infections such as C. difficile or CMV may occur in the cuff, but pseudomembranes are rare in patients with UC and C. difficle infection often look like a recurrence of the underlying colitis. It should also be recalled that occasionally the cuff biopsy may have features of CD, which causes us to go back and review the original resection if available, and also enquire whether the patient has other overt features of CD. However this is also rare.
Crohn’s Disease
Introduction and definition. This disease is traditionally attributed to Burrrill Crohn, Leon Ginsberg, and Bernard Oppenheimer, gastroenterologists from New York, who recognized this disease as being distinct from tuberculosis, and described a series of patients in 1932 with inflammation of the terminal ileum.364 Because of the distribution of the inflammation in the disease, it was entitled regional ileitis or regional enteritis (rather than the “terminal ileitis” it was originally called, to avoid the implication that the disease was “terminal”). The condition, however, had been previously independently identified by others. Probably, the first description of this condition was made by the Italian physician Giovanni Battista Morgagni around 1769, when he diagnosed a young man with a chronic, debilitating illness and diarrhea.365 Later cases were reported in 1898 by John Berg and in 1904 by a Polish surgeon Antoni Lesniowski.142 In the Polish literature, the condition is thus named Lesniowski-Crohn’s disease. In 1913, a Scottish physician T. Kennedy Dalziel (pronounced Dalyeel), at the meeting of the British Medical Association, described nine patients with intestinal obstruction that were likely CD,366 likening the disease to “an eel in the state of rigor mortis.” In 1960, Lockhart-Mummery and Morson described the colonic involvement (Crohn’s colitis), when they distinguished its macroscopic and histologic features from UC.367, 368
CD is a lifelong inflammatory disease of unknown etiology and constitutes one of the two major IBD besides UC. It can involve any segment of the GI tract from mouth to anus. Patients may have ileocolitis (˜40%), disease limited to the small bowel, especially the terminal ileum (˜30%),369 or it may be limited to the colon (˜15%-30%). Colonic CD has three major patterns of distribution: a diffuse colitis that may be difficult to distinguish from UC macroscopically, stricturing disease, and Crohn’s proctitis. These patterns may coexist or change from one to another during the course of a patient’s illness. It is important to remember that the rectum is macroscopically normal in about 50% of cases of Crohn’s colitis.
CD differs from UC largely by distribution of the lesions, the presence of transmural inflammation, and often noncaseating epithelioid cell granulomas. CD is a segmental disease with skip areas and more commonly involves the proximal colon and ileum. Although gross examination of a resection specimen allows its differentiation from UC in almost 90% of the cases, in practice, this distinction needs to be made prior to the surgery. Although there are subtle differences in the histologic features of UC and CD, the differentiation on biopsies is largely based on the finding of granulomas in combination with the presence of features diagnostic for IBD. The presence of
small intestinal involvement, rectal sparing, and skip lesions are other helpful features. The patchiness that may be evident in UC is more subtle than the discrete skip lesions of CD. Inflammation only evident in the rectosigmoid and the splenic flexure with a long segment of normal left colon is more of a skip lesion (and consistent with CD) than patchiness. It is thus appropriate to obtain multiple biopsies from different segments of the colon and the terminal ileum. Furthermore, in difficult situations, biopsies of the upper GI tract may be indicated to assess the extent of the disease and to solve diagnostic problems. These may help if the changes are typical; however, gastroduodenal changes can also be found in UC (see later).
small intestinal involvement, rectal sparing, and skip lesions are other helpful features. The patchiness that may be evident in UC is more subtle than the discrete skip lesions of CD. Inflammation only evident in the rectosigmoid and the splenic flexure with a long segment of normal left colon is more of a skip lesion (and consistent with CD) than patchiness. It is thus appropriate to obtain multiple biopsies from different segments of the colon and the terminal ileum. Furthermore, in difficult situations, biopsies of the upper GI tract may be indicated to assess the extent of the disease and to solve diagnostic problems. These may help if the changes are typical; however, gastroduodenal changes can also be found in UC (see later).
Following therapy, the inflammatory changes in UC may become patchy and preferential resolution of rectal inflammation may lead to impression of rectal sparing making distinction from CD difficult. The endoscopic appearance of involvement or normalcy is not always reliable and histologic confirmation is necessary. Thus in many situations, the best chance of differentiation between UC and CD is at the time of initial diagnosis before treatment is started. It is thus imperative that a proper workup with extensive sampling of colon and terminal ileum be undertaken at this time. It is important to define the junction between involved and normal mucosa, and rectum should be biopsied even if it looks endoscopically normal. Examination of the terminal ileum can be difficult. However, it should be biopsied especially in cases of extensive or pancolitis. In practice, the differentiation between UC and CD is not always possible based on histology alone, and requires clinical correlation with endoscopic, histologic, radiologic, and biochemical findings. Even then it may not be possible.
Clinical features and classification. The symptoms and signs of CD are heterogeneous and depend on the extent and the sites of intestinal and extraintestinal involvement. Crohn’s colitis usually presents with diarrhea, rectal bleeding, or perianal disease; recurrent abdominal pain and intestinal obstruction are much less conspicuous than in small-bowel disease. Anal and perianal disease, including oedematous skin tags, cavitating ulcers, fissures, fistulae, abscesses, and anal canal strictures may occur in 75% of patients with colonic CD at some time during the course of their illness; such manifestations are more common during severe attacks when the colon is extensively involved. They are also associated with other extraintestinal manifestations.370 Abnormalities of the musculoskeletal system are the most common extraintestinal manifestations.
