25 Ulcerative Colitis
Abstract
Ulcerative colitis is a potentially complex and crippling disease process, challenging for the provider and cruel to the patient. While many patients are successfully managed with minimal medical intervention, a significant number will progress to severe, debilitating, and intractable disease, ultimately necessitating a high-risk surgery. Multidisciplinary discussion and formation of a solid therapeutic alliance are critical parts of the treatment process, particularly in refractory cases. It should be kept in mind that chronic ulcerative colitis patients often have an innate understanding of their own disease process that far outweighs clinical data. The risk of any intervention is not insignificant, and ultimately it is the patient who has to live with the outcome. Individualized medical and surgical care should therefore be the standard. In practice, this principle is not easy to follow, as it necessitates thorough and often lengthy counseling. However, the ability to provide a potentially curative surgery to a patient who has been experiencing the pain and humiliation of ulcerative colitis is a privilege worthy of additional effort.
25.1 Introduction
Inflammatory bowel disease (IBD) has been described in various forms and terms since Roman times. It is simultaneously fascinating and exasperating that after two millennia this disease spectrum continues to perplex patients and physicians alike. Certainly the aptly named Soranus (AD 117) could not have imagined that the noncontagious diarrhea he described would remain a mysterious phenomenon in an era where living organs can be grown from stem cells. 1
It was not until 1859 that Sir Samuel Wilks coined the term “ulcerative colitis” (UC). 2 The term itself is a bit of a misnomer, as the inflammation generally emanates from the rectum, and ulceration is not a “sine qua non” of the disease. Knowledge of the gross and histologic manifestations of UC became further refined over the course of the next century, and an increasing interest in defining UC as a unique disease process separate from Crohn’s disease (CD) and ischemic colitis developed. A number of etiologic theories were described, including infection, 3 colonic “mucinases,” 4 allergy, 5 psychological factors, 6 and autoimmunity. 7 , 8 Most physicians currently favor the autoimmune theory, which has shaped the modern framework of how we manage UC. Nevertheless, the debate rages on, as immune-modulating medications have variable and inconsistent impacts on the natural history of the disease. Furthermore, the multiple phenotypic variants in and of themselves likely favor a multifactorial mechanism. It could easily be argued that, despite advances in modern medicine and extensive research in the field of IBD, the most accurate definition to this day was spelled out by F.T. De Dombal in 1968, who described UC as follows 9 :
“An inflammatory disease of unknown origin, characterized clinically by recurrent attacks of bloody diarrhea, and pathologically by a diffuse inflammation of the wall of the large bowel. The inflammatory changes spread proximally from the rectum; and are confined to (or most severe in) the colonic and rectal mucosa”.
These qualities help distinguish UC from CD, which is generally characterized by full-thickness inflammation impacting any portion of the gastrointestinal tract. Confinement to the colonic and rectal mucosa adds surgical options to the armamentarium against UC, with proctocolectomy offering hope of a surgical “cure.” In practice, however, the distinction between the two disease processes is somewhat blurred, and many initially thought to have UC turn out to have CD (and vice versa). Physicians and patients must therefore accept some degree of uncertainty when implying one diagnosis over the other.
The impact of IBD on patient lifestyle and psyche cannot be overstated. Many of these patients have lived for years with chronic pain, debilitating anxiety, and routine humiliation due to fecal incontinence. As such, a therapeutic alliance must be established early, ideally in a multidisciplinary setting. Of utmost importance to this alliance is an upfront and thorough explanation of the natural history of the disease, diagnostic limitations, the role of both medications and surgery, and establishment of realistic therapeutic goals that fit with the patient’s life plan.
25.2 Epidemiology
The global incidence of UC ranges from 0.6 to 24.3 cases per 100,000 life years. 10 , 11 While regional variations exist, it seems clear that UC diagnosis rates have increased across the globe over the last 50 years. 10 Molodecky et al 10 recently performed a systematic review of global IBD epidemiologic data between 1930 and 2008. Sixty percent of UC studies reporting at least 10 years of data showed a statistically significant increasing incidence, with increases ranging from 2.4 to 18.1%. Interestingly, only 23% of UC studies performed after 1980 showed a significant increase, suggesting either a true leveling off or simply the calm after a flurry of new diagnoses secondary to increased disease awareness. Corroborating data presented in previous large series, the study found that Europe and North America had the highest disease incidences, reported at 24.3 and 19.2 per 100,000 person-years, respectively. Asia and the Middle East followed similar trends toward increasing incidence; however, their rates were significantly lower at 6.3 per 100,000 person-years.
The prevalence of UC in the 1980s and 1990s ranged from 6 to 246 cases per 100,000 persons, with the largest populations not surprisingly located in North America (37.5–246 cases per 100,000 persons) and Europe (21.4–243 cases per 100,000 persons). Health insurance data suggest there are approximately 512,000 UC patients in the United States alone, 12 with an extrapolated annual treatment cost of $2.1 billion dollars.
Thorough analysis of global trends reveals that UC rates may be more heavily influenced by industrialization than by regional genotypic tendencies. Central to this argument is the parallel between UC rates and the level of industrialization. Developing countries have exceedingly low rates of UC, but as societies become industrialized higher rates of UC emerge. CD rates tend to lag behind UC in this setting, although the trend is the same. Whether these increased rates of IBD are due to improved health care and diagnostic awareness or some inherent gastrointestinal assault associated with the “Western lifestyle” has yet to be determined. 11
UC is usually diagnosed in the third to fourth decade of life, with the mean age of diagnosis 5 to 10 years later than what is seen with CD. 10 , 13 , 14 , 15 Although it has classically been thought of as having a bimodal age distribution with a second, smaller peak in the sixth to seventh decade, recent epidemiologic studies have not supported this pattern. 11 Pediatric IBD accounts for 7 to 10% of all IBD cases, although UC accounts for a much smaller percentage of this population with an incidence of only 0.8 per 100,000 person-years (compared to 5.2 per 100,000 person-years for pediatric CD). 16 , 17 , 18
While the data are mixed, large regional series have suggested a slight (60%) male predominance in UC. 15 , 19 , 20 Undoubtedly contributing to this is an increasing incidence of UC in males in recent years (with a corresponding decrease in females). 15 Interestingly, the pediatric population seems to exhibit the opposite gender profile in UC, with females carrying the bulk of the disease burden. The shift in differential gender balance, if one exists, seems to occur between the ages of 14 and 17 years. 16
There seem to be variations in IBD incidence and prevalence based on race and ethnicity. People of Jewish descent have higher rates of IBD, and tend to have higher rates than other ethnicities regardless of the geographic location. African Americans were once thought to be at lower risk when compared to whites; however, more recent data suggest similar rates of disease. 21 Conversely, Asian-Americans, Hispanic-Americans, and Native Americans are much less likely to develop IBD. 22 , 23 As with geographic differences, it is unclear how much of an impact culture and lifestyle have on these differential disease patterns. Studies of migrant populations suggest that disease patterns may have more to do with the environment than was previously thought. Evidence of this is seen in South Asians, traditionally with low rates of IBD, who migrate to the United Kingdom and develop an increased risk for IBD when compared to whites. 24 , 25 , 26 Age at migration seems to impact this risk, with the highest disease rates seen in those who move before the age of 15 years. 26
25.3 Etiology
The increased incidence of IBD in “Western,” industrialized countries coupled with 8- to 10-fold higher rates of IBD among first-degree relatives of IBD probands suggests a multifactorial etiology, with contributions from both the environment and the genome. 27 , 28 Specific etiologic theories include aberrant immune regulation, defective mucosal barrier function, defective microbial clearance, persistent specific infection, and dysbiosis (abnormal ratio of beneficial and detrimental commensal microbial agents; ▶ Fig. 25.1). 29
Genetic correlates have come to light in recent years, in large part due to the completion of the Human Genome Project in 2003. A thorough understanding of the genetics of IBD will undoubtedly contribute to the understanding of the various etiologic theories, and will hopefully provide new targets for therapeutic intervention. At this point, however, the relative contributions of associated genes, the environment, the gut microbiome, infection, and any as yet unknown elements to the IBD phenotype are uncertain.
