Ulcerative Colitis

Crohn Disease

Endoscopic Differentiation of UC and CD

Because of the challenges in differentiating UC and CD, IBD-unclassified (IBDU) accounts for around 4% to 6% of IBD diagnosed and seems to be more common in children compared with adults. The term indeterminate colitis is reserved for colectomy specimens where a diagnosis of CD or UC cannot be made based on the histology of the resected specimen. There seems to be significant interobserver variation in the classification of IBD. In a study comparing the histologic diagnosis of colonic IBD made by general pathologist versus that made by specialist gastrointestinal pathologists, 43% of cases initially diagnosed as UC were changed to CD or IC, whereas 17% initially diagnosed as CD were changed to UC or IC. Odze listed several reasons that lead to the diagnosis of IBDU, including failure to recognize unusual variants of UC or CD, insufficient clinical or radiologic information, fulminant colitis, presence of secondary disease such as pseudomembranous colitis, and failure to use hard criteria for CD.

Endoscopic differentiation of CD from UC can be difficult, particularly in the setting of pancolitis. The combination of clinical history, serologic markers, and radiologic imaging may aid the diagnosis. Nevertheless, as many as 12% of patients who underwent surgery following initial diagnosis of either UC or IBDU were subsequently diagnosed as having CD. Initial esophagogastroduodenoscopy (EGD) may be helpful in locating upper gastrointestinal (GI) disease, thereby making CD the more likely diagnosis. Lemberg and colleagues described 25 children with IBDU pancolitis who were eventually diagnosed with CD following EGD. In severe colitis, the most useful discriminatory features for CD were the presence of discontinuous inflammation, anal lesions, and cobblestoning. Erosions and microulcerations within a granular mucosa were believed to be specific for UC.

The terminal ileum is characteristically not involved in UC. However, up to 10% of patients with active pancolitis may develop “backwash ileitis,” which can extend several centimeters into the terminal ileum. The phenomenon has also been described in patients with subtotal or left-sided colitis with only mild disease activity. Features that favor CD rather than UC with associated-backwash ileitis are extensive length of small bowel disease, jejunitis, proximal ileitis separated by skip regions of uninvolved cecum, transmural ileal inflammation with granulomas and neural hyperplasia, greater inflammatory activity in the terminal ileal biopsies, and mucous gland metaplasia of the ileal mucosa. Recognition that CD tends to be transmural and that UC is a superficial mucosal inflammatory process has led to the use of endoscopic ultrasound (EUS) in the differentiation of IBD. However, results seem to be disappointing. EUS has only been effective in the evaluation of perirectal and perianal complications of CD. For more information please refer to discussion elsewhere in this issue on examination under anesthesia, EUS, and MRI evaluation of perianal disease. In UC, EUS can also be used to evaluate the biliary tree. In one trial, a thickened common bile duct greater than 1.5 mm found by transduodenal EUS was found in patients with primary sclerosing cholangitis (PSC) but not those with uncomplicated IBD or choledocholithiasis.

Infection

In a prospective study of patients presenting with acute bloody diarrhea with suspected IBD, up to one-third were found to have an infectious cause. Positive stool cultures are often indicative of infectious colitis. However, positive microbiologic results have been detected in about 20% of patients who present with IBD for the first time, an indication that infection can precipitate a flare-up in latent IBD . Infections known to mimic the endoscopic and histologic appearance of IBD include salmonellosis, shigellosis, campylobacteriosis, tuberculosis (TB), Escherichia coli 0157:H7 infection, yersiniosis , Clostridium difficile infection, gonorrhea, Klebsiella infection, chlamydiosis, syphilis, schistosomiasis, amebiasis, herpes simplex, cytomegalovirus (CMV) and certain fungi. Endoscopically, infectious colitis often has patchy inflammation unlike the continuous inflammation of UC. The discrete ulcers of CD are less common in infection but can occur with TB, CMV, and yersiniosis. Also, rectal ulcers and proctitis can be seen with gonorrhea and syphilis. Isolated ileitis, one of the many presentations of CD, can occur with Yersinia and Salmonella infection and TB. Histologic features that favor IBD rather than acute self-limited colitis (ASLC) are distorted crypt architecture, increased cellularity of the lamina propria, a villous surface, epithelioid granuloma, crypt atrophy, basal lymphoid aggregates, and one or more basally located giant cells. Granulomas can appear in ASLC, but these tend to be poorly formed and are associated with crypt abscesses. The use of EUS in differentiating ASLC from IBD is limited. A small study demonstrated that pathologic perirectal lymph nodes on EUS have also been linked with the development of ulcerative colitis in patients initially thought to have ASLC. In selected cases, EUS of the sigmoid colon helps differentiate between infectious colitis and CD. Patients with active CD have increased submucosal thickening, decreased mucosal thickening, and hypoechoic paracolonic lymph nodes compared with patients with unspecific infectious colitis. However, this pattern was not observed in patients with acute UC compared with those with unspecific infectious colitis.