Because the signs and symptoms of CD are often subtle, the diagnosis is frequently delayed. However, the diagnosis can usually be made on the basis of a careful clinical history, physical examination, and diagnostic testing. The most reliable and sensitive method for differentiating between UC and CD is colonoscopy with endoscopically directed colonic biopsies. However, abdominal MRI or CT scan revealing typical small-bowel disease, intra-abdominal lymphadenopathy, creeping fat, or hypervascularity (the classic “comb’s sign”) can help to clinch a diagnosis of CD.
Based on the macroscopic description of the lesions, different clinical phenotypes of CD have been recognized which have an influence on the presentation and the natural course of the disease. In the Vienna classification of CD proposed by an International Working Party for the World Congress of Gastroenterology in 1998, a distinction was made between inflammatory disease, stricturing disease (occurrence of constant luminal narrowing), and penetrating disease (occurrence of intra-abdominal or perianal fistulas, inflammatory masses, and/or abscesses). However, fistulas are often associated with strictures and therefore the clinical distinction is often limited to two types, the perforating type and the nonperforating type with predominantly mucosal lesions.371 The Vienna classification incorporated the variables, age of onset (A), disease location (L), and disease behavior (B). This resulted in 24 possible different subgroups based upon combination of the variables. However, variables such as disease location and behavior have been found to be highly correlated, making the number of categories in practice much smaller. A modification of the Vienna classification known as the Montreal classification has been adopted as the current working approach in clinical studies. The modifications include introduction of an early age of onset category (diagnosis at 16 years or under), inclusion of a modifier for perianal disease, and the classification of upper GI involvement, which can be added to other site of classification or be the lone site of classification (as is seen in about 5% of cases).155 Both classifications are shown in Table 18-10.
Endoscopic features and gross pathology. The appearances of CD are fundamentally similar at all levels of the GI tract. The terminal ileum and proximal colon are the commonest site, followed by the anorectum and colon. Perianal disease is common although the reported frequency varies between 14 and 76%;372 a population-based study suggested it was closer to 23%.373 Upper GI tract can also be involved (see later), and is usually associated with disease of the terminal ileum. The ileocolonic lesions may however take time to develop and occasionally CD can start as isolated gastric or gastroduodenal disease. Unlike CD in the small bowel, where there is often a clear demarcation between diseased and normal intestine on gross
examination, the transition between affected and unaffected areas in the colon is usually less clear-cut.
examination, the transition between affected and unaffected areas in the colon is usually less clear-cut.
Table 18-10 Vienna and Montreal Classification for Crohn’s Disease | |||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |||||||||||||||||||||||||||||||||||||||
The serosa of the bowel is often hyperemic and erythematous with a slightly dusky blue appearance due to vascular congestion, and there may be a covering of inflammatory exudate. There may be dense fibrous adhesions to other loops of bowel or other intra-abdominal organs. Normally, the mesenteric fat stops at its junction with small intestine, and remaining circumference of the intestine is entirely covered by shiny serosa without any fat. In CD (and any other underlying chronic inflammatory disorder involving the serosa), the mesenteric fat creeps on to the surface of the intestine to partially or completely surround the involved segment and is known as “fat wrapping” or “creeping fat” (Fig. 18-34). At least 50% of the intestinal circumference should be affected for it to be recognized as creeping fat. This phenomenon is observed in 75% of the surgical specimens of CD. It is said to be specific for CD and has not been reported in UC,374 but we have observed it in other diseases causing transmural inflammation such as NSAID-associated ulceration or rarely chronic ischemia, and is almost certainly a reaction to any form of transmural inflammation. Fat wrapping is usually present in the colon; however, it may be difficult to assess as pouches of fat (appendices epiploica) normally cover part of the serosal surface of colon, and of course this will not be assessable in the lower part of the rectum which is normally entirely surrounded by fat. The fat has been shown to produce TNF-a, and is probably actively involved in the inflammatory process.375, 376 The corresponding mesentery is also usually thickened and retracted. Fibrous strands are present in the mesenteric fat, radiating from the intestine and surround thickened, hypertrophic fat lobules. The mesenteric lymph nodes are commonly enlarged. Being a transmural disease, the bowel wall is thickened with involvement of the submucosa, the muscularis propria, the subserosa, and mesenteric fat.315
 Figure 18-34. A,B: Mesenteric fat partially surrounds the intestine, extending from the mesenteric attachment anteriorly and posteriorly corresponding to the involved segment. Variable degree of encroachment of the circumference of the intestine occurs, and as seen as here focally it completely encircles the small bowel. This phenomenon, known as fat wrapping, is fairly specific for CD. |
Mucosal lesions in CD are variable (Figs. 18-35 and 18-36). The pathologic hallmark of colonic or intestinal CD is discontinuous, patchy, or focal disease and this is seen most obviously from the mucosal aspect. Mucosal ulcers are common and may vary