Human single nucleotide polymorphism and candidate gene studies together with studies using mouse models of experimental colitis have identified 200 gene loci, and approximately 300 gene candidates, associated with IBD. 30 , 31 , 32 At least 23 of these loci are specifically associated with UC, although the vast majority is associated with both CD and UC. Most, if not all, of these genes are involved in immunoregulation, mucosal barrier integrity and microbial clearance, and/or homeostasis. 29
25.3.1 Aberrant Immune Regulation
A direct link between IBD-associated genetic abnormalities and the phenotypic expression of altered immunity has not been firmly established. What seems clear, however, is that patients with IBD have activated innate and acquired immune systems, with loss of tolerance to resident enteric bacteria. 33 , 34
Under normal circumstances, intestinal macrophages and epithelial cells have the ability to reside in areas of high bacterial loads without causing inflammation. This is due to the downregulation of bacterial recognition receptors (such as Toll-like receptor [TLR] and cluster of differentiation 14 [CD14]) on these cells. 35 When activated by bacterial recognition, TLR molecules on the surface of innate immune effector cells bind microbial adjuvants and initiate the nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB) pathway. 36 Induction of this pathway ultimately results in an increase in proinflammatory cytokines, which have been directly implicated in the pathogenesis of both forms of IBD. 29 Interleukin-1 beta (IL-1β), tumor necrosis factor-α (TNF-α), IL-6, and IL-18 have all been found to be increased in patients with active UC. A similar pattern is found in CD, although IL-12, IL-23, and IL-27 are also increased. 29 Inhibition of these cytokines attenuates experimental colitis, and represents the foundation for anti-TNF-α therapy.
The reason for the upregulation of the innate immune response including activated macrophages and dendritic cells and the presumed resultant loss of tolerance to commensal gut flora is the subject of considerable debate. Recent studies have suggested that loss-of-function mutations in the IL-10 R anti-inflammatory gene may play a role in the exaggerated immune reaction. While the limited studies have linked this mutation to a severe form of early-onset CD, it has been found to be associated with UC as well. 37 , 38 Certainly it makes intuitive sense for a defect in an anti-inflammatory cytokine to result in exaggerated immunogenicity. Abnormal expression of TLR molecules may be another causative factor, a theory that has gained recent popularity. 39
Adaptive immunity is also altered in IBD, generally through dysregulation of T-helper cell type 1 (Th1) and Th2 pathways. Alterations of Th1 responses have been associated with CD in animal models, while UC is generally thought to be related to abnormalities in the Th2 response. This atypical Th2 response in UC has not been elucidated; however, it is felt to be mediated by natural killer T cells that secrete IL-13 upon activation by antigen-presenting cells. 29
25.3.2 Defective Mucosal Barrier Function
Compromise of the gut mucosal barrier layer has also been theorized to play a role in the pathogenesis of IBD. This mechanism has been postulated to be secondary to an intrinsic defect in patients with IBD, which would allow either primary poor functioning of the barrier or defective barrier repair after insults such as ulcerogenic medications or infections. The theory is at least indirectly supported by genetic studies, with MDR1 (UC) and NOD2 (CD) genes presumably playing a role in maintenance of the intestinal mucosal barrier. MDR1 mediates excretion of xenobiotic molecules, and possibly bacteria, from epithelial cells, while NOD2 potentially plays a role in bacterial clearance and production of antimicrobial α-defensins by Paneth’s cells. 29 , 40 , 41
25.3.3 Defective Microbial Clearance
Some studies have noted the persistence of bacteria in the gut tissue and lymph nodes of CD patients, 29 which is the basis for the defective microbial clearance theory of IBD. Other studies showing significant increases in the number of mucosally adherent enteric bacteria in active UC and CD lend additional support to this theory. 42 It is believed by many that this increase in the adherent bacterial load is secondary to aberrant innate immunity resulting in defective killing of bacteria. 43
It seems counterintuitive that there would be increases in the numbers of adherent gut bacteria in IBD given the previously described intolerance to normal commensal organisms; however, these are not mutually exclusive theories. It may be that intolerance of normal bacteria due to alterations in TLRs and the NF-κB pathway coupled with defective bacterial killing results in the dysfunctional and self-perpetuating nonspecific inflammation we see in IBD. One can easily see how a defective mucosal barrier along with the presence of more virulent commensal organisms would further exacerbate this problem. This complex interplay likely explains the difficulty understanding, much less treating, CD and UC.
A number of IBD-associated genes have functional roles in these pathways, and may in part explain the genetic foundation for the phenotypic variants. NOD2, which is associated with CD rather than UC, is heavily involved in NF-κB activation as well as autophagy and localization of bacteria in autophagolyso-somes. 44 IRGM (UC and CD) and ATG16L1 (CD) are also thought to be involved with autophagy, and therefore the ability to clear bacteria. 31 Another gene thought to be associated with bacterial clearance is IL23R. Interestingly, a mutation in the IL23R gene has been found to be protective against both CD and UC. 27 , 28 , 45 While it is unclear what impact these genetic variants have on the development and course of IBD, the correlations are striking and seem to corroborate prevailing pathogenic theories. 29
25.3.4 Persistent Specific Infection
Gastrointestinal infection can break the mucosal barrier, initiate inflammation, and stimulate innate immunity. Despite these insults, most people who develop an infection ultimately clear the infection without adverse sequelae. Nevertheless, there are several chronic infectious processes that have been associated with CD and, to a lesser extent, UC.
The first of these potential culprits is Mycobacterium avium subspecies paratuberculosis (MAP), which causes a granulomatous infection of the gastrointestinal tract in ruminants that is similar to the granulomatous inflammation of CD. 46 MAP may be encountered in contaminated drinking water, and has been implicated in IBD due to series reporting 55 and 22% incidences of MAP-positive blood cultures in CD and UC, respectively. 47 Additional credence to this theory was added after a series demonstrated that 52% of resected CD specimens contained MAP DNA. 48 In that study, MAP was present in approximately 2% of UC and 5% of control specimens, which argues against a link between MAP and UC. These results are similar to other series suggesting a much stronger link between MAP and CD than MAP and UC.
The other bacterial infection that has been associated with IBD is enteroadherent and invasive Escherichia coli, which has been recovered from 22% of mucosal biopsies from CD patients (compared to just 6% of controls). 49 Interestingly, CD patients with NOD2 polymorphisms and infection with this E. coli strain have decreased expression of TNF and IL-10, a phenotype consistent with ulceration, fistulas, and defective clearance of intracellular infections. 40 , 42 , 50 , 51 No studies have shown a link between E. coli and UC.
Another infection historically linked to IBD, albeit not UC specifically, was measles. This inference was based on post-World War II data from Sweden showing a higher incidence of IBD in people born within 3 months of a measles epidemic. 52 , 53 This finding has not been replicated in subsequent studies, and concern for IBD risk with live measles virus vaccines is unfounded based on available literature. 11
25.3.5 Dysbiosis
Alterations in the homeostasis of the gut microbiome can lead to a relative overgrowth of certain commensal organisms. E. coli, Bacteroides spp., Enterococcus, Klebsiella, and Clostridia have all been implicated in the inflammation of experimental colitis and IBD through this phenomenon, often referred to as dysbiosis. The process may be further exacerbated by the acquisition of virulence factors by these normal resident bacteria. 49 , 54
Diet is often felt to play a role in dysbiosis, and a dietary link could help explain the increased risk of IBD in Western societies. 11 While the validity of this theory remains open for debate, dietary components can certainly alter the macro- and microenvironment of the gut, potentially impacting the composition and virulence of commensal microorganisms. In fact, studies have shown that dietary iron potentiates the growth and virulence of intracellular organisms, while dietary aluminum stimulates the inflammatory response. 55 , 56 , 57 Furthermore, a paucity of colonic short chain fatty acids from a relative scarcity of Lactobacillus or Bifidobacteria may have detrimental effects on maintenance of an effective mucosal barrier. 58
Another potential contributor to dysbiosis is the so-called “hygiene hypothesis,” which suggests that protective immunity is stimulated by early exposure to pathogens and parasites. 29 Whether aware of this theory by name or not, parents who routinely allow their children to eat dirt and/or food off the floor with the thought that it will improve their immune system subscribe to this principle. The thought is that early development of protective immunity against low-virulence organisms prevents a later, more aggressive immunologic response due to a mature, robust immune system at first exposure. Some have suggested that dysbiosis caused by a lack of early pathogen exposure may be a contributing factor to increased rates of IBD in developed countries. The experimental improvement in symptoms with the delivery of pig whipworms to the gut in active UC and CD patients argues in support of this theory. 59 , 60
25.4 Modifiable Risk Factors
Given the high financial, psychological, and physiologic cost and chronicity of IBD, it is important to have a thorough understanding of modifiable risk factors so that we might mitigate the severity of illness. Cigarette smoking, oral contraceptives, and diet have all been theorized to impact the natural history and even development of IBD in the first place.
