Ulcerative Colitis Pathology

Ulcerative colitis (UC) is one of the idiopathic forms of inflammatory bowel disease (IBD), sharing this designation with Crohn disease (CD). Ulcerative colitis is a chronic, relapsing inflammatory disease of the colon, and affected patients may exhibit inflammation from cecum to rectum (see the following image). Symptoms may include abdominal pain, malnutrition, and diarrhea, often bloody.

Chronic architectural changes in ulcerative colitis (UC). One can see crypt branching and irregularity of size and shape, with an increase in chronic inflammatory cells in the lamina propria.

Patients with ulcerative colitis are at increased risk for dysplasia and adenocarcinoma of the colon,[1, 2] and they require lifelong, regular endoscopic surveillance. Approximately 25-30% of patients with this condition will require colectomy at some point in their lives if treatment with medications is not successful or if dysplasia, a precursor lesion to carcinoma, is found. Colectomy is considered curative for ulcerative colitis.[3]

A detailed explanation of treatment for ulcerative colitis is beyond the scope of this discussion. Briefly, the goals of treatment are to decrease symptoms and improve quality of life. Therapeutic interventions may include 1 or more medications, dietary modifications, and supportive care during disease flares. In addition, many affected patients (25-30%) will require surgery at some point in their lives.

See also the following:

• Crohn Disease Pathology

• Inflammatory Bowel Disease

• Complications of Inflammatory Bowel Disease

• Pediatric Crohn Disease

Ulcerative colitis (UC) has 2 incidence peaks, one in adolescents and young adults and the other in middle-aged men and women. Men and women are about equally affected. Overall, inflammatory bowel disease is more common in white individuals than in black or Asian Americans.[4]

Since World War II, the incidence of ulcerative colitis has increased. However, although it is beginning to level off in Western countries, this disease continues to increase in Eastern Europe, Asia, and developing countries. The estimated annual incidence of ulcerative colitis in North America and Europe ranges from 4 to 20 per 100,000 individuals. This condition is more prevalent in northern regions of the United States than in southern regions and more common in urban settings than in rural ones.

Smoking appears to confers protective effects: the incidence of inflammatory bowel disease is decreased in smokers, and when the disease does occur, it tends to be less severe.[3]

The cause of ulcerative colitis (UC) is currently under examination. Although no single definitive cause has been elucidated, many etiologies have been postulated and studied.

It is believed that the 2 idiopathic forms of inflammatory bowel disease (IBD), ulcerative colitis and Crohn disease (CD), develop secondary to complex interactions among genetic predispositions, environmental risk factors, and the immune system. Several genes likely play a role; their products, when combined with environmental factors and dysfunctional immunity, result in a disease spectrum with heterogeneous manifestations and many unique phenotypes.[5] (See Molecular/Genetics.)

One theory proposes that ulcerative colitis occurs as a direct result of an aberrant immune response to "gut luminal antigen" in a genetically susceptible host.[6]

In normal patients' colonic mucosa, phagocytic cells do not respond to the resident bacteria. In patients with inflammatory bowel disease, these phagocytes may begin to respond to gut flora, secreting proinflammatory cytokines that activate a certain population of T cells. This results in mucosal inflammation and damage.[7] In particular, there is speculation that interleukin (IL)–25, which controls certain aspects of T helper cell response, may be abnormal or deficient in some patients with inflammatory bowel disease. This provides the basis for exploring potential treatment options with exogenous IL-25.

Notably, the microbiotic flora in patients with inflammatory bowel disease is less diverse than that in unaffected individuals: healthy subjects have more gut commensals ("good bacteria"), whereas those with inflammatory bowel disease have more gut pathogens ("bad bacteria").[7] Although certain pathogenic microorganisms, including Pectinatus, Sutterella, and Fusobacterium, are frequently found in ulcerative colitis, no pathogenetic link with ulcerative colitis has been proved.[8]

Another hypothesis is that antibodies are formed against some bacteria and bacterial products in the gut and that these antibodies propagate the altered immune response present in inflammatory bowel disease.[9]

In summary, research and literature to date indicate that when a genetic predisposition is combined with certain environmental factors—both in the external and in the internal environment—and acted upon by an altered immune response, inflammatory bowel disease may result. No dietary or environmental changes have yet been identified that can prevent inflammatory bowel disease or definitively improve the course of the disease.