Clinically, ASLC secondary to infection usually presents with an acute transient diarrheal illness; the short, self-limited course differentiates the condition from IBD. However, some infections such as TB run a more protracted and insidious course. Differentiation of CD from intestinal TB often poses a diagnostic challenge. Endoscopically, two factors, area of distribution and endoscopic appearance, can help distinguish between CD and intestinal TB. In terms of distribution, whereas CD and intestinal TB both seem to affect the ileocecal valve in about 60% to 70% of patients, terminal ileal involvement predominates over cecal involvement in CD, and cecal involvement predominates over terminal ileal involvement in intestinal TB. Also, a fixed-open ileocecal valve was believed to be suggestive of TB over CD. Makharia and colleagues found no relevant difference in the involvement of the stomach and duodenum but noted that involvement of the rectum, sigmoid colon, descending colon, ascending colon, and jejunum supported the diagnosis of CD. A small prospective study from South Korea formulated a scoring system to aid the differentiation between CD and intestinal TB based on eight endoscopic features. Anorectal lesions, longitudinal ulcers, aphthous ulcers, and cobblestoning supported the diagnosis of CD, whereas transverse ulcers, patulous ileocecal valve, scars, and involvement of fewer than four segments favored TB. The classical features of caseating granulomas and AFB have been found in about 30% of cases of intestinal TB. Histologic features that suggest TB include confluent caseating granulomas, bands of epithelioid histiocytes lining ulcers, granulomas larger than 400 μm in diameter, 5 or more granulomas in biopsies from one segment, granulomas in the submucosa or in granulation tissue, and disproportionate submucosal inflammation. In contrast, features that favor the diagnosis of CD include infrequent, small granulomas that are poorly organized and isolated and architectural distortion distant from granulomatous inflammation.

Ischemia

Ischemic colitis typically affects older patients with cardiovascular risk factors. However, in younger patients, risk factors such as hypercoagulability, vasculitis, aortic surgery, infection, cocaine use, long-distance running, bowel preparation, and certain medications (eg, estrogen in women, sumatriptan and other selective serotonin agonist in patients with migraines) can all lead to bowel ischemia. Endoscopically, ischemic colitis often spares the rectum because dual blood supply from the mesenteric and iliac arteries makes it resistant to ischemia. Other endoscopic features include loss of vascularity, erythema, friability, granularity, longitudinal ulceration, serpiginous ulceration, confluent ulceration, and submucosal hemorrhage. In chronic ischemia, ulceration with granulation tissue and pseudopolyps have also been described. Unlike IBD, ischemic colitis tends to occur in segments of colon, most commonly affecting the “watershed” regions at the border of the area supplied by the superior and inferior mesenteric arteries. Pancolonic involvement is rare, occurring in less than 10% of cases. Histologically, iron-laden macrophages and submucosal fibrosis are characteristic of ischemic injury.

Drug-Induced Colitis

Diverse drugs have been associated with colitis, with NSAIDs being the most common. Endoscopically, features associated with NSAID-induced colitis include ulcerations, inflammation, erosions, stricture formation, and even neoplastic-like masses. Flat ulcers and the characteristic diaphragmlike stricture appear more commonly in the right colon. Colonic biopsies are often indistinguishable from ischemic colitis and demonstrate patchy erosion, lamina propria fibrosis, and reactive epithelial change. In upper GI biopsies, CD can mimic chemical gastropathy secondary to NSAIDs, particularly if active inflammation, erosions, and granulomas are absent. The paucity of neutrophils at the base of an ulcer or erosion is typical of an NSAID lesion compared with the focal collections of numerous neutrophils that are typical of CD. Follow-up endoscopies can be carried out 4 to 8 weeks following cessation of the offending NSAID to check for endoscopic resolution, which can help confirm the cause.