25.4.1 Smoking
In contrast to CD in which smoking is a risk factor for surgery and recurrence, paradoxically active smoking may impart some benefit in UC. 61 , 62 , 63 , 64 While controversial, given the widely publicized adverse effects of smoking on other body systems, smoking has consistently been linked with a decreased risk of UC. A 1989 meta-analysis concluded that smokers were 40% less likely to develop active UC as opposed to nonsmokers. 62 There have also been reports that smokers with UC are less likely to be hospitalized and have a colectomy when compared to nonsmokers. 61 The reasons for this protective effect are unclear, but are theorized to be related to the impact of nicotine on rectal blood flow, colonic mucus, cytokines, and eicosanoids. 65 , 66 Unfortunately, the beneficial impacts of smoking on UC are not long lasting and patients who smoke and subsequently quit tend to have more active disease, more hospitalizations, and an increased dependence on corticosteroids and immune modulators. 67 Furthermore, former smokers are 70% more likely to develop UC than those who never smoked. 62
25.4.2 Oral Contraceptives
Oral contraceptives have been studied in relation to IBD. While the data are not entirely clear, the bulk of the literature does seem to support an increased risk of developing either CD or UC in women who used oral contraceptives. 68 , 69 , 70 Whether the use of oral contraceptives has an impact on preexisting IBD remains unclear.
25.4.3 Diet
Perhaps the most studied dietary association with IBD is sugar; however, the results of the studies are mixed, with some series showing a relationship but others not. 71 , 72 , 73 , 74 Limited studies attempting to link fiber and fat intake to IBD have also yielded inconsistent results. 74 , 75 , 76 , 77 , 78 Given these varied results, it is not surprising that most studies regarding the impact of dietary modifications on IBD have been disappointing. 79 Recent studies out of Seattle Children’s, however, have yielded some promising results with the utilization of the “specific carbohydrate diet” (SCD) for treatment of CD and UC within the pediatric population. 80 While further studies are needed, their early results suggest that the dietary association is stronger than once thought, and may be utilized to improve symptoms if not induce remission.
25.4.4 Appendectomy
There are numerous other proposed risk factors for UC; however, one that has garnered considerable attention is appendectomy. Early retrospective data, to include a meta-analysis incorporating almost 3,600 cases and 4,600 controls, suggested a significant inverse association between appendectomy and UC, with a 69% decreased risk of UC in those who had previously undergone appendectomy. 81 Subsequent cohort studies have tempered these findings somewhat, although appendectomy for appendicitis does seem to confer a protective effect against the development of UC. This statistic is especially true in patients who undergo appendectomy before the age of 20 years. 82
25.5 Pathology
The most common manifestation of UC is circumferential confluent mucosal inflammation starting in the rectum and extending proximally, with the most substantial disease burden generally seen in the rectum. There may be an element of terminal ileitis in continuity with cecal inflammation, referred to as “backwash ileitis.” Patchy inflammation with “skip lesions” is not the norm, and the presence of this pattern should bias the clinician in favor of CD. Nevertheless, some degree of heterogeneity may be expected, especially in treated UC patients as the mucosa heals. A common finding in patients actively undergoing treatment for UC is an appearance of “rectal sparing.” 83 , 84 , 85 , 86 Kim et al 86 looked at sequential endoscopies in UC patients undergoing therapy and found that 59% of patients had patchy disease, rectal sparing, or both at some point during treatment. The specific pattern of heterogeneity was not associated with the type of therapy (including steroid use and rectal therapy). It is important to be cognizant of this healing pattern lest we mistakenly label all healing UC patients with CD. It is for this reason that the first colonoscopic evaluation for IBD is generally the most accurate for ascertaining UC versus CD.
The colorectal mucosa in early UC may grossly appear edematous, with confluent erythema and a loss of vascular markings. As the disease progresses, the mucosa will develop granularity with micropurulence and bleeding. Advanced UC will be marked by the characteristic pseudopolyp formation, deep (even full-thickness) ulcerations, gross purulence, mucosal bridging, and varied mucosal thickness (▶ Table 25.1). 87 , 88
There are no pathognomonic histopathological features of UC, which makes definitive diagnosis difficult even after the colon and rectum are excised. The diagnosis generally rests on the gross morphology combined with the absence of histopathologic findings more consistent with CD including noncaseating granulomas, vasculitis, and neuronal hyperplasia. Even the thickness of inflammation can be misleading, as severe cases of UC may have full-thickness inflammation. Despite these caveats, routine UC has a fairly predictable microscopic profile. Early disease is generally characterized by mucosal inflammation, crypt distortion, goblet cell mucin depletion, and vascular congestion. Neutrophil infiltration of the lamina propria in intermediate disease is associated with crypt abscesses and loss of mucosa with preservation of the crypts themselves. Crypt destruction, pseudopolyps, and deeper invasion of inflammation are the hallmarks of advanced disease. 88 It is in late disease where we tend to identify dysplasia; however, it can only be interpreted when identified in noninflamed bowel, as many dysplastic features are similar to inflammation (▶ Table 25.1). 89
It should be noted here that there is a population of colitis patients who do not classically fall in line with the routine presentation of either UC or CD. These patients may have rectal sparing disease with confluent colonic inflammation or confluent inflammation of the rectum and descending colon with nonspecific ulcerations elsewhere. Stated differently, these patients may have gross characteristics of both CD and UC. In the absence of pathognomonic histopathologic findings, it is exceedingly difficult to resolve the diagnosis. Rates of this indeterminate colitis are cited at 10 to 15%; however, it is likely even higher, as 4 to 10% of patients who undergo ileal pouch anal anastomosis (IPAA) “develop” CD in the pouch. 89 , 90 As such, all patients with a diagnosis of UC who are considering pouch reconstruction must be counseled with a reasonable degree of uncertainty, especially in cases that are not straightforward.
25.6 Clinical Manifestations
Signs and symptoms of a UC flare depend largely on the severity and location of disease. The most common findings are blood and mucus in the stools. Tenesmus, urgency, increased stool frequency, fecal incontinence, and pain with defecation are common, as the majority of the disease burden is generally in the rectum. Fecal incontinence is certainly the most distressing and anxiety-producing element of the disease, and as with tenesmus and urgency, is due to noncompliance of the rectum and loss of receptive relaxation. While diarrhea is common, up to 25% of patients may complain of constipation with a sense of incomplete evacuation. 89 Left lower quadrant abdominal pain, followed by flank and back pain is the norm as the disease progresses to involve the sigmoid and descending colon, respectively. Abdominal distention may also be present.
The severity of pain generally signifies severity of inflammation, and as the colonic lumen narrows with edema, the intensity of peristaltic pain increases. Nausea is extremely common, and likely reflects either downstream effects of circulating inflammatory mediators or inflammation of the stomach secondary to direct apposition with inflamed colon as opposed to frank obstruction. Weight loss is also common with chronic UC, which is likely a multifactorial phenomenon dependent on protein loss via the inflamed mucosa, avoidance of oral intake, and the metabolic demand of constant inflammation.
Some patients may smolder with advanced disease for some time before manifesting systemic signs of illness such as fevers and tachycardia. Others, however, may manifest these symptoms earlier in the process, possibly indicating a more severe disease phenotype. Regardless, development of these clinical features may be a harbinger of progression to life-threatening colitis, and warrants close attention and aggressive medical therapy. Certainly any further progression to peritonitis or evidence of severe colonic dilation is of grave concern, generally mandating urgent total abdominal colectomy.
Acute complications of UC include progression to life-threatening colitis, severe bleeding, and perforation, each of which is a manifestation of severe disease rather than an indicator of disease chronicity. The term toxic megacolon has largely been replaced, at least in concept if not the vernacular, by life-threatening colitis. This shift simply accounts for the fact that “toxic” patients need not have colonic dilatation. It likely reflects a severe UC phenotype, as 60% of patients who ultimately develop toxic colitis do so in the first 3 years, with half of those occurring in the first year. Severe bleeding may occur in up to 10% of patients over the course of their disease, with approximately 3% developing life-threatening hemorrhage necessitating colectomy. Acute perforation is another dire consequence of acute UC, with a mortality risk of up to 50%. 91 It may occur either with or without the presence of fulminant colitis. Perforation in the absence of fulminant disease generally occurs very early in the disease process, before the colon becomes scarred and resistant to perforation.