within a single specimen from deep serpiginous fissuring ulcers to tiny aphthous ulcers surrounded by normal or mildly edematous bowel (Figs. 18-36 and 18-37). Larger ulcers are usually discrete with edematous, overhanging, slightly violaceous edges, and they are often stellate in outline.367, 368, 377, 378, 379 Other examples have a more linear “tramline” appearance with two, and often three, roughly parallel lines of discrete ulceration running along the length of the colon. These are usually related to the point of entry of mesocolic vessels into the bowel wall. When the linear parallel ulcers are joined by a horizontal ulcer at the top, they resemble “bear-claw” markings or so-called “bear-claw ulcers.” After ulcers heal, tramline indentations of the mucosal surface frequently remain. Another appearance that is typical but not characteristic of CD is cobblestoning of the mucosal surface (Figs. 18-36 and 18-37). This results from areas of marked edema of the mucosa and submucosa separated by crevices that represent narrow fissuring ulcers, or aphthous ulcers that coalesce either horizontally or vertically or both. Fistulae are found in up to 60% of patients, whereas overt perforation of active colonic CD is uncommon, and usually right sided. This probably reflects the fact that the inflammatory process penetrates the tissue planes slowly and causes loops of inflamed bowel to adhere to one another, effectively walling off any perforation or abscess that may have formed. Fistulae, perforations, and abscesses form the base of the fissuring ulcers, where there is extension of the inflammatory process into the serosa and adjacent structures (Fig. 18-38). It can be difficult to find fistulae. Gentle probing, especially if recesses in the mucosa are seen may be helpful. Sometimes, pushing aside inflammatory polyps and looking for openings that may be immediately adjacent to them (sentinel polyps) may be helpful. They may also be visualized when strictures are bread loafed as thinly as possible (about 2 mm thickness). Indeed, it is quite a good practice to do this routinely to find unexpected fistulas and even an occasional carcinoma. In the fixed specimen, serial sections a few millimeters apart going transversely across the bowel but without destroying the integrity of the specimen is a simple and quick way of finding intramural fissures and fistulas.
within a single specimen from deep serpiginous fissuring ulcers to tiny aphthous ulcers surrounded by normal or mildly edematous bowel (Figs. 18-36 and 18-37). Larger ulcers are usually discrete with edematous, overhanging, slightly violaceous edges, and they are often stellate in outline.367, 368, 377, 378, 379 Other examples have a more linear “tramline” appearance with two, and often three, roughly parallel lines of discrete ulceration running along the length of the colon. These are usually related to the point of entry of mesocolic vessels into the bowel wall. When the linear parallel ulcers are joined by a horizontal ulcer at the top, they resemble “bear-claw” markings or so-called “bear-claw ulcers.” After ulcers heal, tramline indentations of the mucosal surface frequently remain. Another appearance that is typical but not characteristic of CD is cobblestoning of the mucosal surface (Figs. 18-36 and 18-37). This results from areas of marked edema of the mucosa and submucosa separated by crevices that represent narrow fissuring ulcers, or aphthous ulcers that coalesce either horizontally or vertically or both. Fistulae are found in up to 60% of patients, whereas overt perforation of active colonic CD is uncommon, and usually right sided. This probably reflects the fact that the inflammatory process penetrates the tissue planes slowly and causes loops of inflamed bowel to adhere to one another, effectively walling off any perforation or abscess that may have formed. Fistulae, perforations, and abscesses form the base of the fissuring ulcers, where there is extension of the inflammatory process into the serosa and adjacent structures (Fig. 18-38). It can be difficult to find fistulae. Gentle probing, especially if recesses in the mucosa are seen may be helpful. Sometimes, pushing aside inflammatory polyps and looking for openings that may be immediately adjacent to them (sentinel polyps) may be helpful. They may also be visualized when strictures are bread loafed as thinly as possible (about 2 mm thickness). Indeed, it is quite a good practice to do this routinely to find unexpected fistulas and even an occasional carcinoma. In the fixed specimen, serial sections a few millimeters apart going transversely across the bowel but without destroying the integrity of the specimen is a simple and quick way of finding intramural fissures and fistulas.
 Figure 18-35. Crohn’s disease. Endoscopic pictures of a small aphthoid ulcer (A) and of large irregular, serpiginous ulcers (B,C). D: Chronic disease with abnormal granular mucosa loss of folds. |
 Figure 18-36. Crohn’s disease. Gross features include the presence of small irregular ulcers which may be the result of confluent aphthoid ulcers (A). They may become longitudinal ulcers (B). |
 Figure 18-36. (Continued) The combination of longitudinal and transverse ulceration in an edematous mucosa is responsible for the characteristic cobblestone appearance (C,D). |
Strictures (Fig. 18-39) in CD usually result from transmural inflammation, fibrosis, and fibromuscular proliferation. They have no particular macroscopic distinguishing features from those due to ischemia, chronic infective disorders, and drug-induced strictures (see below), but they may serve in the distinction from UC, in which they are very uncommon. On the other hand, diffuse colonic involvement with CD, while uncommon, can be difficult to distinguish macroscopically from UC. Useful pointers toward CD include rectal sparing, skip areas, and the presence of anal or perianal inflammation. Mucosal inflammatory polyps may be seen in CD, sometimes adopting giant proportions up to 5 cm in maximum dimension.301, 380 Tall, narrow filiform polyps may also be seen, and rarely a forest of polyps (colitis polyposa or filiform polyposis) may be present. A particular feature of some cases of CD is the presence of a “sentinel” inflammatory polyp on the proximal side of an ulcerated stricture.