Radiation

Radiation effects are related to the total dose of radiation received and frequency of radiation, as well as the total volume of tissue irradiated. Chronic radiation injury can occur many years after initial exposure and in its severe form can cause narrowing of the bowel lumen, leading to obstruction. Acute radiation proctitis seems to peak 2 weeks after the onset of radiotherapy and may stabilize or regress following completion of treatment. Endoscopic features in the acute setting include friability, granularity, pallor erythema, and prominent submucosal telangiectasias. Histologic findings at this time will demonstrate epithelial meganucleosis, fibroblastic proliferation, and absence of mitotic activity. Conversely, chronic histologic changes include telangiectasis of capillaries, platelet thrombus formation, narrowing of the arterioles, lamina propria fibrosis, and crypt distortion. It has been suggested that chronic radiation proctitis is more likely in those with initial severe acute proctitis.

Infection

In a prospective study of patients presenting with acute bloody diarrhea with suspected IBD, up to one-third were found to have an infectious cause. Positive stool cultures are often indicative of infectious colitis. However, positive microbiologic results have been detected in about 20% of patients who present with IBD for the first time, an indication that infection can precipitate a flare-up in latent IBD . Infections known to mimic the endoscopic and histologic appearance of IBD include salmonellosis, shigellosis, campylobacteriosis, tuberculosis (TB), Escherichia coli 0157:H7 infection, yersiniosis , Clostridium difficile infection, gonorrhea, Klebsiella infection, chlamydiosis, syphilis, schistosomiasis, amebiasis, herpes simplex, cytomegalovirus (CMV) and certain fungi. Endoscopically, infectious colitis often has patchy inflammation unlike the continuous inflammation of UC. The discrete ulcers of CD are less common in infection but can occur with TB, CMV, and yersiniosis. Also, rectal ulcers and proctitis can be seen with gonorrhea and syphilis. Isolated ileitis, one of the many presentations of CD, can occur with Yersinia and Salmonella infection and TB. Histologic features that favor IBD rather than acute self-limited colitis (ASLC) are distorted crypt architecture, increased cellularity of the lamina propria, a villous surface, epithelioid granuloma, crypt atrophy, basal lymphoid aggregates, and one or more basally located giant cells. Granulomas can appear in ASLC, but these tend to be poorly formed and are associated with crypt abscesses. The use of EUS in differentiating ASLC from IBD is limited. A small study demonstrated that pathologic perirectal lymph nodes on EUS have also been linked with the development of ulcerative colitis in patients initially thought to have ASLC. In selected cases, EUS of the sigmoid colon helps differentiate between infectious colitis and CD. Patients with active CD have increased submucosal thickening, decreased mucosal thickening, and hypoechoic paracolonic lymph nodes compared with patients with unspecific infectious colitis. However, this pattern was not observed in patients with acute UC compared with those with unspecific infectious colitis.

Clinically, ASLC secondary to infection usually presents with an acute transient diarrheal illness; the short, self-limited course differentiates the condition from IBD. However, some infections such as TB run a more protracted and insidious course. Differentiation of CD from intestinal TB often poses a diagnostic challenge. Endoscopically, two factors, area of distribution and endoscopic appearance, can help distinguish between CD and intestinal TB. In terms of distribution, whereas CD and intestinal TB both seem to affect the ileocecal valve in about 60% to 70% of patients, terminal ileal involvement predominates over cecal involvement in CD, and cecal involvement predominates over terminal ileal involvement in intestinal TB. Also, a fixed-open ileocecal valve was believed to be suggestive of TB over CD. Makharia and colleagues found no relevant difference in the involvement of the stomach and duodenum but noted that involvement of the rectum, sigmoid colon, descending colon, ascending colon, and jejunum supported the diagnosis of CD. A small prospective study from South Korea formulated a scoring system to aid the differentiation between CD and intestinal TB based on eight endoscopic features. Anorectal lesions, longitudinal ulcers, aphthous ulcers, and cobblestoning supported the diagnosis of CD, whereas transverse ulcers, patulous ileocecal valve, scars, and involvement of fewer than four segments favored TB. The classical features of caseating granulomas and AFB have been found in about 30% of cases of intestinal TB. Histologic features that suggest TB include confluent caseating granulomas, bands of epithelioid histiocytes lining ulcers, granulomas larger than 400 μm in diameter, 5 or more granulomas in biopsies from one segment, granulomas in the submucosa or in granulation tissue, and disproportionate submucosal inflammation. In contrast, features that favor the diagnosis of CD include infrequent, small granulomas that are poorly organized and isolated and architectural distortion distant from granulomatous inflammation.