It should be noted that immune suppression might mask peritonitis. Furthermore, IBD patients are at risk for developing a superseding infectious colitis, usually Clostridium difficile, which may be the underlying etiology of the presumed disease exacerbation. 92 , 93 These two points must be at the forefront of the clinician’s thought process when evaluating and managing a patient with new-onset severe signs and symptoms.
25.7 Extraintestinal
Manifestations
Both UC and Crohn’s disease are marked by a variety of extraintestinal manifestations, the severity of which generally parallels the severity of colonic inflammation. Population-based studies report the overall incidence of extraintestinal disease at up to 40%, with 10% of patients having extraintestinal manifestations at the time of IBD diagnosis (▶ Table 25.2). 94 , 95 , 96 These numbers likely underestimate the true incidence, as patients may or may not link their extraintestinal symptoms with their disease, and as such may not report them.
Extraintestinal manifestation | |
Musculoskeletal | Peripheral arthropathy |
Axial arthropathy (ankylosing spondylitis) | |
Osteopenia | |
Osteoporosis | |
Ophthalmologic | Episcleritis |
Scleritis | |
Uveitis | |
Cutaneous | Erythema nodosum |
Pyoderma gangrenosum | |
Sweet’s syndrome | |
Aseptic abscess syndrome | |
Bowel-associated dermatosis–arthritis syndrome | |
Cutaneous vasculitides | |
Malignant melanoma | |
Hepatobiliary | Primary sclerosing cholangitis |
Hematologic | Deep venous thrombosis |
Pulmonary embolism | |
Mesenteric venous thrombosis (postoperative) |
25.7.1 Musculoskeletal
Arthropathies are a common finding in IBD, occurring in approximately 30% of patients. Peripheral arthropathy constitutes the vast majority of these; however, approximately 5% of IBD patients develop axial arthritis (ankylosing spondylitis). The joint pain can be debilitating, and seems to reflect the severity of colorectal inflammation. The importance of this correlation is significant, as a patient with severe urgency, frequent stooling, and difficulty moving due to substantial joint pain will almost undoubtedly become a prisoner of IBD due to fecal incontinence. In addition to arthropathy, IBD patients frequently develop osteopenia and osteoporosis. It is unknown if a primary link between IBD and bone loss exists, as chronic corticosteroid use and deconditioning are both substantial risk factors for bone disease. Regardless, bone disease represents a significant morbidity, and likely explains the 40% increased risk of bone fractures in IBD. 94
25.7.2 Ophthalmologic
Episcleritis, scleritis, and uveitis impact approximately 2 to 5% of patients with UC. Episcleritis manifests as “red eye” and generally does not cause pain or visual disturbance. Unlike other ophthalmologic complications, inflammation of the episclera generally parallels colorectal inflammation. Episcleritis lies in stark contrast to scleritis, which causes severe pain and visual impairment. Uveitis is more common in women with IBD, is associated with arthritis, and causes pain and redness. Neither scleritis nor uveitis appear to reflect the degree of colorectal inflammation, and in fact may be present even in the absence of gastrointestinal symptoms. 97 , 98
25.7.3 Cutaneous
IBD patients are predisposed to a variety of skin conditions. It is important to document these, as many of the medications used to treat IBD can themselves have dermatologic complications and therefore confound the treatment algorithm. Erythema nodosum is the most common inflammatory skin condition associated with IBD; it occurs in 3 to 10% of UC patients. These lesions tend to occur on the extensor surfaces of extremities, especially the anterior tibia, and consist of 1- to 5-mm raised, tender, red or violet subcutaneous nodules. 99 Pyoderma gangrenosum is the second most common skin manifestation in IBD, although it is still a rare condition occurring in only 0.75% of patients. These lesions most commonly occur on the legs but may occur anywhere and have been found adjacent to ileostomies and wounds. 99 , 100 They appear as single or multiple erythematous papules or pustules, which are often preceded by trauma to the skin. Dermal necrosis then leads to deep ulcerations containing sterile purulence. Up to 50% of patients who present with pyoderma gangrenosum have IBD. 101 Erythema nodosum generally parallels gut inflammation, while pyoderma gangrenosum does not. Neutrophilic dermatoses such as Sweet’s syndrome, aseptic abscess syndrome, and bowel-associated dermatosis–arthritis syndrome, and cutaneous vasculitides may also occur in IBD, although they are quite rare. IBD may also be a risk factor for malignant melanoma. 102
25.7.4 Hepatobiliary
Both UC and CD are associated with primary sclerosing cholangitis (PSC), fatty liver, and autoimmune liver disease. 89 PSC occurs in approximately 3% of UC patients, although UC may be present in up to 73% of PSC cases. 103 The risk of both colorectal and hepatobiliary malignancies is thought to be increased in UC patients with PSC. 104
25.7.5 Hematologic
Patients with IBD have a twofold to threefold increased risk of deep venous thrombosis (DVT) and pulmonary embolism (PE) when compared to the general population. 105 The incidence of DVT within this population is 30.7 per 10,000 person-years and is similar between UC and CD (30.0 for UC and 31.4 for CD). Risk of PE is higher in UC than in CD, with incidences of 19.8 and 10.3 per 10,000 person-years, respectively. 106 Patients younger than 40 years suffering from active inflammation and more extensive colonic disease have the highest baseline risk within this group. 106 , 107 The phenomenon cannot be explained by surgery or inflammation alone, as the risk in IBD exceeds that in non-IBD surgery (other than oncologic), rheumatoid arthritis, or celiac disease (other inflammatory conditions). 108 The entire etiology for the unique intrinsic coagulopathy in IBD has not been established; however, the inflammation associated with IBD can certainly shift the hematologic homeostasis toward a prothrombotic state. Implicated in this process are increased plasma levels of thrombotic factors, several of which are acute-phase reactants; decreased levels of natural anticoagulants; reduced fibrinolytic activity; endothelial damage and dysfunction leading to downregulation of the anticoagulant thrombomodulin and endothelial protein C receptor, ultimately affecting the conversion of protein C into its activated form; and an inherent thrombophilia. 109 These factors likely contribute to the 3 to 5% risk of mesenteric venous thrombosis, a feared complication of colectomy in IBD. Initial reports of this complication were associated with mortality as high as 50%; however, heightened awareness in recent years has led to more frequent diagnosis, earlier intervention with anticoagulation, and lower overall morbidity and mortality. 89 , 110 , 111
25.8 Classification
The most utilized clinical assessment tool for UC was described by Truelove and Witts in 1954 (▶ Table 25.3). 112 They described disease as either mild or severe based on bowel frequency, blood in the stool, fever, heart rate, hemoglobin, and erythrocyte sedimentation rate (ESR). Initially used to monitor the impact of cortisone on disease, the simplicity and relevance of the model have allowed some variation of it to be used in nearly all modern grading systems. Modern variations of this system such as the Montreal and Mayo classifications include “moderate” categories as well as colonoscopic appearance, but otherwise use similar clinical criteria. 113 , 114 The Montreal Working Party also felt it necessary to include a parallel classification system stratifying UC by extent of disease, with separate categories for isolated proctitis, left-sided disease, and pancolitis. 113
Mild | Severe | |
Fever | Absent | > 37.5 |
Heart rate | < 90 | > 90 |
Blood in stool | + | + + + |
Stool frequency | < 4 | > 6 |
Erythrocyte sedimentation rate | < 30 | > 30 |
Hemoglobin | > 75% normal | < 75% normal |
Source: Koltun 2011. 89 |
25.9 Natural History
The most common manifestations of UC are blood and mucus in the stool, although other signs and symptoms may be present depending on the location and severity of inflammation. As previously discussed, the involvement of the rectum generally leads to frequent bowel movements, tenesmus, urgency, a sensation of inadequate evacuation, and fecal incontinence. The latter is generally a progressive problem, worsening as the rectum scars and loses capacitance. Abdominal and back pain may be present as well, depending on the presence and severity of inflammation in the intra-abdominal and retroperitoneal colon.