Serology has been touted to distinguish CD from UC. In all comers with CD, ASCA+ ANCA− highly correlates with CD and ASCA− ANCA+ highly correlates with UC. However, the studies touting this include small-bowel CD where serology is unnecessary for diagnosis, as ANCA+ tends to be associated with large bowel-disease and ASCA+ with small-bowel disease. However, in isolated CD colitis pANCA may be positive in 50% and hence in the very situation where
distinction from UC is problematic clinically (unless fistulas or granulomas are present) serology is not helpful.274
distinction from UC is problematic clinically (unless fistulas or granulomas are present) serology is not helpful.274
 Figure 18-37. Crohn’s disease. Aphthoid ulcer-cobblestone phase of CD. A: In this ileocolic anastomosis resection because of a fistula at the anastomosis, recurrent aphthoid ulcers are visible, limited almost entirely to the terminal ileum, stopping abruptly at the colonic margin. They vary from pinpoint areas of redness to small, punched-out ulcers. B: If activity continues, the ulcers enlarge (left), become stellate, and link with each other both transversely and longitudinally, forming islands of mucosa which are the cobblestones. This appearance is accentuated by the marked submucosal edema present in the areas of cobblestoning. The longitudinal ulceration is usually more marked, forming tram track ulcers (right). C: Continuation of the same process results in deep transverse ulcers in addition, with reemergence of cobblestoning, although these areas are now discrete residual mucosal islands (bottom). D: If the process continues, there is marked stricturing resulting from a combination of cobblestoning, thickening of the wall with inflammation, fibrosis, and extensive creeping fat, all of which prevent expansion of the lumen. |
 Figure 18-38. Fistula formation between loop of a small bowel and transverse colon. |
Microscopic pathology. The pathologic features that are characteristic of CD can only be appreciated on resection and only limited features are seen in mucosal biopsies. Thus, the diagnosis is invariably easier on resected specimens than in mucosal biopsies378, 379; however, in practice the approach to diagnosis varies on biopsies compared to resection and both are considered briefly here.
Mucosal Biopsy The most typical histologic feature of CD in mucosal biopsies is patchy or focal chronic active inflammation that includes an erosion at one end of the biopsy, and can occur in the ileum or large bowel (Fig. 18-40A-D). But it is especially
noticeable when multiple biopsies are examined from the same patient, either synchronously or metachronously.277, 381, 382, 383 However, usually the initial exercise is in deciding if the patient has features of IBD (architectural distortion and deep plasma cells, the latter distal to the region of the ileocecal valve, and then decide which type of IBD is present). In deep biopsies it may be seen to extend into the submucosa, where its density may be proportionately greater than in the mucosa. Sometimes the infiltrate contains eosinophils or mast cells, which can be numerous but are not distinctive. Granulomas are a bonus, and great care needs to be taken to ensure that they are not at the site of ruptured crypts, and therefore mucin granulomas. They are not a sine qua non of CD in a patient who clearly has IBD.
noticeable when multiple biopsies are examined from the same patient, either synchronously or metachronously.277, 381, 382, 383 However, usually the initial exercise is in deciding if the patient has features of IBD (architectural distortion and deep plasma cells, the latter distal to the region of the ileocecal valve, and then decide which type of IBD is present). In deep biopsies it may be seen to extend into the submucosa, where its density may be proportionately greater than in the mucosa. Sometimes the infiltrate contains eosinophils or mast cells, which can be numerous but are not distinctive. Granulomas are a bonus, and great care needs to be taken to ensure that they are not at the site of ruptured crypts, and therefore mucin granulomas. They are not a sine qua non of CD in a patient who clearly has IBD.
 Figure 18-39. Stricture may develop in CD as shown here that may be multiple as seen from outside (A) and after opening the small bowel (B). (C) and (D) show strictures in terminal ileum and colon, respectively. |
 Figure 18-40. The key feature of CD on biopsies is patchiness or variability in inflammation, which is not always appreciated or present. A: Ileal biopsies showing parts of the biopsy with normal villous architecture and lack of mucosal inflammation. On closer look (B) one can easily see area with increased mucosal inflammation, loss of villi and crypt distortion. |
 Figure 18-40. (Continued) C: Colonic biopsies in a patient with CD showing three fragments of which one shows diffuse mucosal inflammatory changes while the others are either normal or have focal inflammation. D: Higher magnification of the inflamed fragment to show increased lamina propria inflammatory infiltrate, subtle crypt distortion, and Paneth cell metaplasia. Note that in comparison to UC, the goblet cells are relatively well preserved (less mucin depletion) and crypt architectural distortion is milder. |
 Figure 18-41. Another feature of CD that is strongly suggestive, but not diagnostic is pseudopyloric metaplasia in small-bowel biopsies which can be fairly easy to recognize (A) or be very focal and easy to miss as seen on edge of this biopsy fragment (arrow) (B). Both these examples also show extensive architectural abnormalities of the villi of the ileum, ulcerations, and intense lamina propria inflammation. |
The site of the biopsy can also be very useful. These changes in terminal ileal biopsies, with virtually normal mucosa on the distal side of the ileocecal valve are highly suggestive of CD, especially in Westernized societies, and where chronic infections such as tuberculosis are relatively rare; indeed, other than ensuring that the patient is not taking NSAIDs, the changes are virtually diagnostic. Additional features of previous episodes of ulceration in terminal ileal biopsies (architectural distortion, pseudopyloric metaplasia [Fig. 18-41A,B],
thickened duplicated muscularis mucosae) support this diagnosis, the problem being that virtually everything in the list of etiologies that can also be recurrent is in the differential diagnosis.
