GENETICS

ENVIRONMENTAL FACTORS

It is now almost universally accepted that the pathogenesis of IBD is a result of continuous antigenic stimulation by commensal enteric bacteria, fungi, or viruses, which leads to chronic inflammation in genetically susceptible hosts who have defects in mucosal barrier function, microbial killing, or immunoregulation. Several infectious organisms, including mycobacteria and viruses, have been implicated in the pathogenesis of IBD. No specific infective organism, however, has been isolated consistently from patients with UC, and therefore it is unlikely that the disease is caused by a single common infectious agent.

Numerous clinical and experimental observations have suggested involvement of intestinal bacterial microbiota in the pathogenesis of IBD. The most obvious observation perhaps is that Crohn’s disease and UC preferentially occur in regions of the bowel that contain the highest concentration of bacteria, namely, the terminal ileum and the colon, where bacterial concentrations approach 10¹² organisms per gram of luminal contents. Interestingly, diverting the fecal stream in patients with Crohn’s disease can treat and even prevent disease, whereas reinfusion of ileostomy contents leads to new inflammatory changes within only one week.³⁴ Other human data have shown that antibiotics are useful in the treatment or postoperative prevention of Crohn’s disease and pouchitis. Finally, probiotics (discussed later in this chapter) have been shown to have efficacy in the primary and secondary prevention of pouchitis. The most glaring evidence of the necessary role of bacteria in the pathogenesis of IBD from rodent data is that genetically susceptible mice or rats in a gnotobiotic (germ-free) environment do not have intestinal inflammation; however, these same rodents rapidly develop intestinal inflammation after bacterial colonization.^35–38 Just as in humans, rodent gut inflammation can be treated and prevented with antibiotics and probiotics.^39,40

With respect to the human gastrointestinal microbiome, much has been learned in the last few years. Complex mathematical models have estimated that the human gastrointestinal microbiome contains at least 1800 genera and between 15,000 and 36,000 species of bacteria^41–43; three studies have identified more than 45,000 bacterial small-subunit (SSU) rRNA genes, a technique that captures at most only 50% of the predicted species-level biodiversity.^41–43 Of the bacterial genes identified to date, almost all (more than 98%) can be grouped into four phyla: Firmicutes accounts for 64% of the total and includes the family Lachnospiraceae (e.g., Clostridium groups XIVa and IV) and the subgroup Bacillus (e.g., Streptococcaceae and Lactobacillales). Bacteroidetes account for 23% of the total, Proteobacteria account for 8% of the total and include the family Enterobacteriaceae (e.g., Escherichia coli), and Actinobacteria account for 3%.⁴³

Four general mechanisms have been postulated to explain how components of the normal intestinal microbiome might initiate or contribute to the development of the chronic inflammatory state.⁴⁴ First, microbes can induce intestinal inflammation, either by adhering to or invading intestinal epithelial cells, thereby causing downstream proinflammatory cytokine production or by producing enterotoxins.

Second, a breakdown in the balance between protective and harmful intestinal bacteria, termed dysbiosis, can lead to disease. Most studies comparing the intestinal microbiome in IBD with that in healthy controls show biodiversity in the IBD populations is decreased by 30% to 50%. One study found that this reduction in biodiversity was due to decreased concentrations of Firmicutes (specifically Lachnospiraceae) by 300-fold and Bacteroides by 50-fold.⁴³ The loss of these organisms is important because they are known to produce short-chain fatty acids, such as butyrate, which nourish colonocytes. As a result of their decrease, the relative concentrations of Proteobacteria and Actinobacteria increase in IBD patients relative to controls, although quantitative PCR analysis showed that the absolute numbers of Enterobacteriaceae were not higher in IBD patients than in controls. Loss of protective bacteria, however, could set the stage for overgrowth of pathogenic bacteria. Another interesting finding in this study is that the microbiota of patients with Crohn’s disease and UC were similar to each other. In addition, the study did not identify any individual species that was particularly prevalent in grossly abnormal diseased tissue, and thus no active bacterial etiologic agent was suspected of causing or propagating disease.

The third and fourth ways bacteria could play a role in the pathogenesis of IBD deal with the host itself. Genetic defects in host microbial killing or impaired mucosal barrier function can lead to immune hyper-responsiveness to intestinal bacteria, as the microbes have more exposure to epithelial cells and can trigger the production of high levels of proinflammatory cytokines. Finally, genetic defects in host immunoregulation can lead to a heightened immune response to even nonpathogenic bacteria, such as abnormal antigen processing or presentation, loss of tolerance, or overly aggressive T-cell responses.

In addition to infectious agents, several other environmental factors have been proposed as contributing etiologic factors of UC. The best characterized environmental factor associated with UC is cigarette smoking. Numerous studies have consistently shown that UC is more common among nonsmokers than among current smokers, with the relative risk of UC in nonsmokers ranging from two to six⁴⁵; this association is independent of genetic background and gender. Furthermore, there may be a dose-response relationship, with the disease more common in current light smokers than in heavy smokers. This relationship is consistent with observations that clinical improvement with nicotine therapy in patients with UC appears to be limited to those treated with high doses of nicotine and not those receiving lower doses. This risk of developing UC with smoking is particularly high for former smokers, especially within the first two years of smoking cessation. The rebound effect also is higher for former heavy smokers than for former light smokers.⁴⁶ Smokers also appear to have reduced rates of hospitalization for UC and reduced rates of pouchitis following colectomy.⁴⁷ Studies on the role of passive smoking in UC have yielded conflicting results. A recent meta-analysis of 10 studies examining this issue found no association between passive smoking and future development of UC.⁴⁸

Several mechanisms have been postulated to account for the apparent protective effect of active smoking on UC. These include modulation of cellular and humoral immunity, changes in cytokine levels, increased generation of free oxygen radicals, and modification of eicosanoid-mediated inflammation. Smoking also might have an effect on mucus production by the colonic mucosa, and might alter colonic mucosal blood flow and intestinal motility. No single mechanism, however, can explain the clinical observation of the beneficial influence of smoking on UC and its adverse effect on Crohn’s disease (see Chapter 111).

Other environmental risk factors that have been suggested to influence the development of UC include diet (wheat, maize, cow’s milk, refined sugar, fruits and vegetables, alcohol), oral contraceptives, food additives (silicon dioxide), toothpaste, and breast-feeding⁷; none, however, has been shown conclusively to be associated with UC. Studies have suggested an inverse relationship between appendectomy and the subsequent development of UC^49,50; the mechanisms for this protective effect of appendectomy on UC are unknown. It is possible that removal of appendiceal-associated lymphoid tissues abrogates certain pathologic alterations in mucosal immune responses and therefore prevents the onset of UC.

IMMUNE FACTORS

The prevailing theory of the pathogenesis of UC emphasizes the role of the enteric immune response. The physiologic state of the intestine is one of constant low-grade inflammation in response to environmental stimuli such as bacterial products or endogenous factors. Breaches in this well-regulated mucosal immune system lead to the chronic, uncontrolled mucosal inflammation observed in UC. In this regard, immunologic mechanisms in the pathogenesis of UC involve both humoral and cell-mediated responses.

PSYCHOGENIC FACTORS

Psychosomatic factors first were implicated in the pathogenesis of UC in the 1930s,⁹⁵ but there is no good direct evidence to support this concept. Since the introduction of glucocorticoids for the treatment of patients with UC and the focus on immunologic aspects of the pathogenesis of the disease in the 1950s, this previously widely held notion has diminished in popularity.

Experimental studies have helped identify mechanisms of the proinflammatory potential of stress in animal models of colitis.⁹⁵ When rats are exposed to stress before proinflammatory stimuli are introduced, the severity of colonic inflammation is increased. This particular response has been shown not to be mediated by either vasopressin or corticotropin-releasing factor. In addition, stress has been shown to directly increase intestinal permeability in rats, an action mediated by cholinergic nerves, and to potentiate intestinal inflammation in this particular situation. There are indeed studies reporting that psychosocial stress increases the risk of relapse in patients with quiescent UC.^96,97 Conversely, many of the psychological features observed in patients with UC are likely secondary to this chronic disease process, a phenomenon physicians must be aware of when managing these patients.

SYMPTOMS

SIGNS

Patients with mild or even moderately severe disease exhibit few abnormal physical signs. These patients are usually well nourished and well appearing and show no signs of chronic disease. Caution should be exercised because these patients can appear deceptively well. Weight always should be recorded and, for children and adolescents, both height and weight should be plotted on developmental growth charts. The affected portion of the colon may be tender on abdominal palpation, but tenderness usually is mild and not associated with rebound or guarding. Bowel sounds are normal. Digital rectal examination also is often normal, but the rectal mucosa might feel velvety and edematous; the anal canal may be tender; and blood may be seen on withdrawal of the examining finger.