In order to understand the typical disease progression, we must understand the historical context, to include patient outcomes prior to the development of adequate therapies. The “natural history” of UC was altered dramatically by the widespread use of corticosteroids for UC, which was spawned by the landmark work of Truelove and Witts in 1954. 112 Prior to this time, patients were at the whim of their disease process, as few treatment options were available. In fact, limited series from the early 1930s suggested that as many as 75% of patients died within 1 year of an acute presentation. 115 A larger subsequent series by Edwards and Truelove clearly identified a dichotomy between outcomes of UC flares before and after 1954. That study reported outcomes in 624 patients treated for acute UC exacerbations between 1938 and 1962, and identified an overall mortality rate of 14% prior to 1953 versus 4% thereafter. This included a 34% mortality rate among those referred for severe colitis, which was significant in that 19% of patients had severe disease at the initial attack, and 18% of patients would have a severe attack at some point. 116 The 1954 pivotal study itself compared mortality rates in patients with severe UC who received either cortisone or placebo, and found a significant mortality reduction from 24% in the placebo group to 7% among patients who received cortisone. 112 Despite similar rates of steroid responsiveness in the decades since the initial studies, mortality rates have continued to downtrend and currently reside at approximately 1%. 117 These improved outcomes are presumably due to additional rescue therapies, improved supportive care, and recognition of the importance of surgical intervention in medication-refractory cases. 102
Modern epidemiologic data suggest that the first presentation of UC is usually for an attack of mild severity (73%), although more than one in four patients will initially manifest moderate (27%) or severe disease (1%). 18 , 118 , 119 Duration of an UC “flare” is highly variable, ranging from several weeks to several months, with a small contingent developing intractable disease. This range is independent of treatment, as patients have varying response profiles to the therapeutic options.
Langholz et al 119 published a series of 1,161 UC patients managed from 1962 to 1987 who were followed from diagnosis up to 25 years. They found that the probability of progression to a relapsing course was approximately 90%. Active disease in any given year was predicted by active disease in the preceding year, indicating that sick patients tend to stay sick. In years 3 to 7 after diagnosis, 25% of patients were in remission, 18% had active disease in all years, and 57% had intermittent relapses. Twenty-four percent of patients had a colectomy within the first 10 years. While this study is outdated, and even at the time included patients from several generations of medical therapy, it shows a baseline natural history in patients largely managed by steroid pulses alone. 18
A more recent population-based study by Solberg et al 120 followed 423 UC patients for 10 years (1990–1994 to 2000–2004). Similar to the Langholz study, Solberg et al 120 found that a large percentage of patients develop relapsing disease (83%), although approximately half the patients were relapse free over the last 5 years of the study. The most notable finding in this study was a significantly decreased 10-year rate of colectomy. The 9.8% colectomy rate in the 1990s marked a 58% decrease in colectomy when compared to the years between 1962 and 1987. This change likely reflects advancements in medical management, most notably the widespread use of aminosalicylates and immune-modulating medications. The study was cut off in the year infliximab was approved for use in UC, so biologic therapy likely did not impact the results of this study.
A number of patient and environmental factors have been associated with relapse. Among these are younger age at diagnosis, female gender, and more extensive disease at presentation. Additionally, frequent relapses tend to beget more relapses, again emphasizing that healthy patients stay healthy and sick patients stay sick. Stress has also been linked to relapses, which can lead to a vicious cycle of worsening disease with stress causing relapses, and relapses inducing more relapses. As previously discussed, active smoking and potentially appendectomy confer some degree of protection from relapse. Interestingly, systemic symptoms such as fever and weight loss are associated with fewer relapses, although this is presumably due to a higher likelihood of early colectomy. 14 , 67 , 118 , 120 , 121 , 122 , 123 , 124 , 125
As UC is a dynamic disease, it is important to consider not only the relapse itself, but also the progression of disease extent over time. Gower-Rousseau et al observed a series of pediatric patients with UC and found that 28% of patients have isolated proctitis, 35% left-sided colitis, and 37% extensive colitis at initial presentation. In this cohort, the disease course was characterized by extension of inflammation in 49% of patients. 126 This is not dissimilar to other studies identifying 20 to 53% proximal extension rates, depending on the initial segment involved. 119 , 120 , 127 Patients with proctitis have a ~ 50% chance of extension, and those with disease proximal to the sigmoid colon have a 9% chance of progression to pancolitis. Interestingly, regression may also be observed when the disease is followed over time. In fact, the probability of regression has been reported as high as 76.8% for substantial colitis and 75.7% for pancolitis after 25 years. 119 , 120 , 127
Stricture and neoplasia are feared sequelae of chronic UC inflammation. In 1966, De Dombal et al 9 reviewed their 10-year UC data and found that 11% of UC patients had a stricture within the study period. The majority of these strictures were located in the rectum, which was not surprising as the majority of the patients had rectal disease. Stricture rates in those with panproctocolitis were over 17%, while those with left-sided and isolated rectal disease had 7.5 and 3.6% stricture rates, respectively. Interestingly, the stricture rate did not seem to correlate with duration of disease, as a number of patients had strictures on initial presentation. 9 Stricture rates since the De Dombal study have been reported from 3 to 11%. 128 , 129 , 130 The underlying issue is not, however, the stricture per se, but rather the underlying dysplasia or neoplasia. 129 , 130 While a minority of strictures will have malignant cells on endoscopic biopsy, approximately 40% will return with carcinoma in the resected specimen and an additional 33% will have high-grade dysplasia. 131
Meta-analysis of premillennial UC data suggested that the risk of colorectal carcinoma among UC patients is approximately 3.7%, with pancolitis conferring an even higher risk of 5.4%. The same study found the cumulative risk of colorectal carcinoma by decade was 2% at 10 years, 8% at 20 years, and 18% at 30 years. 132 These findings underlie the recommendation for endoscopic surveillance starting after 8 years of disease duration in pancolitis. More recent studies, however, suggest much lower probabilities, with cumulative risks in the new millennium being reported at 1, 2, and 5% after 10, 20, and more than 20 years of disease duration, respectively. 133 The decreased risk in recent studies likely reflects either the benefits of rigorous endoscopic surveillance or overall improved disease management rather than a shift in disease phenotype. 134 , 135
Whether or not the most recent therapeutic advancements are impacting the natural history of UC remains unknown. Infliximab was the first biologic agent added to the UC armamentarium in 2004. Since then, several relatively small, randomized controlled trials have shown a decrease in early colectomy rates among UC patients started on infliximab. 136 , 137 Despite these favorable trends, any beneficial effect in chronic, refractory UC has been called into question. 138 Unfortunately, no significant long-term epidemiological studies are available to address the impacts of the most modern iteration of medical management.
25.10 Diagnosis
25.10.1 Clinical History
Prior to strategizing a diagnostic approach to any patient complaining of abdominal pain and altered bowel habits, it is important to have a thorough understanding of the differential diagnosis. Thinking in terms of the larger picture will prevent “tunnel vision” with regard to medical and surgical decision making.
CD impacting the colon may present in similar fashion to UC. In fact, it may be difficult to distinguish the two disease processes. While the medical management does not differ significantly, differentiation between the forms of IBD is critical for management of patient expectations and surgical planning. Crohn’s colitis constitutes approximately 30% of CD cases, and most commonly presents with diarrhea and bleeding similar to UC. Patchy “skip” lesions and rectal sparing are commonly found in CD, and may help differentiate the processes; however, treated UC may present with a heterogeneous appearance as the mucosa heals in a disorganized fashion. 86 While “backwash ileitis” may be present in UC patients with pancolitis, any other evidence of small bowel disease, intestinal fistulas, perianal disease, or hallmark histopathologic findings of CD point to a diagnosis of CD. Determining a definitive diagnosis is exceedingly difficult in the absence of these tell-tale findings, and patients should be counseled regarding the degree of uncertainty.