Ischemia

Ischemic colitis typically affects older patients with cardiovascular risk factors. However, in younger patients, risk factors such as hypercoagulability, vasculitis, aortic surgery, infection, cocaine use, long-distance running, bowel preparation, and certain medications (eg, estrogen in women, sumatriptan and other selective serotonin agonist in patients with migraines) can all lead to bowel ischemia. Endoscopically, ischemic colitis often spares the rectum because dual blood supply from the mesenteric and iliac arteries makes it resistant to ischemia. Other endoscopic features include loss of vascularity, erythema, friability, granularity, longitudinal ulceration, serpiginous ulceration, confluent ulceration, and submucosal hemorrhage. In chronic ischemia, ulceration with granulation tissue and pseudopolyps have also been described. Unlike IBD, ischemic colitis tends to occur in segments of colon, most commonly affecting the “watershed” regions at the border of the area supplied by the superior and inferior mesenteric arteries. Pancolonic involvement is rare, occurring in less than 10% of cases. Histologically, iron-laden macrophages and submucosal fibrosis are characteristic of ischemic injury.

Drug-Induced Colitis

Diverse drugs have been associated with colitis, with NSAIDs being the most common. Endoscopically, features associated with NSAID-induced colitis include ulcerations, inflammation, erosions, stricture formation, and even neoplastic-like masses. Flat ulcers and the characteristic diaphragmlike stricture appear more commonly in the right colon. Colonic biopsies are often indistinguishable from ischemic colitis and demonstrate patchy erosion, lamina propria fibrosis, and reactive epithelial change. In upper GI biopsies, CD can mimic chemical gastropathy secondary to NSAIDs, particularly if active inflammation, erosions, and granulomas are absent. The paucity of neutrophils at the base of an ulcer or erosion is typical of an NSAID lesion compared with the focal collections of numerous neutrophils that are typical of CD. Follow-up endoscopies can be carried out 4 to 8 weeks following cessation of the offending NSAID to check for endoscopic resolution, which can help confirm the cause.

Radiation

Radiation effects are related to the total dose of radiation received and frequency of radiation, as well as the total volume of tissue irradiated. Chronic radiation injury can occur many years after initial exposure and in its severe form can cause narrowing of the bowel lumen, leading to obstruction. Acute radiation proctitis seems to peak 2 weeks after the onset of radiotherapy and may stabilize or regress following completion of treatment. Endoscopic features in the acute setting include friability, granularity, pallor erythema, and prominent submucosal telangiectasias. Histologic findings at this time will demonstrate epithelial meganucleosis, fibroblastic proliferation, and absence of mitotic activity. Conversely, chronic histologic changes include telangiectasis of capillaries, platelet thrombus formation, narrowing of the arterioles, lamina propria fibrosis, and crypt distortion. It has been suggested that chronic radiation proctitis is more likely in those with initial severe acute proctitis.

Ulcerative Colitis

Clinical impression of UC severity was found to correlate poorly with endoscopic mucosal findings and histology, with physicians underestimating inflammatory activity in about a third of patients. Osada and colleagues indicated that clinical symptoms were more likely to reflect the activity of distal disease, whereas CRP and erythrocyte sedimentation rate (ESR) reflected the activity of more proximal disease. When determining disease activity and severity, sigmoidoscopy can be inadequate in evaluating UC. This finding was confirmed by Kato and colleagues, who showed that severe inflammatory activity was observed more frequently in patients who had maximum disease activity in the proximal colon compared with the rectum or sigmoid. Also, patients who are receiving rectal therapy may have minimal distal disease. Full colonoscopy is particularly important in the initial mapping of disease extent and severity as well as to investigate any discrepancy between clinical symptoms and rectosigmoid appearance. The extent of colonic involvement in UC is not static and can change over time. Up to a third of patients initially diagnosed with proctitis will have proximal progression of their disease. Although routine colonoscopy is not usually carried out in asymptomatic patients, patients with longstanding disease require reevaluation of their disease extent because this may influence their risk of developing colorectal cancer.