Patients with severe attacks also might appear well, but most are ill with tachycardia, fever, orthostasis, and weight loss. The abdomen typically is soft, with only mild tenderness over the diseased segment. Abdominal tenderness may become diffuse and moderate with more severe disease. Bowel sounds may be normal or hyperactive but diminish with disease progression. In fulminant colitis, the abdomen often becomes distended and firm, with absent bowel sounds and signs of peritoneal inflammation.

There may be aphthoid ulceration of the oral mucosa. Clubbing of the fingernails is a manifestation of chronic disease. Peripheral edema can occur secondary to hypoalbuminemia. Minor perianal disease may be present but is never as severe as is seen in patients with Crohn’s disease. Signs of extraintestinal manifestations also may be present.

LABORATORY FINDINGS

Laboratory findings in UC are nonspecific and reflect the severity of the underlying disease. Patients with active proctitis and proctosigmoiditis often have normal laboratory test results. Patients with limited distal disease often pass visible blood in the stool, but the amount of blood loss typically is small and anemia, if present, is mild. Patients with active extensive disease or severe distal disease can demonstrate laboratory abnormalities. Hematologic changes, including anemia, leukocytosis, and thrombocytosis, reflect active disease. In contrast, patients with quiescent UC typically manifest no laboratory abnormalities. Iron deficiency anemia may be present because of chronic blood loss. Anemia also may be present secondary to bone marrow suppression resulting from chronic inflammation or medications, including azathioprine, 6-mercaptopurine (6-MP), and sulfasalazine.

Mild or moderate attacks rarely are associated with any biochemical disturbance. Hypokalemia, metabolic alkalosis, and elevated serum levels of blood urea nitrogen and creatinine may be present in severe flares of UC, reflecting volume depletion. Hypoalbuminemia may be seen with acute and chronic disease. Minor elevations in serum levels of aspartate aminotransferase or alkaline phosphatase also are commonly associated with severe disease, but these changes are transient and return to normal when the disease enters remission; these abnormalities probably reflect a combination of fatty liver, sepsis, and poor nutrition. Persistently elevated liver biochemical tests, especially serum alkaline phosphatase, are seen in about 3% of patients with UC and should lead to further investigation, particularly to exclude primary sclerosing cholangitis (PSC) (see Chapter 68).

Serum inflammatory markers including erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) may be elevated in active disease. These abnormalities are typically absent or minimal in patients with mildly to moderately active disease. Elevation in these inflammatory parameters is neither sensitive nor specific for UC; measuring them, however, may be useful in clinical practice to assess disease activity in individual patients, particularly if these values are normal during periods of inactive disease. For following clinical changes, CRP is more sensitive than ESR because of the shorter half-life of CRP.

ENDOSCOPY

The diagnosis of UC can be strongly suggested by sigmoidoscopy in most cases. In patients presenting with their first attack of UC, sigmoidoscopy with biopsies usually is sufficient to confirm the diagnosis, thereby allowing initiation of therapy. In patients with active flares, sigmoidoscopy is best performed in unprepared bowel so the earliest signs of UC can be detected without the hyperemia that is often present because of preparative enemas. Colonoscopy is not recommended in patients with severely active disease for fear of perforation; care must be taken to avoid excessive distention. After active disease has been controlled in a patient with newly diagnosed UC, colonoscopy should be performed to establish the extent of the disease and to exclude Crohn’s disease or other disease states that can complicate UC.

Multiple biopsy specimens should be taken from throughout the colon to map the histologic extent of disease and to confirm the diagnosis if there is concern about Crohn’s disease.¹²⁰ In addition, intubation and biopsy of the terminal ileum should be attempted to exclude the presence of Crohn’s disease or other disease states that can mimic IBD.

In patients with an established diagnosis of UC who present with a typical flare, sigmoidoscopy usually is not necessary, although it may be indicated for the rapid diagnosis of pseudomembranous colitis. Sigmoidoscopy combined with histologic evaluation, however, may be useful for assessing disease severity, particularly when therapeutic response is in question. Colonoscopy may be similarly useful, especially in patients whose symptoms seem out of proportion to the known extent of disease. Additionally, colonoscopy is essential for colorectal cancer surveillance (see below).

The hallmark of UC is symmetrical and continuous inflammation that begins in the rectum and extends proximally without interruption for the entire extent of disease. The earliest endoscopic sign of UC is a decrease or loss of the normal vascular pattern, with mucosal erythema and edema (Fig. 112-6); distortion or loss of vascular markings may be the only endoscopic evidence of UC in patients with quiescent disease. As disease progresses, the mucosa becomes granular and friable. With more-severe inflammation, the mucosa may be covered by yellow-brown mucopurulent exudates associated with mucosal ulcerations. In UC, mucosal ulcerations occur in areas of inflammation, vary in size from a few millimeters to several centimeters, and may be punctate, annular, linear, or serpiginous. Finally, severe UC is associated with mucosa that bleeds spontaneously, and, with diffuse colitis, there may be extensive areas of denuded mucosa from severe mucosal ulcerations (see Fig. 112-6). Marked edema can at times lead to narrowing of the lumen.

Figure 112-6. Spectrum of severity of ulcerative colitis. A, Colonoscopic findings in mild ulcerative colitis demonstrating edema, loss of vascularity, and patchy subepithelial hemorrhage. B, Colonoscopic findings in severe ulcerative colitis demonstrating loss of vascularity, hemorrhage, and mucopus. The mucosa is friable, with spontaneous bleeding as well as bleeding after the mucosa is touched by the endoscope. C, Histopathology showing a severe acute and chronic inflammatory process, with multiple crypt abscesses. D, The colonic architecture is distorted, with a loss of crypts and abnormal branching of the crypts. Recognition of the disordered architecture is useful in differentiating acute from chronic colitis. E, Colonoscopy findings in a patient with chronic ulcerative colitis undergoing surveillance for colorectal cancer. The ascending colon (top left), transverse colon (top right), and descending colon (bottom left) are normal, but the sigmoid colon shows active inflammation (bottom right). F, A biopsy specimen of the normal-appearing colon demonstrates abnormal architecture with shortened crypts and no active colitis.

In patients with long-standing UC, pseudopolyps may be present. Inflammatory pseudopolyps develop in active disease and result from inflamed, regenerating epithelium that is interposed among ulcerations. These inflammatory pseudopolyps may give the colonic mucosa a cobblestoned appearance. With repeated inflammation that is followed by healing, these pseudopolyps remain during the quiescent phase of disease and usually do not regress with treatment. Endoscopically, pseudopolyps typically are small, soft, pale, fleshy, and glistening; however, they may be large, sessile, or pedunculated and may have surface ulcerations. Differentiation of these benign pseudopolyps from neoplastic polyps may be difficult and require histologic confirmation.

There is a loss of normal colonic architecture with long-standing inflammation that is characterized by muscular hypertrophy, loss of the normal haustral fold pattern, decreased luminal diameter, and shortening of the colon; a resultant featureless appearance of the colon in chronic UC gives rise to the lead pipe appearance seen on barium enema. Strictures can occur in patients with chronic UC and result from focal muscular hypertrophy associated with inflammation. Malignancy must be excluded in patients with UC who have strictures, particularly long strictures without associated inflammation and strictures proximal to the splenic flexure (see later).

RADIOLOGY

Plain Films

Patients with a severe attack of UC should have a supine plain film of the abdomen. The presence of intraperitoneal air may be missed on plain abdominal films, however, and CT has demonstrated a better diagnostic yield than plain abdominal radiography for detecting disease complications and extent. In the presence of severe disease, the luminal margin of the colon—the interface between the colonic mucosa and the luminal gas—becomes edematous and irregular. Thickening of the colonic wall often is apparent on a plain film, and prognostic signs such as islands of residual mucosa surrounded by extensive deep ulcerations, distention of the small bowel, and dilatation of the colon can be detected (Fig. 112-7).

Figure 112-7. Plain abdominal film of a patient with severe ulcerative colitis. The transverse colon is dilated (arrow), the colon wall is thickened, and mucosal islands are visible. In addition, distended loops of small bowel are apparent.

Plain films also are useful for detecting the presence of fecal material. Inflamed colons seldom contains feces, and no fecal material is present when the whole colon is involved. It is common, however, for a patient with left-sided disease to have proximal constipation (Fig. 112-8). Thus, a plain film can give considerable information with respect to the extent of disease. The presence of marked colonic dilatation suggests fulminant colitis or toxic megacolon. A plain abdominal film also can detect unsuspected free air and is especially useful in following the daily progress of a patient on high-dose glucocorticoid therapy in whom such a complication may be masked.

Figure 112-8. Plain abdominal film of a patient with mild left-sided ulcerative colitis showing a stool-filled proximal colon.

Barium Enema

With the advent of endoscopy, barium studies have been used less often in the care of patients with UC. Barium studies of the colon remain important, however, and may be superior to colonoscopy for certain specific scenarios, such as evaluation of colonic strictures; barium enema provides information on their location, length, and diameter and allows visualization of the entire colon when the presence of strictures precludes advancement of the colonoscope. Upper gastrointestinal barium study and small bowel follow-through with air-contrast visualization of the terminal ileum should be performed to exclude Crohn’s disease.