thickened duplicated muscularis mucosae) support this diagnosis, the problem being that virtually everything in the list of etiologies that can also be recurrent is in the differential diagnosis.
Interpreting ileal biopsies with “chronic active inflammation” with or without erosions is challenging. The easy way out is to report them descriptively without interpretation or a differential diagnosis. The addition of “clinical correlation is required” (the job of the clinician anyway) in the report, just throws it back in their court, although if there is no history or information provided one’s options may be limited. However if findings are isolated to the terminal ileum, a standard comment such as “In our society by far the most common cause of these changes is CD, especially if chronic infections and NSAID-associated disease can be excluded” is helpful and gives the clinician permission to treat, if needed. There is a degree of paranoia about what NSAIDs can or cannot do in the terminal ileum, and data is severely limited. However, until better data are available this remains a consideration.
The edges of erosions and ulcers can be seen in biopsies. Fissuring ulcers that are more typical of CD are difficult to appreciate on biopsies. Aphthous ulcers may be seen. These are thought to represent the earliest manifestation of CD. Often, it is represented by focal loss of epithelium overlying a lymphoid follicle with minimal exudate or neutrophilic infiltrate. This finding can be subtle and easily overlooked (Fig. 18-42A,B). These ulcers were thought to have a predilection for the epithelium overlying the lymphoid follicles, although in practice they occur anywhere. These are by no means specific for CD (Table 18-11) (Fig. 18-35). Some studies suggest biopsies of aphthous ulcers seen endoscopically are more likely to show the characteristic granulomas, which can be diagnostically very helpful.384, 385, 386 The aphthous lesions result from patchy epithelial necrosis or mucosal microulcerations (loss of 1-6 cells),387, 388 and in the small intestine may be over the dome area overlying a mucosal lymphoid follicle with loss of M cells.389, 390 Ulcers may also result with neutrophils streaming into the bowel lumen coming from the base of crypts, leading in a later phase to mountain peak ulcers, or damage to small capillaries with subsequent loss of surface epithelial cells (summit lesion) (Fig. 18-43A,B).271, 391
 Figure 18-42. A: Gross aspect of aphthoid ulcers in CD. They appear as small, round mucosal depression with yellow or grey base surrounded by a reddish halo in an otherwise normal or edematous mucosa. B: Histology of aphthoid ulcers showing the classical site of the ulcer over a lymphoid nodule developing in the M-cell region of the large bowel. These were incidental histologic findings in a mucosa with occasional red, pinpoint-like areas. Note the lack of involvement of crypts immediately adjacent to the ulcer, which appear normal. |
Active disease is manifested by the presence of neutrophils, and typically these also seem to have a patchy or focal distribution, such that apparently single crypts, or even segments of crypts, may be acutely inflamed while their neighbors are apparently unaffected.392 This so-called focal active colitis in a background of patchy chronic inflammation is highly suggestive of CD, but it must be distinguished from “isolated” focal active colitis (when there is no background chronic inflammation), which can be found in
a number of other conditions.393, 394 Migration of neutrophils into crypt lumina may result in crypt abscess formation, and rupture of inflamed crypts releases mucus into the surrounding lamina propria. As described above in the section on UC, this can sometimes result in pericryptal aggregates of macrophages (crypt rupture granulomas or mucin granulomas), and it is important to recognize these for what they are and not attribute to them the same diagnostic relevance as “proper” granulomas (Fig. 18-18A,B).284, 395, 396 While these may be more common in CD, they are also seen in genuine UC or active colitis from any other cause.20, 397 Its diagnostic value is therefore limited.
a number of other conditions.393, 394 Migration of neutrophils into crypt lumina may result in crypt abscess formation, and rupture of inflamed crypts releases mucus into the surrounding lamina propria. As described above in the section on UC, this can sometimes result in pericryptal aggregates of macrophages (crypt rupture granulomas or mucin granulomas), and it is important to recognize these for what they are and not attribute to them the same diagnostic relevance as “proper” granulomas (Fig. 18-18A,B).284, 395, 396 While these may be more common in CD, they are also seen in genuine UC or active colitis from any other cause.20, 397 Its diagnostic value is therefore limited.