Infectious colitis or proctitis must be excluded prior to immunosuppressive treatment for a UC flare. Not only can these infections present in similar fashion to UC, but they may also occur more frequently in IBD due to recurrent hospitalizations, administration of immunomodulatory and antimicrobial agents that disturb the intestinal flora, and a decreased nutritional status. These risk factors are especially important for C. difficile infections, which are present in 5 to 19% of patients admitted with an IBD exacerbation. 92 , 93 Other bacteria that may cause inflammatory diarrhea include Shigella spp., enterohemorrhagic or enteroinvasive E. coli, Campylobacter jejuni, Salmonella spp., Yersinia enterocolitica, M. tuberculosis, Vibrio parahaemolyticus, and Chlamydia trachomatis. Parasitic infections may also mimic a UC flare, and include Entamoeba histolytica, Schistosoma spp., Balantidium coli, and Trichinella spiralis. Cytomegalovirus (CMV) is a fairly ubiquitous pathogenic virus with low virulence. In the immune suppressed, however, it may produce “CMV colitis,” which can worsen rapidly and sometimes necessitate urgent colectomy. Different pathogens are responsible for isolated proctitis. These include sexually transmitted diseases such as Neisseria gonorrhoeae, Herpes simplex virus, and C. trachomatis. Other proctitis causing organisms are Treponema pallidum and CMV. 139 A thorough review of infectious diarrhea is outside the scope of this chapter; however, the importance of ruling out such an infection before starting pulse immunosuppression cannot be overstated. At a minimum, all patients with an inflammatory or infectious colitis require laboratory evaluation of the stool to include stool culture, fecal leukocytes, fecal ova, and parasites, and the presence of C. difficile toxin. In patients who are immune suppressed, it is prudent to take colonic mucosal biopsies to rule out CMV colitis.
A number of other conditions can mimic the clinical profile of UC. These include lymphoma, chronic mesenteric ischemia, radiation colitis, diversion colitis, solitary rectal ulcer syndrome, graft versus host disease, diverticular colitis, and medication-associated colitis. In addition to clinical history, the diagnosis of UC requires direct visualization of the colon via endoscopy, radiographic imaging studies, and evaluation for serologic markers specific to IBD or UC.
25.10.2 Endoscopy
Endoscopy is the gold standard for diagnosis of UC, as the entire extent of the disease process can effectively be surveyed. Additional imaging modalities only need be utilized if the clinical history is such that additional diagnoses, such as CD, need to be entertained. Colonoscopy allows for tissue biopsy as well as assessment of the gross appearance and specific pattern of inflammation. Complications of IBD such as strictures, fistulas, bleeding, or neoplasm may be readily identified. Colonoscopic examination alone is generally enough to diagnose UC in the right clinical setting, especially if tissue biopsies return without sarcoid-type giant cell granulomas, which would indicate CD. Colonoscopic findings consistent with mild UC include edema, confluent erythema, and loss of vascular markings, generally starting in the rectum. Moderate disease is marked by mucosal granularity, bleeding, and micropurulence, and may show further proximal extension of disease. Severe disease may be noted by ulcerations with deep penetration into the bowel wall, pseudopolyp formation, frank purulence, variable thinning and thickening of the mucosa, strictures, and mucosal bridging. 88 , 89 Deep ulceration at colonoscopy predicts worse outcomes and higher risk of surgery. 140 The findings in severe disease may have significant overlap with CD, which makes definitive diagnosis difficult. Another caveat to the “classic” findings is that the confluent erythema emanating from the rectum may be absent in a UC patient with healing disease. As previously discussed, a significant number of UC patients on treatment will appear to have rectal-sparing and/or patchy disease. 83 , 86
The decision to perform endoscopy in the setting of severe disease should not be taken lightly. Despite studies reporting its safety, the general feeling is that the risk of perforation is increased. Nevertheless, patients with intractable, severe, disease will require tissue biopsies and stool samples to exclude concurrent infection. Rigid or flexible proctoscopy is generally undertaken in these situations, with biopsies taken below the peritoneal reflection. 89
25.10.3 Imaging
Plain Film X-Rays
X-rays still play a role when considering the patient who presents acutely with abdominal pain. Upright abdominal films will reveal the presence of free air or “toxic megacolon,” a term that has all but fallen out of favor due to the fact that patients can be “toxic” without having “megacolon.” Nevertheless, colonic dilatation ≥ 6 cm or cecal diameter greater than 9 cm in the setting of fever, tachycardia, and abdominal pain is a grave concern that mandates urgent surgical intervention. 141 , 142 More subtle findings on plain films include nodularity of the mucosa suggesting the presence of pseudopolyps, mucosal edema (“thumbprinting”), and an ahaustral, “lead-pipe” colon, which is suggestive of chronic disease.
Contrast Enema
Double-contrast barium enema is not required for the diagnosis of CD; however, it is useful to help exclude CD when there is any question of small bowel involvement. The specific benefit of rectal contrast over small bowel follow-through is that it more clearly delineates the terminal ileum as there is less interference from proximal intestinal loops.
Enteroclysis
Enteroclysis is preferred over small bowel follow-through for detailed assessment of the small intestine. In fact, this technique is more sensitive than computed tomography (CT) for the detection of small bowel fistulas and early mucosal abnormalities. 143 Nevertheless, this study is not widely utilized as it is labor intensive and requires nasointestinal tube placement, which is never desirable for the patient. The study entails placement of a postpyloric nasal tube, followed by repetitive small boluses of barium contrast, which will coat the bowel wall. Each bolus is followed by insufflation of air to distend the small bowel, and a spot film is taken. The additional detail offered by this technique is not generally worth the added time and labor. As such, if a contrast study is obtained to evaluate the small bowel, it is generally either standard small bowel follow-through or CT.
Computed Tomography
CT of the abdomen and pelvis is the most commonly obtained radiographic study in the acute evaluation of IBD. As with plain films, the role of CT in UC is negligible when the diagnosis of UC is secure. The benefit of CT is twofold: first, it has the ability to evaluate all intra-abdominal organs simultaneously and, second, it can evaluate the thickness of the entire intestinal wall. It must be emphasized here that a CT scan to evaluate the bowel wall specifically is severely limited if intraluminal contrast is not used. We therefore recommend a full oral contrast load with consideration of rectal contrast if a CT scan is to be performed for any diagnostic study in IBD. In UC specifically, signs that may be identified on CT include increased perirectal and presacral fat, heterogeneous colonic thickening, target or “double-halo” sign of the colon, and strictures. 144
Magnetic Resonance
The role of magnetic resonance imaging (MRI) in IBD is still evolving. While less efficient and more costly than CT for an abdominal survey, 143 its ability to objectively quantify severity of inflammation in the bowel and mesentery may hold promise for monitoring response to medical therapy. 145 , 146 It may also have a unique role in characterizing perianal fistulas in CD, although this technology would have limited use in UC unless the diagnosis is in question. In addition, the lack of radiation exposure is appealing.
Ultrasound
Ultrasonography may be used as an adjunctive measure to assess response to therapy in IBD patients. A transabdominal wall ultrasound (TABS) can assess for strictures or other bowel wall thickening. Some functional data can be obtained as well, as peristalsis can be observed. 147 The technique is obviously operator dependent, and the data obtained are nonspecific, which may account for the lack of enthusiasm for TABS in the United States.
Nuclear Medicine
Nuclear imaging techniques have shown some promise in distinguishing CD from UC, and may be utilized to assess response to therapy as well. These studies rely on injection of the patient’s own radionuclide-labeled white blood cells, and have been shown to accurately delineate the anatomic extent of inflammation. Newer techniques use Technetium-99 m haxamethyl-propylamine-oxime (HMPAO), which has improved sensitivity over older, Indium-111 labeling. While the specificity is inadequate to differentiate infectious versus idiopathic or autoimmune inflammation, studies have shown 93 to 98% accuracy in differentiating CD from UC. 89 , 148 , 149 These findings are based on the differential anatomic patterns of inflammation between CD and UC, so it may be that nuclear imaging provides little if any information that is not readily obtained by more traditional techniques. Given its high sensitivity for identifying inflammation, nuclear medicine may have a future role in assessing response to therapy. Whether this methodology offers any benefit beyond standard MRI has yet to be determined.
25.10.4 Capsule Endoscopy
Wireless capsule endoscopy is an additional imaging modality used in IBD. The primary benefit of this technique is that it allows visualization of the small bowel when CD is suspected. 150 There is no role for this technology in classic cases of UC, although it provides additional diagnostic information in indeterminate cases.