The earliest radiologic change of UC seen on barium studies is fine mucosal granularity (Fig. 112-9). The mucosal line becomes irregular and is not as sharp as that of a normal colon. With increasing severity, the mucosal line becomes thickened and irregular, and superficial ulcers are well shown en face. Deep ulceration can appear as collar-stud or collar-button ulcers in tangent, which indicates that the ulceration has extended through the mucosa to the muscularis propria (Fig. 112-10).

Figure 112-9. Film from a double-contrast barium enema examination in a patient with long-standing ulcerative colitis as indicated by a marked loss of haustration. The mucosa is finely granular throughout the colon, consistent with mildly active disease. The terminal ileum is normal.

Figure 112-10. Film from a double-contrast barium enema examination in a patient with active ulcerative colitis. This localized view of the splenic flexure shows multiple ulcers. At the flexure itself there is deep ulceration appearing as a collar-button ulcer (arrow).

Haustral folds may be normal in mild disease but become edematous and thickened as disease progresses. Loss of haustrations also can occur, especially in patients with long-standing disease (Fig. 112-11). Because the left colon may normally lack haustration, this sign is relevant for only the ascending and transverse colon. With long-standing disease, loss of haustration can lead to a featureless and tubular appearance of the colon.

Figure 112-11. Film from a barium enema examination showing postinflammatory polyposis in a shortened sigmoid and descending colon in a patient with active ulcerative colitis.

Other chronic changes are shortening of the colon and widening of the presacral (retrorectal) space as seen on a lateral film of the rectum. Pseudopolyps may be present and often are filiform. In the presence of active changes, these pseudopolypoid changes can resemble a cobblestone pattern (see Fig. 112-11).

CROHN’S DISEASE

Crohn’s disease should be excluded in all patients given a diagnosis of UC, and colonoscopy with multiple biopsies is important in this regard. The presence of skip lesions or granulomas supports the diagnosis of Crohn’s disease. It is important to recognize that muciphage granulomas may be present in patients who do not have Crohn’s disease. Other endoscopic features distinguishing UC from Crohn’s disease are listed in Table 112-4.

Table 112-4 Endoscopic Differentiation of Ulcerative Colitis and Crohn’s Disease

It is not uncommon for patients with ileal Crohn’s disease also to have rectal involvement, and these patients might present with symptoms of proctitis rather than symptoms of small intestinal involvement. In patients with proctitis or diffuse pancolitis, it may be impossible to differentiate UC and Crohn’s disease. Thus it is advisable to obtain radiologic assessment of the small intestine in all patients with colonic disease, particularly in those with pancolitis or proctitis on colonoscopy and elevated inflammatory markers or hypoalbuminemia on laboratory testing. A definitive diagnosis of Crohn’s disease may not be able to be made until the development of small bowel disease or perianal complications.

“Indeterminate” colitis is diagnosed in approximately 10% of patients when the distinction between Crohn’s disease and UC cannot be made. These patients generally are initially managed as if they had UC, until more characteristic features of Crohn’s disease appear. The key points in the differential diagnosis between UC and Crohn’s disease are summarized in Tables 112-4 and 112-5.

INFECTION

Another major category of differential diagnosis for UC is infection (see Chapter 107). As mentioned previously, newly diagnosed UC can manifest as part of a well-documented episode of infectious colitis. It is unknown if the infection prompts the UC or simply unmasks underlying UC that previously had subclinical activity. Patients with documented UC in clinical remission also can develop acute infectious colitis and present with symptoms of a flare of UC. Thus, infections need to be excluded with each episode of disease exacerbation.

The most common organisms causing infectious colitis are Salmonella, Shigella, and Campylobacter. Patients with infectious colitides usually have a more acute onset of symptoms than do patients with a flare of UC, and they have a prominence of abdominal pain; they also might report diarrheal illness in one or more of their contacts. The sigmoidoscopic appearance of infectious colitis may be indistinguishable from that of UC, but the histologic appearance often is helpful in differentiating infectious acute colitis from a more chronic condition. The presence of a chronic inflammatory infiltrate, architectural disturbances, and basal lymphoid aggregates favors a diagnosis of UC, and these features distinguish infectious colitis from UC with a probability of 80%, albeit with considerable interobserver variation. One would not be able to identify bacterial superinfection on a background of chronic UC on histology, however, because the additional changes would be nonspecific; only viral (especially cytomegalovirus) or amebic superinfection would be readily identifiable on biopsy specimens by the presence of intranuclear inclusion bodies or the actual organisms themselves, respectively (see later).

Other bacterial infectious colitides include infection with E. coli O157:H7, which can occur in outbreaks or sporadically. Patients with this infection, particularly children and the elderly, usually present with bloody diarrhea and can develop associated hemolytic-uremic syndrome or thrombotic thrombocytopenic purpura. Because the diagnosis requires a special culture medium and cannot be made on routine stool cultures, clinicians need to have a high index of suspicion and specifically request such a test. Development of molecular probes might facilitate the ability to establish this diagnosis.

Yersinia infections can cause enteritis, enterocolitis, or colitis and can last for several months before resolving spontaneously. The diagnosis is made on the basis of stool culture or a rising titer of serum antibody. Other, less-common bacterial infections causing colitis include Aeromonas hydrophila and Listeria monocytogenes; the former is usually associated with drinking untreated water and the latter is often associated with consumption of unpasteurized milk.

A history of antibiotic use suggests pseudomembranous colitis associated with C. difficile (Chapter 108). This infectious colitis, which is among the most common infections in UC patients, manifests with diarrhea and may be superimposed on or even lead to a relapse of UC. In a national survey of hospitalized patients in the United States, the prevalence of C. difficile among UC patients (3.7%) was eight times higher than that of either non-IBD gastroenterology patients or all hospital-discharge patients.¹²¹ This study also showed that prevalence of C. difficile infection in UC patients has doubled over the last seven years. C. difficile infection can occur in the absence of antibiotics, especially in the elderly and in patients who are taking proton pump inhibitors or who have had placement of a postpyloric tube. Thus, it is important to exclude C. difficile infection in hospitalized patients whose colitis does not respond to medical therapy as expected or who suddenly lose their initial response to treatment.

Patients with C. difficile infections often present with watery diarrhea (usually without rectal bleeding) and have characteristic pseudomembranes at colonoscopy. This infection can cause severe colitis that progresses to toxic megacolon and bowel perforation. Thus, appropriate stool studies for toxin analysis are necessary to exclude superimposed C. difficile infection, even in patients with established UC who present with an exacerbation.¹²²

In patients from endemic areas, certain protozoan and parasitic infections need to be considered (see Chapters 109 and 110). Amebic colitis tends to have a more prolonged course than do most bacterial colitides, but amebiasis is not a cause of chronic colitis. Schistosomal colitis may be chronic and diffuse, exhibit pseudopolyps, and involve the rectum. The presence of characteristic ova in a biopsy specimen confirms the diagnosis.

Other infectious causes of a bloody diarrhea include opportunistic infections of the colon in immunosuppressed patients (see Chapters 33 and 34).

Cytomegalovirus (CMV) infection has been reported in patients with UC, typically those with long-standing disease who are being treated with glucocorticoids or immunosuppressants; the diagnosis of CMV infection should be considered whenever patients who have UC and are taking glucocorticoids either fail to respond as expected or lose their response to treatment. CMV colitis in steroid-naive patients also has been described.¹²³ Patients with CMV colitis often present with abdominal pain and bloody diarrhea and have discrete deep ulcers on colonoscopy; however, CMV colitis can manifest with diffuse inflammation and resemble UC. Because the clinical presentation of CMV colitis may be indistinguishable from a flare of UC, a high index of suspicion is needed to make the diagnosis. Endoscopic biopsies should be obtained from both the ulcer bed and adjacent mucosa; careful histologic examination for giant cells with intranuclear inclusion bodies is important to confirm the diagnosis.

Mycobacterium avium complex usually causes patchy rather than diffuse inflammation.

Sexually transmitted causes of proctitis, including gonorrhea, Chlamydia, and lymphogranuloma venereum, usually do not cause diarrhea and are associated with large volumes of watery pus, especially gonorrhea. These diagnoses are suspected clinically and confirmed by appropriate cultures as well as histologic appearance on rectal biopsy specimens.

OTHER CAUSES

Noninfectious causes of colitis that should be considered in the differential diagnosis of UC include diverticulitis, ischemia, radiation, collagenous colitis, lymphocytic colitis, and drug-induced colitis. Diverticulitis and ischemic colitis usually present acutely or (less commonly) subacutely, but most of the noninfectious colitides have prolonged presentations that can extend for several months.

Acute diverticulitis most commonly occurs in the sigmoid colon and does not involve the rectum (see Chapter 117). When the inflammation does extend to the rectum, it tends to be patchy and involves only the upper rectum. This appearance is more likely to be confused with Crohn’s disease than UC.