Table 18-11 Common Causes of Endoscopic Aphthoid Ulcers | ||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||
 Figure 18-43. A: Early lesions in CD can appear as mucosal ulcers with a stream of leucocytes toward the lumen. They have been called “mountain peak ulcers.” Although this lesion is suggestive of a diagnosis of CD, it is not specific. B: Neutrophils infiltrating into the bases of multiple crypts (cryptitis) as well as lamina propria as seen here represent active disease. |
In some biopsies of Crohn’s colitis the acute inflammation is more diffuse, especially when the disease is highly active with ulceration, and in these cases the pattern can occasionally be indistinguishable from UC. Nevertheless, the inflammation often fails to induce marked mucous (goblet cell) depletion and the crypt architecture is often surprisingly preserved (Fig. 18-40D). In long-standing cases the epithelium may show metaplasia, with the appearance of either Paneth cells or pseudopyloric metaplasia (ulcer-associated cell lineage) (Fig. 18-36),398 although the latter is less common in colonic CD than in the small bowel.399
GRANULOMAS. The finding of granulomas is a key histologic feature of CD400 (Fig. 18-44). It should however be remembered that granulomas can occur in other conditions (Table 18-12), especially at the site of ruptured crypts, in infectious diseases such as tuberculosis, but also in Campylobacter and Yersinia colitis, in the rectum Chlamydia, and even occasionally in drug-induced colitis.401 A granuloma is defined as a collection of histiocytes/macrophage cells and often contains other inflammatory cells.402 They are arranged in clusters. Because of the epithelial cell-like morphology of the histiocytes, they are called epithelioid cells, which are less phagocytic in nature but release more cytokines. According to some authors, a genuine granuloma in CD contains five or more epithelioid cells, with or without giant cells. Sometimes, a small granuloma is referred to as “microgranuloma” (Fig. 18-44F). The number of histiocytes in the microgranuloma is fewer than in the granuloma (7-18 vs. 25-90). Granulomas can be closely packed together and have a sarcoid-like
aspect, but a “loose” expanded form of granuloma is more common in CD. Granulomas can be detected in otherwise endoscopically normal-appearing mucosa or in inflamed tissues. Some cases of CD tend to show scattered mucosal granulomas without a background of mucosal inflammation and IBD-like changes, and this always raises a differential diagnosis of infections and sarcoidosis. Central necrosis and caseation are unusual and should raise the suspicion of tuberculosis. Giant cells may contain calcified conchoid bodies.384 Associated inflammatory cells are lymphocytes (usually CD4+ T cells and the epithelioid cells may express MHC class II antigens).403
aspect, but a “loose” expanded form of granuloma is more common in CD. Granulomas can be detected in otherwise endoscopically normal-appearing mucosa or in inflamed tissues. Some cases of CD tend to show scattered mucosal granulomas without a background of mucosal inflammation and IBD-like changes, and this always raises a differential diagnosis of infections and sarcoidosis. Central necrosis and caseation are unusual and should raise the suspicion of tuberculosis. Giant cells may contain calcified conchoid bodies.384 Associated inflammatory cells are lymphocytes (usually CD4+ T cells and the epithelioid cells may express MHC class II antigens).403
 Figure 18-44. Another key feature of CD on biopsies is the presence of granulomas which can be solid and well formed, and easily recognized even on low power (A-C), or they can be subtle (arrows) (D,E). Sometimes, only loose aggregates of few epithelioid histiocytes (arrows) may be seen (microgranuloma) (F), which are not really pathognomonic for a diagnosis of CD. |
 Figure 18-44. (Continued) G-I: Tangentially sectioned crypts can be confused with granulomas as shown here and only become evident on serial sectioning. |
The reported frequency of finding granulomas in CD varies between 15 and 85% for surgical resections and between 3% and 56% for biopsy, but in our experience is rarely higher than 50% in untreated CD in biopsies.387, 404, 405, 406, 407, 408, 409, 410, 411 The results depend highly on the extent of tissue sampling (number of biopsies, number of sections examined, endoscopic or surgical samples). In pediatric patients, the incidence of granulomas is twofold higher compared to adults, but it decreases after the second year of the disease. The lowest incidence is seen in surgical resection in older patients.407, 408, 409, 410, 411, 412 In general, granulomas are more common in the distal colon and rectum.400 The number of colonoscopic examinations and extent of sampling can influence the frequency of detection of granulomas.386 A granuloma was found in 23% and 47% of the patients in whom one (2.5 ± 1.4 biopsy samples) and four colonoscopies were performed (8.0 ± 1.8 samples), respectively. The practice of taking more than one biopsy during colonoscopy increases the yield clearly. For one to six biopsy samples, the frequency varies between 11% and 47%. Six biopsy samples seems an optimal number. When multiple serial sections are examined in the pathology laboratory the frequency of detection further increases to 50%.386
While the diagnostic value of the granuloma in CD is generally accepted, its clinical and prognostic significance remains unclear. Several studies have examined the relation between the presence of granulomas and prognosis with conflicting results. In 8 of 14 studies, the presence of granulomas had no influence upon the outcome. In three series the recurrence of the disease was decreased in the presence of granulomas, while in three other series recurrence rate was increased.384, 410, 413, 414, 415 The presence of granulomas has been associated with the need of surgical resection, but not immunosuppressive therapy.404 In a meta-analysis, granulomatous CD appears to be associated with a higher number of recurrences and reoperations, and a shorter time to recurrence and reoperation compared to nongranulomatous CD.342 Although granulomas occur more often in young patients, they are not related to polymorphisms of the CARD15/NOD2 and TLR 4 genes.405, 416 The R702W polymorphism of the CARD15/NOD2 gene has been found to be associated with the stricturing phenotype of CD and the presence of intestinal granulomas.417
Resection Specimens In the majority of resection specimens, the most striking histologic feature is ulceration that characteristically takes the form of deep, knife-like fissures that are lined by ulcer slough
and surrounded by inflammatory granulation tissue (Fig. 18-45A). Fissuring ulcers extend into and often through the colonic wall to form sinuses or fistulae, and may terminate in an extramural abscess, or communicate with other fissuring ulcers extending laterally to produce a complex network of sinuses and fistulae (Fig. 18-45B-D). In other cases, there may be more widespread mucosal ulceration with large, deep, but still discrete, mucosal defects with overhanging edematous edges. Despite the severity of the ulceration, it is often remarkable that in CD, the adjacent mucosa within a few millimeters may be virtually normal. Ulcer healing may lead to submucosal entrapment or misplacement of epithelium which may form mucusfilled cysts (colitis cystica profunda). The other mucosal inflammatory changes are the same as described in the previous section, except that the patchiness or skip lesions are better appreciated in resection.