Multiple clinical and endoscopic scoring methods have been designed to classify the severity of UC. The simplest clinical scoring system, introduced in 1955, is based on stool frequency, presence of fever, tachycardia, anemia, and elevated ESR. Scoring systems that include endoscopic features include the Baron score, Mayo score, Powell-Tuck, Rachmilewitz endoscopic index, and the UC Disease Activity Index. Most endoscopic scoring systems are similar in their description of inactive, mild, moderate, and severe UC. Patients with inactive UC have normal mucosa with clearly visible vascular pattern. Mild disease usually consists of mild erythema, decreased vascular pattern, and mild friability. Moderate disease is defined as moderate friability with mild contact bleeding, no spontaneous bleeding, and granularity. Severe colitis is seen as spontaneous bleeding from the mucosa with pronounced mucosal damage and visible ulceration.

In a study by Kiesslich and colleagues, newer techniques such as chromoendoscopy (CE) and magnification endoscopy (ME) have been found to be more accurate in predicting mild disease (87%) compared with conventional endoscopy (54%), P = .0002. The same study demonstrated that ME and CE (84.5%) were superior to conventional endoscopy (37%) in predicting the extent of UC disease activity when correlated to histologic diagnosis, P <.0001. For more information on CE, please refer to discussion, “CE in IBD surveillance,” elsewhere in this issue.

Crohn Disease

The Crohn Disease Endoscopic Index of Severity, developed in the late 1980s, is based on the extent of disease in each of the five bowel segments (rectum sigmoid and descending, transverse, ascending and ileum) and type of lesion observed (superficial or deep ulceration, ulcerated and nonulcerated stenosis). However, poor correlations were found between clinical symptoms, endoscopic scoring, and treatment outcome. This scoring system was later simplified by Daperno and colleagues into the Simple Endoscopic Score for Crohn Disease (SES-CD) and evaluated ulcer size rather than depth of penetration. The SES-CD scoring system achieved better interobserver consistency and correlation with clinical symptoms as well as CRP. However, neither of these scores takes into account upper GI involvement. Following curative resection, Rutgeerts’ score is the most widely used scoring system for evaluating postsurgical recurrence. It takes into account the presence of aphthous ulcers, inflammation, involvement of the terminal ileum, the extent of colonic involvement, and stricturing disease.

Ulcerative Colitis

Endoscopic and especially histologic remission typically lag behind clinical response to treatment. Endoscopic remission has been shown to predict later symptomatic outcome and clinical remission. Early studies demonstrated that 40% of patients who achieve endoscopic remission after treatment remained symptom-free during a 1-year follow-up compared with 18% if endoscopic abnormalities persisted. This finding has lead to the use of endoscopic scoring, usually in combination with patient symptoms as a clinical end point in most pharmacologic trials evaluating treatment response. These trials include the ASCEND trials and a study for patients previously treated with 5-aminosalicilyc acid (5-ASAs) by Kamm and colleagues. More recently, the Acute Ulcerative Colitis Treatment (ACT1 and ACT2) study group evaluated the induction of mucosal healing by infliximab. It was demonstrated that complete mucosal healing or a Mayo score of 0 at week 8 had a 73.7% chance of achieving symptomatic remission at week 54 and predicted a lower need for colectomy. Other than the evaluation of response to medical therapy, endoscopy is an important modality to diagnose postoperative disease such as pouchitis. (See “Evaluating pouch problems” elsewhere in this issue.)

Crohn Disease

In CD, it has been shown that there can be poor correlation between clinical and endoscopic remission. The GETAID group demonstrated that high-dose steroid therapy failed to induce mucosal healing in 29% of patients who had achieved clinical remission. Also, endoscopic appearance actually worsened in 9% of patients treated with steroids despite symptomatic improvement. Endoscopic evidence of mucosal healing in CD also seems to have prognostic implications. Baert and colleagues demonstrated that 70.8% of patients with a SES-CD of 0 at 2 years were found to be in steroid-free remission at years 3 and 4. In a study by Bjorkesten and colleagues, mucosal healing at 3 months was successful in predicting mucosal healing at 12 months. Also, mucosal healing in CD has been shown to lead to decreased incidence of abdominal surgery and hospitalizations. Please refer to “Evaluation for Postoperative Recurrence of CD” elsewhere in this issue for the role of endoscopy in assessing postoperative CD patients.