Ischemic colitis usually occurs in the elderly (see Chapter 114). The classic distribution is segmental involvement in the watershed areas around the splenic flexure or sigmoid colon; ischemic proctitis also has been described.

Radiation colitis usually occurs in patients who have been given radiation therapy for uterine, cervical, or prostate cancer. The location of disease depends on the sites irradiated but typically involves the rectosigmoid. The onset of symptoms often temporally corresponds to the radiation therapy but can develop years afterward (see Chapter 39).

Microscopic colitis, including lymphocytic and collagenous colitis, manifests with diarrhea and should be distinguished readily from UC by lack of rectal bleeding, the normal endoscopic appearance, and characteristic histopathology (see Chapter 124).

A drug history must always be taken in a patient with colitis, because NSAIDs, gold, and penicillamine among many other drugs may induce colonic inflammation.

Patients with UC can present with symptoms similar to those of IBS (Chapter 118), specifically diarrhea, abdominal pain, fatigue, and poor general well-being. Lack of rectal bleeding and laboratory markers of inflammation, as well as a normal endoscopic and histologic appearance, helps to distinguish IBS from active UC. Patients with quiescent UC can have concomitant symptomatic IBS in the absence of active inflammation. The physician must be able to recognize these subtleties in order to provide appropriate management for the patient. Patients with UC also can present with symptoms mimicking colonic neoplasm (Chapter 123), solitary rectal ulcer syndrome (Chapter 124), diverticular disease (Chapter 117), and factitious diarrhea (Chapter 15). These diagnoses also do not give rise to diffuse inflammation in the colon and, therefore, should be distinguished easily from UC on colonoscopy.

MEDICAL

The goals of therapy of UC are to induce remission, to maintain remission, to maintain adequate nutrition, to minimize disease- and treatment-related complications, and to improve the patient’s quality of life. Current management strategy focuses on using appropriate medical therapy and optimizing timing of surgery.

Several factors should be considered in determining optimal therapy for patients with UC (Fig. 112-12). Current therapeutic strategies can be classified broadly, based on disease activity, into those that treat active disease (induction therapy) (Table 112-9) and those that prevent recurrence of disease once remission is achieved (maintenance therapy) (Table 112-10). This concept of induction and maintenance of remission forms the basis of our evaluation of the efficacy of a specific therapy. The extent of disease is an important consideration that helps determine the route of administration of medication. Thus, for example, proctitis may be treated with suppositories or foam preparations as well as with oral therapy, and enema preparations may be used alone or in combination with systemic therapy for patients with left-sided disease. Other important factors to consider are a patient’s prior response to or side effects from a specific medication and compliance with medication. These factors might favor or preclude the use of a specific agent. Given the chronic nature of UC, medications need to be efficacious and well accepted by patients from the standpoints of safety and ease of administration. The mainstay of medical therapy focuses on regimens that alter host response to decrease mucosal inflammation. Therapies that target other aspects of the systemic inflammatory process or manipulate the enteric flora also have been developed to treat UC.

Figure 112-12. Factors that should be considered in the choice of medical therapies for ulcerative colitis.

Table 112-9 Induction Therapy for Ulcerative Colitis Based on Disease Severity

Aminosalicylates

Oral

Sulfasalazine consists of an antibacterial component, sulfapyridine, bonded by an azo bond to a salicylate, 5-aminosalicylic acid (5-ASA, mesalamine) (Fig. 112-13).¹³¹ The drug was synthesized by Nana Svartz in 1938-1939 and its benefit for the treatment of IBD was discovered serendipitously in 1941-1942 by her when patients with UC receiving this medication for a presumed diagnosis of rheumatoid arthritis noted improvement in colitis symptoms¹³²; in retrospect, these patients had peripheral arthropathy associated with their IBD. Research subsequently established that 5-ASA is the principal therapeutic moiety of sulfasalazine in IBD and that the sulfapyridine component of the parent drug serves as an inactive carrier, largely preventing absorption of 5-ASA in the small intestine and allowing it to be released in the colon.^133,134 Approximately 90% of sulfasalazine reaches the colon, and only a small amount is absorbed in the small intestine. On reaching the colon, the enzyme azoreductase, which is elaborated by colonic bacteria, cleaves the azo bond to release the active constituent moiety, 5-ASA. After 5-ASA is absorbed from the colon, 20% of the compound undergoes hepatic acetylation, forming N-acetyl 5-ASA, and is excreted in the urine. Sulfasalazine is one of several agents in the class of 5-ASA compounds that is considered to be the first line of therapy for inducing remission in patients with mild to moderate UC.^131,135 Mesalamine derivatives have not been evaluated in a randomized, controlled fashion in patients with severely active disease. At a dose of 3 to 6 g/day, sulfasalazine induces remission in 39% to 62% of patients with mild to moderate UC, about twice the remission rate of placebo-treated patients.^136,137

Figure 112-13. Molecular structures of 5-aminosalicylate (5-ASA) preparations.

Various formulations and controlled-release systems (Table 112-11; see Fig. 112-13) have been developed to deliver 5-ASA to specific sites of the gastrointestinal tract without the sulfapyridine moiety, which is thought to be responsible for most of the side effects. Olsalazine (Dipentum) is a 5-ASA dimer linked by an azo bond and is formulated in gelatin capsules. Balsalazide (Colazal) consists of a 5-ASA monomer linked to a biologically inactive carrier molecule, 4-aminobenzoyl-β-alanine. Similar to sulfasalazine, 5-ASA is released from olsalazine and balsalazide in the colon upon cleavage of the azo bond via the bacterial enzyme azoreductase. Approximately 99% of the drug is delivered intact to the colon, and its metabolites are cleared rapidly in the urine.

Table 112-11 Oral 5-Aminosalicylic Acid Preparations and Sites of Delivery in the Gastrointestinal Tract

DRUG	FORMULATION	SITE OF DELIVERY
Prodrugs
Sulfasalazine	Sulfapyridine + 5-ASA	Colon
Olsalazine	5-ASA dimer	Colon
Balsalazide	4-aminobenzoyl β-alanine + 5-ASA	Colon
Mesalamine Preparations
Asacol, Claversal, Salofalk	pH sensitive, resin-coated; delayed release	Distal ileum, colon
Rowasa	Enema	Distal colon
Canasa	Suppository	Rectum
Pentasa	Ethylcellulose-coated microgranules; controlled release	Duodenum to colon
Lialda	pH sensitive, multi-matrix and polymethacrylate coated; delayed and slow release	Distal ileum, colon

5-ASA, 5-aminosalicylate.

Three commonly used mesalamine preparations allow delivery of 5-ASA before the drug reaches the colon: Pentasa, Asacol, and Lialda. Pentasa uses ethyl cellulose-coated microgranules that release mesalamine from the duodenum throughout the small bowel and the colon; about 50% of 5-ASA is released in the small intestine, and the remainder is released in the colon. Asacol is a Eudragit-S-100–coated mesalamine tablet that is released at a pH greater than 7, usually in the distal ileum and the colon. With Asacol, about 15% to 30% of mesalamine is released in the small intestine. Lialda (MMx mesalamine) is a novel mesalamine formulation that uses a multimatrix structure composed of an inner lipophilic matrix and an outer hydrophilic matrix. It is coated with a pH-dependent polymethacrylate film to allow the delayed release of mesalamine in the terminal ileum and colon at a pH greater than 7. This technology also allows mesalamine to be released slowly and in close proximity to the colonic mucosa.

These oral 5-ASA derivatives (mesalamines) have been shown to be superior to placebo for mildly to moderately active UC.^137–142 Meta-analyses have demonstrated that the mesalamines are as efficacious as sulfasalazine, and the various mesalamine preparations appear to be comparable in efficacy.^137,140 Balsalazide has been shown to have superior efficacy and a more rapid response compared with traditional mesalamine agents.^143,144 In a RCT, balsalazide 6.75 g/day, a dose equivalent to mesalamine 2.4 g daily, achieved higher rates of remission and had better tolerance compared with pH-dependent mesalamine 2.4 g/day.¹⁴³ It has been suggested that the greatest benefit of balsalazide is in patients with newly diagnosed left-sided UC.¹⁴⁴

More important than the specific 5-ASA preparation is the dose-dependent response when 5-ASA is used as an induction therapy for active UC.^137,140 For this indication, mesalamine is not effective at doses lower than 2 g daily, and there is an increased response at doses of 4 to 4.8 g daily. The ASCEND I and II trials showed that mesalamine at doses of 2.4 and 4.8 g/day have similar efficacy for patients with mildly active disease, but the higher dose (4.8 g/day) was more efficacious in patients with moderately active disease.^145,146 This dose of mesalamine is comparable to 12 g/day of sulfasalazine, which is impractical in clinical practice because of the high probability of intolerance. No RCT has evaluated the use of aminosalicylates for severely active UC, but these agents are generally thought not to be effective in severely active disease.