Ulcerative colitis (UC) is usually limited to the colon. Typically, inflammation begins at the rectum and continuously involves all or part of the colon. The transition to normal areas may be either abrupt or gradual, and it may occur anywhere from the rectum to the cecum. Some cases only involve the rectum—often called ulcerative proctitis. Less commonly, isolated or discontinuous proximal segments of disease may be identified at presentation—also called patch colitis. Similarly, the appendix may be involved in a discontinuous fashion, sparing all or part of the cecum.

Although most cases of ulcerative colitis have inflammation that is limited to the colon, some patients with severe disease may have an incompetent ileocecal valve that allows retrograde flow of colonic contents into the ileum. The resulting ileal inflammation is commonly referred to as backwash ileitis.[10] In addition, the appendix may be involved in a discontinuous fashion, sparing all or part of the cecum. In addition, there are a few reports of inflammatory alterations in the small intestine in some patients with ulcerative colitis.[11]

Ulcerative colitis is considered primarily a gastrointestinal (GI) disease, but it has important associations with diseases in other organ systems. About 2-7% of affected patients will also suffer from primary sclerosing cholangitis,[12] a rare liver disorder characterized by chronic inflammation and fibrosis of bile ducts leading to cholestasis and eventual cirrhosis. The clinical course is variable, and liver transplantation is offered to patients with end-stage disease.[13]

In addition, 13-23% of patients with ulcerative colitis suffer from peripheral arthritis, 1-2% develop ankylosing spondylitis, and 4-5% show skin manifestations, including pyoderma gangrenosum and erythema nodosum.

In one study, 7% of patients with ulcerative colitis had 1 of 6 major extraintestinal diseases, and 0.3% had multiple extraintestinal diseases. The 6 diseases studied were primary sclerosing cholangitis, ankylosing spondylitis, iritis, uveitis, pyoderma gangrenosum, and erythema nodosum.[12]

Ulcerative colitis (UC) is a chronic, episodic disease with relapsing symptomatic periods that often alternate with disease-free intervals. During flares, or periods of activity, patients may suffer from some or all of the following:

• Diarrhea, possibly with blood and/or mucus in the stool

• Nocturnal diarrhea

• Fecal incontinence

• Constipation/Obstipation

• Fecal urgency

• Tenesmus

• Abdominal pain

• Malnutrition

• Weight loss

• Nausea/vomiting

• Fever

• Fatigue

• Extraintestinal manifestations

A thorough clinical history must be taken to rule out other possible causes of these symptoms. It is also important to inquire into the patient's travel history, medication usage, smoking history, and family history and to find out whether the patient has other medical problems.

In addition, a complete physical examination should be performed. Patients should be carefully assessed for evidence of anemia, malnutrition, and distress. Abdominal examination may reveal a mass, distention, pain, tenderness, or altered bowel sounds.

Entities that may be in the clinical differential diagnosis of ulcerative colitis (UC) include infectious diarrhea, medication effect (eg, drug-induced injury, especially injury due to nonsteroid anti-inflammatory agents [NSAIDs]), ischemia, diversion colitis, and ischemic colitis.

Other conditions that should be considered when evaluating a patient with suspected ulcerative colitis include the following:

• Behçet Disease

• Clostridium Difficile Colitis

• Collagenous and Lymphocytic Colitis

• Crohn Disease

• Cytomegalovirus Colitis

• Diverticulitis

• Irritable Bowel Syndrome

The diagnosis of ulcerative colitis (UC) is made on the basis of clinical information and hematoxylin and eosin (H&E) light microscopy. Multiple laboratory tests may be performed if clinically indicated, including, but not limited to, stool examinations for ova and parasites, Clostridium difficile, and leukocytes; complete blood cell (CBC) count; erythrocyte sedimentation rate (ESR); electrolyte concentrations; iron studies; and liver enzyme assays.

Immunohistochemistry (IHC) does not play a role in the diagnosis of this disease, except to exclude certain viral infections, such as cytomegalovirus (CMV) infection, which may be associated with an acute flare of the disease or which may develop in the setting of immunosuppressive therapy.

Radiologic and endoscopic studies significantly augment clinical assessment of putative ulcerative colitis patients. These studies include abdominal plain films, computed tomography (CT) scanning, colonoscopy, and upper endoscopy.[3]

It is of the utmost importance for the clinician to be aware that without adequate clinical information, including endoscopic and imaging data, the pathologist is essentially blind. Correct diagnosis of inflammatory bowel disease (IBD) depends on clinical-pathologic correlation.