25.10.5 Serologic Testing
A number of serum tests may be obtained in IBD in order to assess the degree of inflammation and aid in diagnosis. C-reactive protein (CRP) and ESR are the only two widely used tests for assessing general systemic inflammation. Neither measurement is sensitive or specific for IBD, but they may give an idea as to the level of inflammation in an IBD flare if other causes of inflammation are ruled out. ESR is felt to correlate more with colitis than small bowel disease irrespective of CD versus UC. The results of ESR and CRP may be misleading, as elevation can indicate anything from a perirectal abscess to pneumonia. A normal level can also be somewhat misleading, as the degree of inflammation may not correlate with a patient’s symptoms. Such a scenario could involve a symptomatic stricture without significant active inflammation. 151
An ever-increasing number of IBD-associated serologic markers have been identified. Among these are antineutrophil cytoplasmic antibodies (pANCA) and anti-Saccharomyces cerevisiae antibodies (ASCA), which were the first two widely used markers. It is felt that the presence or absence of these markers may help differentiate CD from UC in cases that are clinically in question. pANCA is more frequently identified in UC patients (50–70% of UC vs. 20–30% of CD) and may indicate a more medically refractory phenotype. 152 ASCA is more frequently associated with CD, although it may be present in 10 to 15% of UC patients as well. Other serologic markers such as anti-outer membrane porin C (anti-OmpC), anti-CBir1, anti-A4-Fla2, and anti-Fla-X have since been associated with IBD and are in clinical use; however, their utility in diagnosis is not as well established. In addition to aiding in diagnosis of IBD, both pANCA and antiCBir1 have been found to predict the risk of pouch complications after IPAA. 153
25.10.6 Genetic Studies
A number of genetic markers have been identified that may help differentiate CD from UC. Out of 200 known DNA loci associated with IBD, at least 23 are associated with UC alone. 30 , 32 ATG16L1, ECM1, NKX2-3, and STAT3 are currently in clinical use, although of those only ECM1 is UC specific. There is a test on the market that combines these four genetic markers with several associated serologic and inflammatory markers. Reliance on such technology is variable, and at this point it is unclear what additional diagnostic benefit is gained from such testing. However, it likely has a future role in tailoring not only medication therapy, but also the decision to proceed with proctocolectomy with IPAA versus end ileostomy in medically refractory UC patients.
25.11 Medical Management
As with CD, UC is a chronic and debilitating disease with no definitive curative nonsurgical options. This realization alone is both humbling for the provider and terrifying for patients. As such, it is preferable to hold discussions with UC patients in a multidisciplinary setting, where the gastroenterologist and surgeon can establish cohesive therapeutic goals with specific input from a well-counseled patient.
The medical management of UC is predicated on a combination of disease distribution, disease severity, and clinical response to treatment (▶ Table 25.4). Unpredictable responses to individual medication classes mean that management is a dynamic process, constantly being tailored to a patient’s disease phenotype. In addition to the fluidity built into treatment algorithms themselves, there are different philosophies with regard to upfront, aggressive management with biologic agents (“top-down” approach) versus a “step-up” approach for severe UC. Unfortunately, despite all algorithms and best intentions, it may appear that we are pursuing trial-and-error medicine in patients who ultimately have medically refractory UC, which is not far from the truth. With each intervention requiring 6 to 17 weeks to be deemed a failure, this approach may be particularly costly to the patient’s lifestyle and trust in the medical community. It is therefore imperative that a therapeutic relationship be developed early on, with a thorough explanation of the potential limitations of medical management.
25.11.1 Induction of Remission
Proctitis
Mild to moderate UC confined to the rectum is generally treated by topical therapy alone, although oral adjuncts may be used as well. Mesalamine (5-ASA) and steroids are the mainstays of therapy; both indications are available in a variety of formulations. Suppositories are the preferred method of delivery for isolated proctitis as they are easier to administer, leak less, and have been shown to target the site of inflammation more effectively than liquid or foam enemas. 154 , 155 Randomized controlled trials suggest that topical mesalamine is more effective than monotherapy with topical steroids, oral mesalamine, or placebo at inducing remission, with 63 to 94% of patients experiencing a clinical response within 4 to 6 weeks. 102 , 156 , 157 , 158 , 159 , 160 The dosage administered should be 1 to 2 g daily, although there is no increased dose response above 1 g. Divided daily doses are no more effective than once daily. 161 , 162
Topical steroids are four to five times more likely to induce remission than placebo, 163 although they are considered second-line therapy due to randomized controlled trials showing inferiority to topical mesalamine. 156 , 157 , 158 , 159 , 160 Nevertheless, topical steroids continue to play a role in those who do not tolerate mesalamine or in patients who are refractory to topical mesalamine monotherapy. Formulations and dosing are variable in the studies, which may account for the wide-ranging response rate of 35 to 70% at 4 weeks. 156 , 157 , 158
If patients are not responsive to topical monotherapy, combination therapy with topical mesalamine and budesonide may be more efficacious than either therapy alone. A randomized controlled trial comparing topical mesalamine (2 g/d) plus topical budesonide (3 g/d) to either therapy alone noted clinical, endoscopic, and histologic improvement at 4 weeks in 100, 76, and 71% for combination therapy, mesalamine, and budesonide, respectively. Endoscopic remission followed a similar trend favoring combination therapy, with 37, 30, and 10% of patients showing complete endoscopic healing at 4 weeks. 156
While not as effective as topical mesalamine, oral mesalamine may be utilized as monotherapy or in combination with other agents for distal UC. When used as monotherapy, a 4-g daily dose of pH-dependent release oral mesalamine is more effective than placebo or lower doses. 154 , 164 , 165 , 166 The combination of oral (2.4 g/d) and topical mesalamine, however, was found to be more efficacious than either application alone in patients with disease progression halting within 50 cm of the anal verge. 167 , 168
Ultimately, patients with mild to moderate ulcerative proctitis should be started on mesalamine suppositories. Combination therapies with topical steroid and/or oral mesalamine should be considered in refractory patients. If inflammation is ongoing, systemic therapy with oral steroids is indicated in order to induce remission. It is this final group of patients that may go on to require immune modulation, biologic therapy, or surgery if the rectal inflammation cannot be controlled. 102
Left-Sided Colitis
The treatment algorithm for mild to moderate left-sided UC is similar to that for ulcerative proctitis, with several minor distinctions. First-line therapy consists of combination therapy with oral (2–4 g/d) and topical mesalamine, which was found to have improved disease activity indices (–5.2 for combination therapy; −4.4 for topical mesalamine; −3.9 for oral mesalamine) and clinical improvement at 8 weeks (86% for combination therapy vs. 68% for monotherapy) for left-sided disease. 102 , 168 , 169 Remission rates at 8 weeks were similarly improved by combination therapy, with 64% remission in those on combination therapy versus 43% of those on oral mesalamine only. 169 The argument for this combination is strengthened when data from the treatment of more extensive UC are extrapolated to left-sided UC. 170
The bulk of the literature suggests the superiority of topical 5-ASA over topical steroids for left-sided UC, although the data are not as strong as they are for isolated proctitis. 163 , 171 , 172 Topical steroids are therefore a reasonable second line, especially in the case of 5-ASA hypersensitivity (bleeding and urgency 3–5 days after starting therapy; resolves in 72 hours with discontinuation of 5-ASA). Additionally, combining steroid and mesalamine topical therapy with oral mesalamine in cases refractory to oral and topical mesalamine alone may result in clinical improvement. While no comparative studies of this triple therapy exist, there is evidence to suggest that dual topical therapy results in improved remission rates when compared to either therapy alone. 156 Regardless of whether mesalamine or steroids are being used, topical agents should be delivered via enema (foam or liquid) as opposed to suppository form. 102 Neither foam nor liquid enemas have shown superiority over the other. 173 Low- and high-volume doses also show equivalent efficacy, although the low-volume dose is better tolerated. 174
Although oral mesalamine alone is better tolerated than is oral sulfasalazine, it is no more effective. 175 As described earlier, a single daily 2- to 4-g dose of pH-dependent release oral mesalamine is more effective than placebo when used as monotherapy for left-sided colitis. 164 , 166 While the literature is somewhat mixed, most agree that any dose-dependent effect is negligible at doses over 2 g/d. 154 , 165 , 166 Combined analysis of the studies limited to pH-dependent release oral mesalamine has shown 8-week remission rates of approximately 40%, which marks a 100% increase in efficacy over earlier formulations. 166 Once daily dosing with 3-g oral mesalamine has been associated with improved remission rates when compared to 3-g divided daily dosing (86 vs. 73%). 176 Once-daily oral mesalamine dosing has not been comparatively evaluated as part of combined modality therapy.