Once remission is achieved, sulfasalazine and other 5-aminosalicylates are effective in maintaining it.^147–150 This benefit appears to be dose dependent for sulfasalazine, with a dose of 2 g/day often used to balance efficacy and adverse side effects.¹⁴⁷ Such a dose-dependent response, however, has not been found with the other 5-ASA preparations, and at doses of 1.5 to 4.8 g/day, remission can be maintained in more than 50% of patients.¹⁵¹ One meta-analysis has suggested that sulfasalazine might have a slight but statistically significant therapeutic superiority relative to the newer 5-ASAs in maintaining remission when considering trials of six months’ duration; however, when these trials were combined with those of 12 months’ duration, this statistically significant benefit was lost.¹⁵¹ A double-blind RCT comparing two doses of balsalazide (1.5 g twice daily and 3 g twice daily) with mesalamine 0.5 g three times daily for six months reported a remission rate of 77.5% with the higher dose of balsalazide compared with remission rates of 56.8% and 43.8% with mesalamine and the lower dose of balsalazide, respectively.¹⁵² In general, the same dose of 5-ASA derivative that induces remission is recommended for maintenance therapy, although this recommendation has not been formally tested in a randomized, placebo-controlled fashion.

Common side effects of sulfasalazine include fever, rash, nausea, vomiting, and headache (Table 112-12). Other, less-common but important side effects of sulfasalazine include hypersensitivity reactions, reversible sperm abnormalities, and impairment of folate absorption. Approximately 15% of patients taking sulfasalazine develop significant side effects that require discontinuing the medication. Up to 90% of patients who are intolerant to sulfasalazine, however, can tolerate mesalamine.¹⁵³ In clinical trials, the newer 5-ASA preparations and balsalazide have been shown to be better tolerated than sulfasalazine,^140,154,155 although the adverse event profiles during maintenance therapy appear to be similar for 5-ASA preparations and sulfasalazine.¹⁵¹ Sulfasalazine can impair folate absorption (by competitively inhibiting the jejunal enzyme, folate conjugase) thereby contributing to anemia, and folate supplementation should be prescribed to patients receiving sulfasalazine. Olsalazine is associated with drug-induced diarrhea in up to 10% of patients, which often limits its use. It has been noted that if olsalazine is ingested with meals and is continued despite the diarrhea, the incidence of this side effect can be lessened substantially to 3%. A systematic review of oral 5-ASA for maintenance of remission in UC found olsalazine to be significantly inferior to sulfasalazine, and this reduced efficacy was related mostly to a significantly higher rate of withdrawals because of adverse events.¹⁵¹ Oral mesalamine preparations do not appear to have significant dose-dependent toxicity.

Table 112-12 Side Effects of Sulfasalazine and 5-Aminosalicylates

Dose-Related

Alopecia Anorexia Back pain Folate malabsorption (with sulfasalazine) Headache Nausea, vomiting, dyspepsia

Non–Dose-Related

Agranulocytosis, aplastic anemia

Arthralgia

Colitis

Fever

Fibrosing alveolitis, pulmonary eosinophilia

Hemolytic anemia (Heinz bodies)

Hepatitis

Hypersensitivity skin rashes (occasionally with photosensitivity)

Male infertility (with sulfasalazine)

Pancreatitis

Pericarditis, myocarditis

Topical

Topical aminosalicylates can be administered in the form of 5-ASA enemas, 5-ASA suppositories, and, in Europe, 5-ASA foam. The use of enemas allows the medication to be delivered up to the level of the splenic flexure in about 95% of patients, and suppositories can be used to treat disease up to 15 to 20 cm from the anal verge.

Topical mesalamine derivatives may be used as an alternative monotherapy or as an adjunctive therapy to oral agents in patients with left-sided colitis or pancolitis. They are effective for inducing remission in patients with mildly to moderately active distal UC, without a clear dose-response effect^156,157 in nonrefractory patients. The standard dosing regimens used to induce remission are 1 to 4 g of 5-ASA in the form of an enema nightly, or mesalamine suppositories 1 to 1.5 g either nightly or in divided doses throughout the day. Mesalamine enemas have been shown to be comparable to oral sulfasalazine in the treatment of active distal UC, with fewer side effects.¹⁵⁵ Similar efficacies have been demonstrated for mesalamine enemas regardless of whether the 1-, 2-, or 4-g formulation is used for inducing remission in patients with mild to moderate left-sided UC not requiring concurrent glucocorticoids or immunomodulators. In fact, mesalamine enemas are perceived to be even more effective than topical glucocorticoid enemas in this setting.^157,159 A combination of topical and oral mesalamine also may be more effective than either agent alone in patients with left-sided colitis or pancolitis, suggesting a dose-response effect.^160,161 In patients with proctitis, mesalamine suppositories, 500 mg administered twice daily, have been shown to be beneficial for treating active disease.¹⁶² Mesalamine foam has a more uniform distribution and longer persistence in the distal colon compared with mesalamine enemas. The foam preparation has been shown to have better patient acceptance than the enema preparation,¹⁶³ but mesalamine foams currently are not available in the United States.

Topical mesalamine preparations also are effective for maintaining remission in left-sided UC or proctitis.^156,157 The effective maintenance dosing interval ranges from nightly to every three days. Topical mesalamine is as effective as oral mesalamine,¹⁶⁴ and the combination of topical and oral mesalamine may be more effective than oral mesalamine alone as a maintenance regimen.¹⁶⁵

Systemic

At doses equivalent to 40 to 60 mg/day of oral prednisone, glucocorticoids are effective first-line therapy for moderate or severe flares of UC.^{124,166–170} The use of doses higher than 60 mg/day is associated with increased side effects without appreciable clinical benefit and thus should be avoided. The addition of sulfasalazine to corticosteroids in moderately to severely active UC does not offer any incremental benefit. Although no study has directly compared the efficacy of oral and parenteral glucocorticoids, the latter commonly are used in severe disease. No adequately designed controlled study has been performed to confirm the clinical impression that continuous infusion of parenteral glucocorticoids is superior to pulse therapy.

The use of adrenocorticotropin (ACTH) has been suggested as an alternative to conventional glucocorticoid therapy of active UC in small studies.¹⁷¹ One double-blind RCT suggested that intravenous ACTH was more effective than intravenous hydrocortisone for the treatment of severely active UC only in steroid-naïve patients¹⁷²; this observation has not been confirmed. Because most patients with severely active flares have been treated previously with glucocorticoids, ACTH rarely is used in clinical practice. A noteworthy complication of ACTH therapy is bilateral adrenal hemorrhage.

Glucocorticoids have no maintenance benefits in patients with UC. Steroid-dependent patients, or patients who are unable to taper off glucocorticoids without experiencing disease exacerbation, benefit from the addition of steroid-sparing agents. There has been no trial to date assessing mesalamine therapy and its efficacy in maintaining remission induced with glucocorticoids. The long-term remission rate in patients who require parenteral glucocorticoids for severe UC is approximately 50%.¹⁷³ Immunomodulatory agents, as discussed, should be considered in patients who are dependent on steroids, who require two courses of glucocorticoids for induction of clinical response or remission within one year, or who require parenteral glucocorticoids to induce remission. In addition to the use of immunomodulatory agents, one should consider using infliximab for steroid-dependent patients.

Glucocorticoids are associated with many mild and serious side effects in patients with IBD (Table 112-13). These side effects occur commonly and involve nearly every organ system. Every effort should be made to minimize glucocorticoid use and exposure.

Table 112-13 Side Effects of Glucocorticoids

Azathioprine and 6-Mercaptopurine

Of the various immunomodulatory agents, the most widely used are azathioprine and 6-MP. These two agents are purine analogs that interfere with nucleic acid metabolism and cell growth and exert cytotoxic effects on lymphoid cells. They are inactive prodrugs with subtle structural differences. Azathioprine is nonenzymatically converted to 6-MP, which is then metabolized through a series of enzymatic pathways to active and inactive metabolites (see Fig. 111-8). The two primary metabolites of 6-MP are 6-thioguanine nucleotides (6-TGNs) and 6-methylmercaptopurine (6-MMP). The 6-TGN metabolites are thought be responsible for the immunomodulatory action of azathioprine and 6-MP and their bone marrow suppression property, whereas hepatotoxicity is thought to be related to 6-MMP. One key enzyme involved in the biotransformation of 6-MP is thiopurine methyltransferase (TPMT), which converts 6-MP to its inactive metabolites, 6-MMP and 6-methylmercaptopurine ribonucleotides.

There is a population polymorphism in the TPMT gene: 89% of the population have homozygous wild-type TPMT, and 11% and 0.3% of the population have heterozygous and homozygous mutations, respectively.¹⁸⁷ Persons with heterozygous and homozygous TPMT mutations have decreased to absent enzyme activity. The clinical significance of this genetic polymorphism is that inherited differences in TPMT may be responsible for most of the variability in drug response observed among individual patients.