A study by Theede et al investigated the association between the level of fecal calprotectin and mucosal healing and clinical disease activity in patients with ulcerative colitis since it is difficult to determine mucosal healing, a goal of treatment, by clinical evaluation and endoscopy alone. The study reported that the level of fecal calprotectin identifies patients with ulcerative colitis who have endoscopic and histologic features of mucosal healing and correlates with endoscopic and histologic inflammatory activity.[14, 15, 16]

As noted earlier, ulcerative colitis (UC) may diffusely involve the entire colon, or it may be limited to the rectum (ulcerative proctitis).

Classic gross findings

In classic ulcerative colitis, the gross resection specimen usually shows a continuous pattern of changes, beginning with the rectum and progressing proximally, terminating anywhere in the colon up to the ileocecal valve. The mucosa may be hyperemic and nodular and may exhibit fine granularity (see the images below). Erosions and ulcerations may be seen. The ulcers are typically broad-based and may be linear or geographic; blood, pus, or mucus may overlie the mucosa. Generally, the colon is diffusely involved up to the point where the mucosa transitions to normal regions.

Note the darkened mucosa in this ulcerative colitis specimen. The specimen has been fixed in formalin, which causes hyperemic areas to appear brown-black.

Another gross specimen illustrating ulcerative colitis.

Nonclassic gross findings

On occasion, rectal sparing or patchy rectal inflammation may occur, suggesting that repair has occurred in these areas. The transition from involved to normal colon may be abrupt or gradual. In addition, although so-called skip lesions have traditionally been considered more typical of Crohn disease (CD) than of ulcerative colitis, patients with ulcerative colitis may have a cecal or appendiceal patch associated with colitis that is otherwise most severe in the distal left side.

It should be kept in mind that the pattern of involvement and the gross appearance may be altered by treatment. For instance, use of steroid enemas may lead to the spurious appearance of "rectal sparing."

Findings under colonoscopy

On colonoscopy, all of the features recognized in the mucosa of resection specimens may be seen; again, transitions to normal-appearing mucosa at the proximal extreme of disease may be gradual or abrupt.

Notably, mucosa that appears normal at the transition on either inspection of the resected specimen or colonoscopic examination may nonetheless exhibit varying degrees of active or chronic inflammatory injury on histologic analysis. This finding emphasizes that macroscopic disease requires significant mucosal injury and that reliance on macroscopic features for diagnosis thus may underrepresent the true distribution of disease.

This gross-microscopic disparity also informs the histologic evaluation of Crohn disease as well as quiescent ulcerative colitis. (So-called microscopic colitis, which includes collagenous colitis, lymphocytic colitis, and focal active colitis, owes its name to this phenomenon.)

Generally, the submucosa, muscularis propria, and serosa are normal, because ulcerative colitis is not a transmural process and remains limited to the mucosa; however, thickening and foreshortening may occur, as seen in the image below. An important clinical exception to this general rule arises in the setting of acute fulminant ulcerative colitis, in which transmural suppurative inflammation is commonly seen.

Bowel-wall thickening and foreshortening are apparent in this colectomy for ulcerative colitis. In addition, the mucosa is hyperemic, with focal nodularity and ulceration.

Because the risk of adenocarcinoma is increased in ulcerative colitis (see Dysplasia and Carcinoma in Ulcerative Colitis), one must conduct a thorough gross examination and sample all suspicious lesions, as well as submit sections of flat mucosa every 10 cm in the colon and every 5 cm in the rectum. If a stricture is found, malignancy is suggested, and microscopic study of these areas is indicated.[17]

An effective biopsy strategy is critical for correct diagnosis of ulcerative colitis (UC), including providing appropriate clinical details. Without clinical information, the pathologist may misinterpret biopsy or resection findings, particularly in patients who have received previous medical or surgical treatment. For instance, steroid enemas may eliminate the distal colonic and rectal inflammation so that these areas may appear normal in the specimen. Such patients may appear to have right-side disease that, in association with backwash ileitis, may lead to the incorrect diagnosis of Crohn disease.

Oral treatment with anti-inflammatory drugs, including corticosteroids, may also quiet the inflammation and give rise to areas where the colitis heals, altering the distribution pattern and creating discontinuous areas of involvement that mimic Crohn disease. Pediatric patients may also show no evidence of rectal involvement at their initial presentation with ulcerative colitis.[18]

Biopsy collection

Biopsy specimens from different locations should be separately submitted and designated. Sampling of both endoscopically abnormal and endoscopically normal areas should be submitted, and biopsies should include sampling of aphthous ulcers and polyps.