and surrounded by inflammatory granulation tissue (Fig. 18-45A). Fissuring ulcers extend into and often through the colonic wall to form sinuses or fistulae, and may terminate in an extramural abscess, or communicate with other fissuring ulcers extending laterally to produce a complex network of sinuses and fistulae (Fig. 18-45B-D). In other cases, there may be more widespread mucosal ulceration with large, deep, but still discrete, mucosal defects with overhanging edematous edges. Despite the severity of the ulceration, it is often remarkable that in CD, the adjacent mucosa within a few millimeters may be virtually normal. Ulcer healing may lead to submucosal entrapment or misplacement of epithelium which may form mucusfilled cysts (colitis cystica profunda). The other mucosal inflammatory changes are the same as described in the previous section, except that the patchiness or skip lesions are better appreciated in resection.
Table 18-12 Granuloma-associated Diseases in the Gastrointestinal Tract | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Transmural inflammation is the histologic hallmark of CD, and is manifested as round or ovoid lymphoid aggregates that may measure up to a few millimeters in diameter. These may be found in all layers of the bowel wall, but they are most obvious in the submucosa and when they line up along the outer aspect of the muscularis propria to form a “Crohn’s rosary” (Figs. 18-46 and 18-47).271 When marked, these changes are so apparent that they can be seen macroscopically looking at the slide.
Granulomas develop in all layers of the intestines from the mucosa to the serosa but are most frequent in the submucosa. They are also common in draining lymph nodes being present in approximately 20% to 50% of the patients (Fig. 18-48). Rarely does the granulomatous inflammation affect extraintestinal sites such as the skin, liver, lungs, eyes, and ovaries, and has been reported in the literature with the unfortunate designation of “metastatic Crohn’s disease.”418, 419 Associated with the transmural inflammation is gross bowel wall thickening, involving all the layers, with edema and fibrosis being especially prominent in the submucosa. Lymphangiectasia is also a common feature, which is best appreciated in the submucosa and subserosa (Fig. 18-49).
Connective tissue changes of CD affect all layers of the bowel wall379, 420, 421 and include thickening and disruption and duplication of the muscularis mucosae, fibrosis, focal muscularization of the submucosa, fibrous scarring of the muscularis propria, and marked neuronal hyperplasia of the intramural and extramural nerve fibers (Fig. 18-50A).420, 421 Perineural chronic inflammation may be seen in the submucosal (Meissner’s) plexus, as well as in the myenteric (Auerbach’s) plexus (Fig. 18-50B,C). The presence of the latter at resected margins is associated with an increased risk of disease recurrence.422, 423
In patients who have received anti-TNF-α therapy, we have become aware that while parts of the resection may show typical histology, there are often areas with dense hypocellular hyalinized submucosal fibrosis relatively devoid of inflammation (Fig. 18-51). Granulomas are also rare. We suspect that these are the results of therapy.424 This is in contrast to most other forms of therapy in CD that may improve symptoms, but appear to do little to the pathology.
Patients are described in whom the histologic features of CD are limited to the mucosa and the submucosa.425 This has been termed “superficial Crohn’s disease” and is very rare in our experience. Amyloidosis (AA-type) is a rare complication of CD with a
frequency of <1%. It usually involves the kidney, but amyloidosis can also be seen in the bowel wall rarely426 (Fig. 18-52).
frequency of <1%. It usually involves the kidney, but amyloidosis can also be seen in the bowel wall rarely426 (Fig. 18-52).