The efficacy of azathioprine in the treatment of UC is a matter of debate. Four RCTs have evaluated azathioprine for inducting remission in active UC (Table 112-14).^188–191 These four studies were small, heterogeneous in design, used different outcome definitions for response, and reached different conclusions. Two of the studies involved steroid-dependent patients,^190,191 one other study used steroids for induction,¹⁸⁸ and two studies used 5-ASAs as a comparator group rather than placebo.^189,191 Only one study showed a significant benefit with azathioprine compared with 5-ASA for induction therapy in steroid-dependent disease.¹⁹¹

With respect to the use of azathioprine for maintenance of remission in UC, four RCTs have been performed (see Table 112-14).^{188,192–194} Just as with studies of induction therapy, these four studies also had small sample sizes, used heterogeneous designs with different outcome definitions of response, allowed for various cotherapies, and again reached different conclusions. One of the studies was in steroid-dependent disease,¹⁹² another allowed the use of steroids for relapse,¹⁸⁸ one study used 5-ASA as a comparator group rather than placebo,¹⁹⁴ and another included patients who were mostly taking 5-ASAs and was actually a study of azathioprine withdrawal.¹⁹³ Only this withdrawal study showed a benefit with continued azathioprine.

Thus, for the purpose of induction or maintenance therapy for UC, our use of azathioprine is largely based on its established efficacy in Crohn’s disease rather than any proven benefit in UC. One subset of patients, however, has been shown to obtain benefit with the use of azathioprine, specifically patients who have severely active UC and who are able to attain induction of remission with intravenous followed by oral cyclosporine. In these patients, maintenance therapy with azathioprine has been reported to decrease colectomy rates (see later).

The optimal dose of azathioprine or 6-MP for treating UC is unclear, and no formal dose-ranging study has been reported in the literature. The effective doses for 6-MP and azathioprine generally are 1 to 1.5 mg/kg/day and 2 to 3 mg/kg/day, respectively.¹⁹⁵ At these doses, however, there still may be nonresponders and, for them, higher doses may be necessary. Induction of leukopenia had been advocated for dose optimization,¹⁹⁶ but this practice was not supported by subsequent studies.^197–199 Monitoring metabolite levels may be beneficial in determining the optimal dose of azathioprine or 6-MP.

To date, at least 13 studies examining response in IBD with respect to 6-TGN level have been published. A meta-analysis of the first 12 of these studies found that the studies were similar in that they were retrospective and the majority of patients were adults with Crohn’s disease, but they were heterogeneous with respect to sample size, the proportion of patients in remission, and the activity indices used to assess response.²⁰⁰ Of the seven studies that reported data on 6-TGN threshold levels, a pooled analysis of the first six studies showed a three-fold significantly higher rate of remission among patients with a 6-TGN level of greater than 230 to 260 pmol/8 × 10⁸ red blood cells. Incorporation of 6-TGN metabolite measurement into the management regimen of patients receiving azathioprine or 6-MP therapy for IBD is not mandatory and it is a subject of continuing controversy.

Currently, 6-TGN measurement appears to be most useful for identifying reasons for nonresponse to therapy and for suspected noncompliance. If used, metabolite levels should be determined at least two weeks following any dose adjustment to allow sufficient time for the metabolites to reach steady-state.

Currently, it is recommended in the package insert and by the U.S. Food and Drug Administration (FDA) to determine TPMT genotype or phenotype before initiating therapy. The active metabolites, 6-TGNs, also are responsible for myelosuppression with therapy, and patients with TPMT mutation or decreased TPMT enzyme activity are more likely to experience this toxicity because of preferential shunting of 6-MP metabolism toward the excessive production of 6-TGN.²⁰¹ Thus, identifying TPMT polymorphism before initiating azathioprine or 6-MP therapy can decrease the risk of myelotoxicity. Patients with homozygous wild-type TPMT or normal (to high) TPMT enzyme activity level may receive these agents starting at the weight-based optimal dose of 2.5 mg/kg/day for azathioprine or 1.5 mg/kg/day for 6-MP. It has been suggested by some investigators that in patients with heterozygous TPMT mutation or intermediate enzyme activity level, 6-MP or azathioprine should be started at 50% of the weight-based optimal dose. Alternative therapy should be considered in patients with homozygous mutations for TPMT. Regardless of whether a patient’s TMPT genotype or phenotype is known, continued frequent monitoring of complete blood counts remains necessary, because only 27% of all patients with leukopenia have TPMT mutations.²⁰² In addition, two studies have reported that TPMT testing may be cost effective.^203,204

Azathioprine and 6-MP therapy have a delayed onset of action. The mean time to clinical response with azathioprine or 6-MP therapy in patients with UC has been reported to be three to four months in uncontrolled studies,^205,206 a figure that is similar to the 17 weeks’ response time to clinical benefit in placebo-controlled trials of azathioprine or 6-MP therapy for active Crohn’s disease.²⁰⁷ Intravenous loading of azathioprine at 40 mg/kg for 36 hours does not shorten the time required for a therapeutic response in patients with Crohn’s disease.²⁰² Such practice presumably would have the same results if attempted in patients with UC.

Because azathioprine or 6-MP therapy is associated with a number of potentially significant toxicities, its duration of therapy should be determined by weighing clinical benefit against these potential toxicities. The optimal duration of maintenance therapy with azathioprine or 6-MP currently is unknown in patients with UC. In patients with Crohn’s disease, the maintenance benefit of azathioprine or 6-MP can be observed for at least five years.^208,209 Based on these data in Crohn’s disease and the paucity of alternative maintenance therapies, in patients with UC in whom remission is maintained with azathioprine or 6-MP, treatment generally is continued indefinitely as long as there is no significant adverse side effect.

Common side effects of azathioprine and 6-MP therapy include nausea, vomiting, bone marrow suppression, pancreatitis, allergic reactions, and infections (Table 112-15).^210,211 Bone marrow suppression occurs in 2% to 5% of patients.^210,212 It is dose dependent and manifests primarily as leukopenia, although all three cell lines may be affected. This hematologic toxicity can increase with concurrent use of sulfasalazine or mesalamine compounds.^{198,213–215} It is known that mesalamine can interact with the enzyme TPMT, leading to increased levels of 6-TGN, and that this interaction has been associated with leukopenia. Bone marrow suppression is managed by reducing the dosage of immunomodulator or withdrawing the medication. Routine monitoring of complete blood count with differentials is necessary for patients receiving azathioprine or 6-MP and should be continued for the entire duration of therapy. Allergic reactions to azathioprine or 6-MP usually manifest as fever, rash, and arthralgia and resolve following discontinuation of these medications.^212,216 Recurrence of similar reactions occurs with medication challenge, although patients who develop allergic reactions to one agent may be able to tolerate subsequent challenge with the other.²¹⁷ Pancreatitis also is idiosyncratic and independent of dosage.^212,218,219 It usually occurs during the first month of therapy and is reversible upon withdrawal of the drug.

Table 112-15 Side Effects of Azathioprine and 6-Mercaptopurine

Patients using azathioprine or 6-MP therapy can have abnormal liver biochemical tests, but these usually resolve following drug withdrawal.²²⁰ Because liver biopsy is not performed routinely in these patients, their pattern of hepatic injury, if any, is unknown. Cholestasis with inflammation, nodular regenerative hyperplasia, and peliosis hepatis have been reported with azathioprine and 6-MP therapy.^212,220 As is the case for complete blood counts, routine monitoring of liver biochemical tests is recommended. An increased risk of malignancy, primarily lymphoma, has been reported, but not consistently. A meta-analysis of six studies examining this risk reported a four-fold elevated risk of lymphoma with 6-MP/azathioprine.²²¹ The lymphoma that develops in patients who have IBD and receive these immunomodulatory agents appears to be associated with Epstein-Barr virus.²²²

Cyclosporine

Cyclosporine A is a potent inhibitor of cell-mediated immunity. Its use in UC is primarily in patients with severe, steroid-refractory disease.

There has only been one randomized, placebo-controlled trial evaluating the efficacy of intravenous cyclosporine in severe UC. In this study of 20 patients who did not respond to at least seven days of intravenous hydrocortisone, nine (82%) of the 11 patients receiving continuous intravenous infusion of cyclosporine at 4 mg/kg/day responded, compared with none of the nine patients receiving placebo therapy.²²³ The time to clinical response was rapid, at a mean of seven days. After the intravenous route of therapy was converted to oral cyclosporine, 44% of those patients who responded initially required colectomy during the six-month follow-up period.

Intravenous cyclosporine monotherapy may be as effective as intravenous glucocorticoids in patients with severely active UC; its use thus potentially minimizes the toxicities of combination therapy.²²⁴ The addition of azathioprine or 6-MP in patients who have responded to intravenous cyclosporine has been shown in other studies to reduce the rate of relapse or colectomy.^225,226 Thus, cyclosporine can be considered a bridge therapy to control active disease in patients with steroid-refractory UC while waiting for elective surgery or the onset of action of azathioprine or 6-MP.