Ileoscopy with ileal biopsies is essential to differentiate ulcerative colitis from Crohn disease (CD). Sodium phosphate bowel preparations and nonsteroidal anti-inflammatory drugs (NSAIDs) should be avoided before the biopsies, as they may cause mucosal changes that confound interpretation.

In patients undergoing surveillance for dysplasia, it is recommended to take a minimum of 32 biopsies, sampling all 4 quadrants every 10 cm from cecum to rectum and every 5 cm in the rectum. If nodules are identified, they are removed and submitted in 1 container, with additional biopsies of the flat mucosa surrounding the polyp, to help differentiate a dysplasia-associated lesion or mass (DALM) from sporadic adenoma.[19]

Histologic changes

To make the diagnosis of inflammatory bowel disease (IBD), which includes both ulcerative colitis and Crohn disease, it is imperative to see both active/acute and chronic histologic alterations.

Chronic features include the following:

• Alteration of crypt architecture (see the images below): Branching crypts with marked deviation of the crypt axis from the perpendicular; variation in crypt size and/or shape; crypt dropout; foreshortening of crypts, with bases of crypts lifted off the muscularis mucosae

  Chronic architectural changes in ulcerative colitis (UC). One can see crypt branching and irregularity of size and shape, with an increase in chronic inflammatory cells in the lamina propria.

  Chronic architectural changes in ulcerative colitis. Note the trifid crypt.

• Increased chronic inflammatory cells in the lamina propria (see the following image): Lymphocytes; plasma cells, especially below the base of crypts (so-called basal plasmacytosis [see the third image below]; this feature may be accompanied by a more generalized mucosal plasmacytosis).

  Basal plasmacytosis. Plasma cells separate the crypt bases from the muscularis mucosae.

Active/acute features include the following:

• Neutrophilic inflammation, including neutrophilic infiltration of the surface epithelium and neutrophilic cryptitis (neutrophils infiltrating crypt epithelium), both evidence of damage to cellular junctions

• Crypt abscesses (see the images below): Neutrophils within crypt lumens

  In addition to basal plasmacytosis, crypt abscesses are visible, characterized by neutrophils and apoptotic debris in crypt lumens.

  Higher power view of the crypt abscesses from the previous image.

  High-power view of a crypt abscess shows the crypt to be dilated and filled with neutrophils and debris.

• Erosion or ulcer

Ulcerative colitis versus Crohn disease

To distinguish ulcerative colitis from Crohn disease, a number of factors are considered. Ulcerative colitis is typically a disease that is limited to the colon (with the exception of backwash ileitis). This condition generally involves the entire colon diffusely (with the exceptions noted earlier), including the rectum, and inflammation may become worse as it progresses from the right side of the colon to the left. Perianal disease is uncommon, and mesenteric adenopathy is only occasionally seen.

Despite its name, ulcerative colitis often does not present with macroscopic ulcers (although histologic evidence of erosions and ulcers is common in untreated disease), and fissuring disease is typically absent. In contrast, Crohn disease is overtly ulcerative, with characteristic fissures.

Unlike Crohn disease, ulcerative colitis does not exhibit transmural inflammation, and serositis is absent, except in the setting of fulminant colitis or toxic megacolon. Instead, the inflammation in ulcerative colitis is typically limited to the mucosa and submucosa. Granulomas are not a characteristic feature of this disease, although granulomas may be seen with mucous extravasation in sites of crypt abscess or crypt destruction. The muscularis mucosa is usually thickened, and mucosal atrophy may be seen.

Dysplasia

Ulcerative colitis patients are subject to a risk of dysplasia that increases with the duration of disease as well as an increased likelihood of adenocarcinoma in comparison with age-matched cohorts without inflammatory bowel disease. Histologic recognition of both dysplasia and carcinoma is thus of critical importance.

Dysplasia is defined as an unequivocal neoplastic alteration of colonic epithelium.[20] It may consist of architectural disorganization of colonic crypts, cytologic atypia (including nuclear hyperchromasia), pseudostratification, pleomorphism, clumped chromatin, and prominent nucleoli, with increased mitotic activity, particularly in the upper portion of the crypts (see the following images).