 Figure 18-45. Deep fissuring ulcers (A,B) can become deeper to go through the entire thickness of the muscle layer and join each other to form a complex network of sinuses. Eventually they may perforate (C) and form fistulas between involved intestinal segments and adjacent organs or nearby uninvolved loops (D). |
Vascular Lesions Inflammatory cell infiltration of blood vessels and obliterative lesions have been observed in surgical samples from patients with CD in a number of studies (Fig. 18-53). The reported frequency varies between 3% and 85%. The lesions consist mainly of intimal fibrous thickening, hypertrophy of the tunica media, and adventitial fibrosis. The vascular changes are typically located in the outer part of the bowel wall. Changes of the intima and media and inflammatory cells infiltrating the adventitia of arteries were noted in 16% of 56 cases.427 Granulomas are commonly found alongside blood vessels, and especially adjacent to lymphatics, and sometimes there may be a full-blown granulomatous lymphangitis, phlebitis, or even arteritis.428 Granulomatous involvement of serosal and mesenteric arteries and veins was seen in 3 cases out of a series of 100.429 In another study of 70 patients, chronic inflammation, acute and chronic inflammation, obliterative changes, or granulomatous inflammation was found in more than 20% of the patients.430 In prospective studies of surgical specimens prepared by heparin-saline perfusion, occlusive fibrinoid
lesions of arteries supplying areas of intestine affected by CD were demonstrated. Granulomas were identified in the wall and lumen of the blood vessels and the vascular lesions were present in all 15 patients examined. Granulomatous angiitis was later confirmed by the same authors in 24 consecutive surgical samples. Based on these observations, multifocal GI infarction was proposed as the pathogenetic mechanism for CD but this hypothesis has not been confirmed.291, 431 While the nature and the exact frequency of granulomatous
vasculitis remain to be determined, their occurrence and diagnostic significance are unquestionable. Occasionally, giant cell arteritis has also been observed in CD and rare cases of concomitant occurrence of Takayasu’s arteritis and CD have been reported.432
lesions of arteries supplying areas of intestine affected by CD were demonstrated. Granulomas were identified in the wall and lumen of the blood vessels and the vascular lesions were present in all 15 patients examined. Granulomatous angiitis was later confirmed by the same authors in 24 consecutive surgical samples. Based on these observations, multifocal GI infarction was proposed as the pathogenetic mechanism for CD but this hypothesis has not been confirmed.291, 431 While the nature and the exact frequency of granulomatous
vasculitis remain to be determined, their occurrence and diagnostic significance are unquestionable. Occasionally, giant cell arteritis has also been observed in CD and rare cases of concomitant occurrence of Takayasu’s arteritis and CD have been reported.432
 Figure 18-46. CD with transmural lymphoid infiltration. A,B: This is usually found only in areas of gross macroscopic disease as shown here with ulcerating small-bowel disease with focal ulceration. Particularly in the submucosa, there are numerous aggregates of lymphocytes without germinal centers, primarily expanding the submucosa, although smaller aggregates are present immediately beneath the muscularis propria in the subserosa, creating the full picture of transmural inflammation. Even at this magnification, this appearance is virtually diagnostic of CD. C: Small-bowel disease with lymphoid aggregates, but here forming three rosary beadlike rows: one in the submucosa, a middle row in the region of the myenteric plexus, and a third in the subserosa. D: As seen here, sometimes prominent lymphoid aggregates are present in the submucosa with relative sparing of the muscularis propria. The pattern of inflammation can be extremely variable. E: Uncommonly, the most extensive disease is limited to the submucosa. This is found particularly in resections of defunctioned bowel with CD. F: Appendiceal involvement by CD as shown here on low magnification showing transmural lymphoid aggregates. The mucosal changes tend to be similar to colonic disease. |
 Figure 18-47. Granulomas in resection specimens are seen most commonly in the submucosa as shown in a resection of small bowel at low magnification (A) and higher magnification (B). They can also be seen in the muscularis propria (C), or rarely as small whitish subserosal nodules (miliary CD) (D). The histology of the subserosal nodules shows granulomas. |
 Figure 18-48. A: Granulomas can appear in mesenteric lymph nodes. B: Higher magnification. |
 Figure 18-49. One of the more common histologic features is the occurrence of dilated lymphatics or lymphangiectasias in the mucosa (A,B) but also in the submucosa or rarely in subserosa. |
 Figure 18-50. A: Abnormalities of the intrinsic enteric nervous system are common and characteristic. They include an increase in size of submucosal nerve fibers, sometimes called neuromatous hyperplasia and as seen here can resemble a traumatic neuroma. They are associated with granulomatous inflammation (B) or show infiltration by chronic inflammatory cells (C). |
 Figure 18-51. Ileal resection in a patient treated with a TNF-α antagonist. Note the dense fibromuscular changes occupying the entire submucosa, the virtual absence of the usual submucosa, myenteric plexus, and subserosal lymphoid hyperplasia, with only vestigial remnants and minimal inflammation in the lamina propria. |
Evaluation of Resection Margins in Crohn’s Disease Intraoperative histologic examination of the resection margin is of no practical value, since studies have shown that extended resections aiming at disease-free margins have little effect on the recurrence rate, although the disease tends to recur more quickly if the margins are involved grossly.433, 434 Surgeons may typically try to resect all overtly evident disease at the resection margin if possible, because of the view that the anastomosis may be healthier if it is not made across overtly inflamed bowel. However, aphthoid ulcers may be extremely extensive so that just being clear of overt gross disease appears to be the most practical approach. There is no indication to do quick (frozen) sections on resected margins.
Related posts:
Disorders of Endocrine Cells
Gastrointestinal Manifestations of Extraintestinal Disorders and Systemic Disease
Mesenchymal Tumors
Inflammatory Disorders of the Esophagus: Reflux and Nonreflux Types
Enteric Infections and Associated Diseases
Stomach and Proximal Duodenum: Inflammatory and Miscellaneous Disorders
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