With the addition of azathioprine, long-term remission at one year may be more likely in patients who initially respond to intravenous cyclosporine monotherapy than in those who respond to intravenous glucocorticoids. A European retrospective cohort study of 142 patients who were treated with cyclosporine, 118 of whom responded initially, reported the probability of avoiding colectomy to be 63% at one year, 41% at four years, and 12% at seven years; overall, 54% of patients required colectomy at some point.²²⁷ Patients who were already taking 6-MP or azathioprine at the time cyclosporine was initiated continued taking their current dose, and those who were naïve to 6-MP or azathioprine were started at target doses at the time of response to cyclosporine during their hospitalization. The authors found that 59% of patients previously taking 6-MP or azathioprine required eventual colectomy, compared with 31% for patients naïve to these drugs (P < 0.05).

Because most of the serious adverse effects associated with the use of cyclosporine are dose-dependent, intravenous doses lower than 4 mg/kg that still can achieve efficacy are desirable. One RCT has shown that a dose of 2 mg/kg is as effective as 4 mg/kg given intravenously in patients with severely active UC, judged by clinical response rates, time to response, and short-term colectomy rates.²²⁸ The mean plasma cyclosporine levels were 237 ng/mL in patients receiving the 2 mg/kg dose and 332 ng/mL in patients receiving the 4 mg/kg dose. Thus, initiating therapy at 2 mg/kg may be reasonable, but regardless of the dose used, careful monitoring of plasma cyclosporine trough levels is necessary.

Cyclosporine has been associated with many adverse effects, including paresthesias, tremors, headache, hypertrichosis, and gingival hyperplasia (Table 112-16). Other potentially serious toxicities include hypertension, seizures, electrolyte and liver biochemistry abnormalities, nephrotoxicity, anaphylaxis, and opportunistic infections. These complications are mostly dose-dependent. Severe complications have been reported with cyclosporine in up to 12% of patients with UC,²²⁹ and two large series have reported death rates of 1.8% to 2.8% with cyclosporine, more than half of which were due to infections acquired while taking the drug.^227,229

Table 112-16 Side Effects of Cyclosporine

Careful monitoring for adverse effects is critical during cyclosporine therapy. Baseline serum electrolytes, creatinine, cholesterol, and liver biochemical values should be measured. Cyclosporine therapy should be avoided in patients with an impaired creatinine clearance to minimize the risk of severe nephrotoxicity. Patients with serum cholesterol lower than 120 mg/dL should receive nutritional support to improve the level before initiating cyclosporine therapy, because a low cholesterol level is associated with an increased risk of seizures. During intravenous therapy, cyclosporine levels should be monitored daily, and the dose should be adjusted to achieve a trough concentration (measured one hour before dosing) between 200 and 400 ng/mL, determined by high-pressure liquid chromatography. Serum electrolytes and serum creatinine levels should be monitored daily or every other day. The dose of cyclosporine also should be decreased when the serum creatinine increases by 20% to 30% over baseline.

If patients respond to intravenous cyclosporine, the route of administration can be changed to oral therapy with 2 mg of oral agent for each 1 mg of intravenous cyclosporine. The drug can be administered in two divided doses daily. Drug monitoring during oral cyclosporine therapy includes weekly trough cyclosporine levels and weekly to biweekly electrolyte and creatinine levels. Oral cyclosporine should be continued for three to six months, while waiting for surgery or for azathioprine or 6-MP to take effect. Patients on long-term cyclosporine therapy should receive Pneumocystis carinii pneumonia prophylaxis with trimethoprim-sulfamethoxazole.

Methotrexate

Methotrexate is a folic acid antagonist and has antimetabolite and anti-inflammatory properties. Although early reports suggested potential benefit of methotrexate administered intramuscularly or orally in UC, the only randomized, placebo-controlled trial failed to demonstrate its efficacy for the treatment of active UC.²³⁰ In this study of 67 patients with chronic active UC, oral methotrexate at 12.5 mg/wk for nine months was comparable to placebo therapy in the rate of achieving first remission, time to first remission, relapse following remission, and the mean glucocorticoid dose. It is unknown if methotrexate at higher doses administered intramuscularly or subcutaneously may be beneficial in inducing or maintaining remission in UC. Given the absence of data supporting its efficacy, methotrexate cannot at this time be considered a standard therapy for UC.

Other Immunomodulators

Alternative immunomodulators have been explored for patients who do not tolerate or have not responded to the previously mentioned immunosuppressants. Mycophenolate mofetil has pharmacodynamic properties similar to those of azathioprine and 6-MP but a more rapid onset of action. A pilot study of patients with chronic active UC receiving concomitant prednisolone found azathioprine to be superior to mycophenolate mofetil throughout the one-year study period, with remission rates at one year of 100% and 88%, respectively.²³¹ Uncontrolled studies reported less than 50% remission rates with mycophenolate mofetil therapy in patients with steroid-dependent UC^232,233 and the intolerance rate was high.²³² A substantial number of patients developed adverse effects necessitating drug withdrawal, including recurrent upper respiratory tract infection, bacterial meningitis, depression, and migraine headache.^231,233

Tacrolimus is another immunosuppressant with actions similar to those of cyclosporine. In contrast to cyclosporine, it has a 100-fold greater potency and a more-rapid onset of action. A number of small uncontrolled studies have suggested benefit of oral or intravenous tacrolimus for the treatment of patients with refractory UC. The only randomized, placebo-controlled trial of tacrolimus in UC involved 63 Japanese patients with either steroid-dependent or steroid-refractory disease who were randomized to receive either initial oral tacrolimus at 0.05 mg/kg or placebo twice daily.²³⁴ Patients in the high-trough concentration (10 to 15 ng/mL) tacrolimus group had a significantly higher rate of response and nonsignificantly higher rate of remission than those in the placebo group at week two, and a number of patients demonstrated response or remission (or both) after an additional 10 weeks of open-label therapy. As with cyclosporine, tacrolimus can result in a number of toxicities including nephrotoxicity, electrolyte abnormalities, nausea, diarrhea, headache, tremors, paresthesias, insomnia, alopecia, hirsutism, and gingival hyperplasia.^235,236 Thus, given the limited data and potential for harmful adverse events, the use of these alternative immunomodulators currently is not incorporated into standard practice.

Anti-Tumor Necrosis Factor Antibodies

TNF is a key proinflammatory cytokine that has been demonstrated to play a role in several disease states, including IBD. Elevated TNF concentrations have been found in inflamed intestine in patients with Crohn’s disease and UC, and stool and mucosal concentrations of TNF in patients with IBD have been shown to correlate with clinical disease activity. Infliximab (Remicade) is a chimeric monoclonal antibody of IgG₁ subclass directed against human TNF-α. It consists of 75% human and 25% murine components (Fig. 112-14). The efficacy of infliximab in Crohn’s disease is well established, and it is approved by the FDA to treat Crohn’s disease and UC. Infliximab is thought to operate in Crohn’s disease via a multitude of mechanisms, including antagonizing the activity of TNF-α,^285,286 initiating cytotoxicity on immune cells,²⁸⁷ and inducing T-cell apoptosis.^288,289

Figure 112-14. Structural diagram of anti–tumor necrosis factor antibody. Normal human immunoglobulin G (IgG) antibody is shown on the left and infliximab is shown on the right. Infliximab is 75% human and 25% murine.

Results from two large, multicenter, randomized, double-blind, placebo-controlled trials (ACT 1 and 2) showed efficacy of infliximab therapy in UC.²⁹⁰ In these two similarly designed trials, 728 patients with moderately to severely active UC who failed conventional therapy with glucocorticoids alone or in combination with thiopurines (ACT 1) or glucocorticoids alone or in combination with thiopurines and 5-aminosalicylates (ACT 2) were randomized to placebo, infliximab 5 mg/kg, or infliximab 10 mg/kg at weeks 0 and 2 and then every eight weeks through week 46 (ACT 1) or week 22 (ACT 2). With respect to clinical response at week 8, in ACT 1 69% and 61% of patients receiving infliximab at 5 and 10 mg/kg, respectively, had a clinical response, compared with 37% of patients receiving placebo (P < 0.001 for both comparisons). In ACT 2 at week 8, 64% and 69% of patients receiving infliximab at 5 mg/kg and 10 mg/kg, respectively, had a clinical response, compared with 29% of patients receiving placebo (P < 0.001 for both comparisons). With respect to clinical remission at week 8 in ACT 1, 39% and 32% of patients receiving infliximab at 5 mg/kg and 10 mg/kg, respectively, attained remission, compared with 15% of patients receiving placebo (P < 0.003 for both comparisons). In ACT 2 at week 8, 34% and 28% of patients receiving infliximab at 5 mg/kg and 10 mg/kg, respectively, attained remission, compared with 6% of patients receiving placebo (P < 0.001 for both comparisons). The results for clinical remission at week 30 (ACT 1 and 2) and week 54 (ACT 1) were very similar for all groups, with highly significant greater than two-fold higher remission rates for the infliximab-treated patients. The proportions of patients with a sustained clinical response or remission also were significantly higher in the infliximab groups. Treatment with infliximab also was shown to have steroid-sparing and mucosal healing properties.