This is an example of low-grade glandular dysplasia in a patient with longstanding ulcerative colitis. Note the loss of mucin, nuclear hyperchromasia, and nuclear pseudostratification.

High-grade dysplasia in the same patient as the previous image. There is significant cytologic atypia, with rounding of the nuclei and a greater degree of pseudostratification.

Histologic section from another location in the same patient as the previous image. This field shows glands that are suspicious for invasive carcinoma.

In the classic paper that defined the basis of separation of biopsies that are reactive from those that are positive for dysplasia, Morson stated that in low-grade dysplasia, "[a]rbitrarily, the nuclei should be confined largely to the basal half of the cells...."[21, 22] With cytologic similarity to tubular adenomas of the colon, it is distinguished from high-grade dysplasia largely on differences in cytologic atypicality.

In high-grade dysplasia, nuclear stratification extends to the upper (luminal) aspect of the cells, and there is loss of nuclear polarity, including loss of the orientation of the long axis of the nucleus from perpendicular to the basement membrane, that may vary from cell to cell, accompanied by nuclear rounding.[20]

Accepted categories include indefinite for dysplasia, low-grade dysplasia, and high-grade dysplasia. It may be difficult to distinguish dysplasia from the reactive/reparative atypia that occurs after periods of disease activity. These diagnoses must be made with caution, because the implications for the patient are serious. Low-grade dysplasia warrants at least consideration for colectomy, whereas high-grade dysplasia or carcinoma is an indication for colectomy.[19]

In periods of quiescence, one may see completely unremarkable colonic mucosa. Alternatively, one may appreciate an overall decrease in the number of crypts, irregular spacing, and other chronic architectural changes. There may be fibrosis in the lamina propria, thickened muscularis mucosae, and Paneth cell metaplasia. Active inflammation may be seen in clinical quiescence, but significant active crypt injury is not expected.[20]

See Dysplasia and Carcinoma in Ulcerative Colitis.

Genetic factors have been strongly implicated in inflammatory bowel disease (IBD). Family studies have shown that the incidence of inflammatory bowel disease is higher in patients who have a family history than in those who do not,[23] and twin studies show a higher disease concordance in monozygotic (but not dizygotic) twins.[24, 25]

Most genetic associations identified thus far have been found with Crohn disease (CD); however, ulcerative colitis (UC) has recently begun to receive more attention, and considerable overlap with Crohn disease is being uncovered.[26]

Studies have found a strong association between the major histocompatibility complex (MHC) region and susceptibility to ulcerative colitis.[27] This region contains immunoregulatory genes, including those for human leukocyte antigen (HLA) and tumor necrosis factor (TNF)-alpha, which may have important treatment implications in the future. The HLA-DRB1*0103 locus has been implicated as a possible predictor of corticosteroid-dependent refractory disease in ulcerative colitis.

The IBD-5 locus on chromosome 5q1 has been associated with both Crohn disease and ulcerative colitis. A regulator of interleukin (IL)-23, IL-23R (found on chromosome 1), is also associated with Crohn disease and ulcerative colitis. IL-23 shares the p40 subunit with IL-12; both interleukin moieties affect the differentiation of naive T cells into effector T helper type 1 cells (Th1) or Th17 cells. The latter are thought to play a role in the pathogenesis of ulcerative colitis, and medications directed against p40 are being explored as possible treatments for this disease.

About 3-5% of patients with ulcerative colitis (UC) will develop adenocarcinoma, and this outcome is thought to be the result of a progression of a precursor dysplastic lesion in most cases. Accordingly, screening endoscopy in ulcerative colitis patients focuses on dysplasia, a lesion that may be quite focal in its early presentation, despite what appears to be a clonal overgrowth (or "field defect") of the chromosomally unstable cell population that ultimately exhibits dysplastic histology.

Hence, patients with ulcerative colitis will require screening colonoscopies more often than do patients without inflammatory bowel disease (IBD). The risk of malignancy in ulcerative colitis is definitely higher than in patients unaffected by idiopathic inflammatory bowel disease, and it also appears to be higher than the risk in patients without inflammatory bowel disease who have had multiple adenomas, adenomas with high-grade dysplasia, or a previous colonic malignancy.