These data have led to the approval of infliximab by the FDA for patients with moderately to severely active UC who have had an inadequate response to conventional therapy. Infliximab is now accepted as part of the standard treatment options in patients with UC. Two other anti-TNF agents, adalimumab and certolizumab pegol, have shown efficacy for the induction and maintenance of remission in Crohn’s disease but have not yet been studied in patients with UC.

Anti-Adhesion Molecules

Several agents directed at blocking small adhesion molecules have been evaluated for the treatment of UC. These molecules are glycoproteins expressed on the surfaces of endothelial cells and lymphocytes. Adhesion molecules are important in cellular trafficking in IBD and other diseases, in which immune and inflammatory cells from the periphery are recruited into sites of inflammation. Among these, natalizumab is a humanized IgG₄ monoclonal antibody against lymphocyte adhesion molecules, α₄ integrins. A pilot study of 10 patients with active UC suggested clinical benefit with a single infusion of 3 mg/kg of natalizumab.²⁹¹ Natalizumab currently is approved for treating patients who have Crohn’s disease and in whom anti-TNF therapy has failed; its use in UC is now undergoing evaluation.

Another anti-adhesion molecule agent is MLN-02 (formerly called LDP-02), a humanized IgG₁ monoclonal antibody to α₄β₇ integrin. In a phase 2 study, two infusions of 0.5 mg/kg of MNL-02 administered 29 days apart were found to be effective in achieving clinical remission and response at six weeks after the initial infusion in patients with moderately active UC.²⁹²

Others

Although historically considered more important in the inflammation of Crohn’s disease, as it is produced by Th1 cells, IL-2 also has been implicated in UC inflammation (see earlier). Two agents designed to block the binding of IL-2 to its receptor have been examined for potential efficacy in UC. Daclizumab, a humanized monoclonal antibody against the IL-2 receptor (CD25), was suggested to be beneficial in patients with refractory UC in a small open-label pilot study.²⁹³ A potential clinical benefit also has been reported with basiliximab, a chimeric monoclonal antibody to the IL-2 receptor, in a small, uncontrolled study of patients with steroid-refractory UC.²⁹⁴ Along with the emphasis on T-cell–mediated immune response in the pathogenesis of UC, a humanized monoclonal antibody to CD3, visilizumab, has shown promise in an open-label phase 1 study of 32 hospitalized patients with UC whose disease failed to respond to intravenous glucocorticoids.²⁹⁵

Other biological therapies include agents targeted at tissue repair and restitution following mucosal injury. In this regard, epidermal growth factor (EGF) is a potent mitogenic peptide that stimulates cell proliferation in the gastrointestinal tract. A preliminary study showed that EGF enemas at a dose of 5 µg daily for two weeks was effective in treating mild to moderate left-sided UC when administered along with oral mesalamine.²⁹⁶ In contrast, another potent stimulant of intestinal epithelial cells, repifermin (keratinocyte growth factor 2) was not found to be more effective than placebo when administered intravenously in patients with active UC in a phase 2 dose-ranging study.²⁹⁷ Further studies clearly are necessary to confirm some of these early promising findings.

Colectomy

Subtotal colectomy with ileostomy is the least extensive of these operations. Most of the colon is removed, a Hartman pouch or a mucus fistula is created for the remaining colon, and an end-ileostomy is created. This surgery typically is performed in patients requiring emergent surgery for severe or fulminant colitis, and it has the advantage of allowing restorative surgery in the future. Colectomy with ileorectal anastomosis is similar to subtotal colectomy with ileostomy, but it maintains bowel continuity. Many patients continue to have attacks of proctitis and the retained rectal stump is at risk for developing colorectal cancer. Thus, lifelong endoscopic surveillance of the rectum is necessary for patients who elect this type of operation.

Total proctocolectomy with a permanent Brooke end-ileostomy was one of the earliest operations performed for UC. Removal of the entire colon and rectum eliminates any future disease and risk of colorectal cancer. The primary disadvantage of this operation is the presence of the permanent ileostomy, which might not be acceptable from the standpoint of quality of life for some patients. This is the operation of choice for elderly patients, those with anal dysfunction, and those who do not wish to have a restorative proctocolectomy. Proctocolectomy with continent ileostomy (Koch pouch) was developed as an alternative to the conventional end-ileostomy.³¹² In this operation, loops of small bowel are used to create an intra-abdominal pouch with an intussuscepted (nipple) valve. This pouch allows storage of stool contents and is attached to the abdominal wall with a flush ostomy opening. The stool contents in the pouch are emptied by inserting a suction-catheter through the stoma. Because of technical challenges associated with this operation (e.g., slippage of the nipple valve) and the development of restorative procedures, proctocolectomy with continent ileostomy rarely is performed.

Proctocolectomy with Ileal Pouch-Anal Anastomosis

Restorative proctocolectomy with IPAA currently is the operation of choice for most patients with UC who require elective colectomy. In this procedure, the entire colon and rectum are removed, the anal sphincters are preserved, and a pouch is constructed from approximately 20 cm of the distal ileum (see Fig. 112-16). Bowel continuity is established by anastomosing this pouch with the anal canal. An IPAA usually is performed as a two-stage operation, during the first stage of which a temporary diverting ileostomy is created to allow the ileal pouch to heal. This operation can be performed as a single-stage operation; however, there have been reports suggesting a higher rate of bowel obstruction and sepsis when this is done.³¹³ The ileostomy is reversed after approximately two to four months.

Proctocolectomy with IPAA presents technical challenges and might not be suitable or technically feasible for all patients. Most reports suggest satisfactory quality of life following IPAA surgery.³¹⁴ Mean stool frequency ranges from four to nine bowel movements per day, including one or two nocturnal stools. Nocturnal seepage occurs in approximately 20% of patients in the early postoperative period but is infrequent after the first year.

Rates of complications following IPAA surgery vary widely. In a series from the Cleveland Clinic of more than 1000 patients undergoing restorative proctocolectomy and IPAA, most of whom had UC, the overall morbidity rate was 63% (early complications 28%, late complications 51%).³¹⁴ A larger case series from the Mayo Clinic included 1885 patients who had proctocolectomy with IPAA for UC and who were followed for up to 20 years (mean follow-up 11 years).³¹⁵ Pouch success rates were reported of 96.3%, 93.3%, 92.4%, and 92.1% by 5, 10, 15, and 20 years, respectively, however, complication rates also were high and included pouchitis (48% by 10 years, 70% by 20 years), small bowel obstruction (42% by 20 years), anastomotic stricture (39% by 20 years), abscess (16% by 20 years), and fistula (14% by 20 years), as well as pelvic sepsis. Other complications of proctocolectomy with IPAA are covered in Chapter 113 and include fecal incontinence and sexual and urinary dysfunction.

This surgery is best performed in centers with considerable experience with the operation and with managing pouch dysfunction. These high rates of complications will likely decrease over time as experience grows with this type of surgery.

A widely performed modification of proctocolectomy with IPAA is proctocolectomy with ileal pouch-anal transition zone anastomosis. This technically less-complex surgery involves stapling the ileal pouch to the distal rectum in close proximity to the dentate line (1 to 4 cm), thereby eliminating the need to perform rectal mucosectomy. This type of surgery might carry a lower risk for fecal incontinence and may be performed as a single-stage operation without a temporary diverting ileostomy.

Laparoscopic approaches to the proctocolectomy with IPAA have become popular among both physicians and patients. Because this modality is relatively new, comparison data versus open proctocolectomy with IPAA are limited. A case-matched series from the Mayo Clinic, however, which included 100 consecutive laparoscopic IPAA case patients matched to 200 open IPAA control patients, reported that although the median operative time was significantly longer for the laparoscopic group by 103 minutes, laparoscopic patients had significantly improved recovery, as witnessed by lower median time to regular diet (three vs. five days), time to ileostomy output (two vs. three days), length of stay (four vs. seven days), and intravenous narcotic use (all P < 0.05).³¹⁶ In addition, postoperative morbidity and hospital readmission rates were similar between the two groups and no mortalities were observed. Fewer patients required reoperation within three months with the laparoscopic approach (3% vs. 6.5%, P < 0.2). A follow-up survey in this same cohort of patients one year after operation showed that patients reported high cosmetic, body image, quality of life, and sexual function scores irrespective of the type of operation.³¹⁷

A smaller randomized trial from Amsterdam of 60 patients who underwent laparoscopic or open IPAA found significantly higher operative time for the laparoscopic approach (by 77 minutes) but similar morbidity (20% vs. 17%), length of stay (10 vs. 11 days), time to regular diet, narcotic requirement, and quality of life at three months after surgery in both groups.³¹⁸ Larger RCTs will need to be conducted before any definitive conclusions can be made regarding the superiority of one approach over another. The experience of the particular surgeon and surgical center is likely to be an important determinant of success rates irrespective of procedure.