The American College of Gastroenterology (ACG) guidelines, updated in 2010, recommend annual or biannual surveillance colonoscopy after 8-10 years of colitis.[28] Patients with primary sclerosing cholangitis (PSC) are recognized to be at greater risk for colonic neoplasia than other ulcerative colitis patients are and thus should begin surveillance at the time of diagnosis of PSC and yearly thereafter.[4, 19, 28]

It is recommended that the endoscopist take 4 biopsies at each level every 10 cm in the colon and every 5 cm in the rectum, with additional sampling of any polypoid or raised lesions.

Sporadic adenomas versus DALMs

Patients with ulcerative colitis may also have sporadic adenomas, which must be distinguished from the dysplasia associated with ulcerative colitis (dysplasia-associated lesions or masses [DALMs]), because sporadic adenomas do not appear to carry the same implications for carcinoma risk in these patients. It is helpful for the endoscopist to submit biopsies of the flat "shoulders" of raised lesions; specimens should be in separate containers and specifically labeled.

The absence of flat dysplasia near polypoid dysplasia is helpful for distinguishing sporadic adenomas from dysplasia that may involve endoscopically flat mucosa. If dysplasia is seen in an area of active disease and is flat, one may favor a DALM; however, if the patient is in remission or if a polypoid lesion is seen in a noninflamed area without adjacent flat dysplasia, one may diagnose a sporadic adenoma, especially in an older patient who has had adenomas before.[20]

Unfortunately, it often is not possible for the pathologist to make a specific distinction, because there may be insufficient sampling of flat mucosa near the polyp or because the pathologist may not have been provided with the endoscopic or clinical information.

Surgery in low-grade versus high-grade dysplasia

If low-grade dysplasia is diagnosed in ulcerative colitis, colectomy is considered; if high-grade dysplasia or carcinoma is found, a colectomy is usually performed.[19] Of note, adenocarcinomas in ulcerative colitis tend to invade earlier than sporadic adenocarcinomas, and some may not be visible as a polypoid lesion.

Corroborating pathologic review

Pathologist agreement in the grading of dysplasia is only fair to good, with the highest levels of disagreement in the categories of indefinite for dysplasia and low-grade dysplasia.[17] Given the challenges associated with imperfect interobserver reproducibility in the pathologic diagnosis, and given the clinical implications of a diagnosis of dysplasia in ulcerative colitis, the ACG recommends that the diagnosis be confirmed by a second gastrointestinal pathologist experienced with assessing dysplasia in the setting of inflammatory bowel disease.[19, 28]

Although the overall risk of carcinoma in a patient with ulcerative colitis (UC) is 3-5% over the patient's lifetime, this risk is not uniform across all affected patients: those with longstanding disease and pancolitis are at the highest risk. As noted earlier, dysplasia is a histologic risk factor for carcinoma, and high-grade dysplasia is usually an indication for colectomy.[20]

Age at diagnosis, smoking status, and extent of disease may predict disease severity and progression over time. In addition, complications can occur and may be life-threatening. These include toxic megacolon, colonic perforation, sepsis, pulmonary embolism, disseminated intravascular coagulation, hemorrhage, hypoproteinemia, and severe electrolyte disturbances.[17]

The reader is directed to the following references for additional information regarding the pathology of ulcerative colitis:

• Bernstein CN, Kraut A, Blanchard JF, Rawsthorne P, Yu N, Walld R. The relationship between inflammatory bowel disease and socioeconomic variables. Am J Gastroenterol. Jul 2001;96(7):2117-25. [PMID: 11467642].

• Brentnall TA, Haggitt RC, Rabinovitch PS, et al. Risk and natural history of colonic neoplasia in patients with primary sclerosing cholangitis and ulcerative colitis. Gastroenterology. Feb 1996;110(2):331-8. [PMID: 8566577].

• Caruso R, Sarra M, Stolfi C, et al. Interleukin-25 inhibits interleukin-12 production and Th1 cell-driven inflammation in the gut. Gastroenterology. Jun 2009;136(7):2270-9. [PMID: 19505427].

• Rankin GB, Watts HD, Melnyk CS, Kelley ML Jr. National Cooperative Crohn's Disease Study: extraintestinal manifestations and perianal complications. Gastroenterology. Oct 1979;77(4 Pt 2):914-20. [PMID: 467943].

• Romberg-Camps MJ, Dagnelie PC, Kester AD, et al. Influence of phenotype at diagnosis and of other potential prognostic factors on the course of inflammatory bowel disease. Am J Gastroenterol. Feb 2009;104(2):371-83. [PMID: 19174787].