While most patients with left-sided UC will experience symptomatic improvement within 2 weeks, it may take up to 16 weeks for mesalamine responders to achieve remission. 177 , 178 , 179 , 180 , 181 , 182 Within this time frame, the patient will have ongoing symptoms to some degree, and may not respond at all. As such, a discussion regarding the timing for escalation of therapy often takes place prior to declaration of failure of first-line treatments. Oral steroids are the next line of therapy, and induce remission in the majority of patients. 102 , 183 , 184 Unfortunately they are not without complications, and an extensive discussion should be undertaken in which the patient is thoroughly counseled regarding the desired time-to-response versus steroid-induced side effects. Regardless, oral corticosteroids are likely needed in cases of clinical deterioration, continued rectal bleeding beyond 14 days, and inability to achieve remission after 40 days of appropriate mesalamine and/or topical steroid therapy. 102
Extensive Ulcerative Colitis
Extensive UC, also referred to as pancolitis, refers to inflammation extending proximal to the splenic flexure. This disease pattern requires oral therapy as the proximal extent of the inflammation cannot effectively be reached by topical agents. It should be noted that the data driving the treatment of both left-sided colitis and extensive colitis are largely drawn from the same studies, which include a heterogeneous patient population. As such, much of the treatment of mild-to-moderate extensive colitis has already been reviewed.
First-line therapy is a single daily dose of 2- to 4-g oral mesalamine in addition to a 1-g mesalamine enema. This combined therapy effectively targets both the proximal and distal extents of the disease, and results in an 8-week remission rate of 63%. 102 , 168 , 169 , 175 , 180 , 185 Topical steroids are therefore a reasonable alternative to topical mesalamine, especially in the case of 5-ASA hypersensitivity.
Patients who do not respond to first-line therapy, or who have a relapse of disease while on maintenance therapy, require oral steroids. Truelove and Witts initially reported the benefit of cortisone for the treatment of UC in 1954. 112 A follow-up study used a regimen of a 20-mg divided daily dose of prednisolone in addition to topical hydrocortisone. The remission rate in this group of patients was 76%, which represented a significant improvement over the 52% seen with sulfasalazine alone. 184 The oral steroid regimen for induction of remission is 40 to 60 mg of prednisone or prednisolone daily. These dosages are based largely on a 1962 study by Baron et al, which showed a 50% revision rate for patients with pancolitis regardless of whether they received 60- or 40-mg daily prednisone doses. 186 Most prefer a 40-mg dose, as the side-effect profile worsens with little or no additional therapeutic benefit above this dose. 186
Given the considerable side-effect profile, steroids are not intended as maintenance therapy. As such, a reasonable taper must be implemented such that steroid dependence is recognized early, but not at the cost of early relapse. Steroid courses less than 3 weeks in duration have been associated with early relapses. As such, an 8-week course starting with 40-mg prednisolone per day for 1 week and reducing by 5 mg/d every week is recommended by the most recent European consensus. 102
Budesonide is an alternative oral steroid with limited systemic effects due to first pass metabolism. While attractive in concept due to the lack of side effects, numerous studies have failed to identify any benefit of budesonide over mesalamine. A newer formulation with a colonic release mechanism (budesonide multimatrix [MMX]) has improved remission rates over placebo (15 vs. 7%), and is most effective in left-sided patients who are mesalamine responsive. Nevertheless, a recent Cochrane review found no definitive data to suggest a benefit of budesonide MMX over standard budesonide, much less over mesalamine. 187
Severe Ulcerative Colitis
Severe UC is defined as at least six bloody stools daily, abdominal tenderness with signs of systemic toxicity including fever (> 37.5 °C), tachycardia (> 90 bpm), anemia (hgb < 75% normal), and increased ESR (> 30 mm/h). 188 Historically, these cases carried a high mortality burden, with up to 75% of patients dying within the first year of a severe flare. The era of corticosteroid salvage was ushered in by the landmark study of Truelove and Witts in 1954; outcomes have improved dramatically since that time. Nevertheless, a significant number of severely inflamed patients progress to fulminant, life-threatening colitis.
While the clinical urgency of a toxic patient with presumed fulminant UC must be acknowledged, uninformed decision making can be potentially catastrophic. In the moment, it is of dire importance that intestinal infection be ruled out or at least treated empirically before starting immune suppression. Patients with inflammatory bowel disease (IBD) are at increased risk for C. difficile and CMV infections for a variety of reasons, and failure to address these as potential confounders, if not primary etiologies, could mean the difference between life and death. 92 , 93 , 189 , 190 , 191 , 192 , 193 , 194 Clostridial infection may easily be determined by stool studies, which should be performed on admission for any UC flare. Biopsies are needed to effectively rule out CMV colitis. As such, an unprepared, limited flexible sigmoidoscopy with biopsies should be undertaken in truly refractory cases that do not require urgent surgical intervention. While the decision to perform endoscopy in the face of acute inflammation is never taken lightly, the information obtained could be of vital importance. Infected patients should be treated with appropriate antimicrobials, and consideration should be given to withholding immunosuppression (when able). 194 , 195 In addition to excluding or treating infection, a number of other measures should be undertaken to ensure the patient is medically “optimized” 102 :
Liberal fluid resuscitation and correction of electrolytes should be undertaken, as hypokalemia and hypomagnesemia can promote toxic colonic dilation. 91
DVT chemoprophylaxis should be administered, as patients with IBD are at increased risk for venous thromboembolism compared to controls. The risk further increases during an acute flare independent of other risk factors. 196 , 197
Nutritional support should be undertaken, particularly in patients who suffer from malnutrition at baseline. Enteral feeding is preferable to parenteral, as bowel rest with intravenous (IV) nutrition has not shown any benefit and is associated with significantly more complications (35 vs. 9%). 198 , 199
Narcotics, anticholinergics, antidiarrheals, and nonsteroidal anti-inflammatories should be limited as they may increase the risk of progression to toxic colonic dilatation. 91 , 200 , 201
Topical mesalamine or corticosteroids may be given if tolerated; however, there are no comparative studies to suggest a benefit in this setting. 184 , 202
Appropriate antibiotics should be given empirically if infection is suspected. There is no therapeutic advantage to antibiotics in the absence of infection. 203 , 204 , 205
Patients admitted for severe UC should be given IV steroids; both methylprednisolone 60 mg/24 h and hydrocortisone 100 mg every 4 hours are effective. Higher dosages, continuous infusions, and prolonged courses beyond 7 to 10 days do not confer added benefit. 102 , 117 , 206 , 207 Approximately 67% of patients will respond to IV steroids, although 29% of patients will ultimately stall or deteriorate, necessitating a colectomy. 117
Cyclosporine may be an alternative to IV steroids in this setting, although it is less studied. A small, randomized controlled trial comparing cyclosporine 4 mg/kg/d to methylprednisolone 40 mg/d in 30 patients with severe acute UC resulted in equivalent (if not improved) efficacy in obtaining a clinical response in the cyclosporine group (67% for cyclosporine vs. 53% for methylprednisolone). It should be emphasized here that this is not the standard of care, and should only be considered in patients in whom steroids should be avoided (history of steroid psychosis, osteoporosis, poorly controlled diabetes).
Clinical response should be continuously assessed, and decision making should be undertaken in a multidisciplinary setting with a surgeon and a gastroenterologist, with input from the patient. Steroid-refractory patients must be identified early and projected to either rescue therapy or surgery, as delayed colectomy is associated with high morbidity. 208 , 209
Intravenous Steroid Refractory Ulcerative Colitis
If declaration of success or failure has not already been made, the third day of treatment marks a critical junction in decision making. Continued administration of ineffective high-dose IV steroids in this setting risks progression of disease, delayed colectomy, and increased morbidity. 208 , 209 Any discussion regarding salvage attempts using cyclosporine, infliximab, or tacrolimus should be framed such that expectations are appropriately managed. Ideally, the patient has already met with the surgical team and enterostomal therapist, as it is important to have already established a foundation of trust should the patient deteriorate. Certainly if significant clinical improvement is not seen by day 3, a surgical consultation must be sought even as salvage therapy is undertaken. Colectomy is indicated if improvement is not seen after 4 to 7 days of salvage therapy. 102
A number of clinical, biochemical, and radiologic/endoscopic markers have been used to predict the likelihood of steroid failure and progression to urgent colectomy. Fever, tachycardia, and elevation of inflammatory markers such as ESR and CRP are all associated with increased colectomy rates despite aggressive steroids. ESR greater than 75 mm/h and fever greater than 38 °C at the time of admission confers a 4.6- and 8.8-fold increased risk of colectomy on that admission. 210 Days 2 to 5 are critical for repeat assessment of risk for steroid nonresponse. Greater than 12 bowel movements in 24 hours on day 2 of steroids is associated with colectomy in 55% of patients. 211 Greater than eight bowel movements, three to eight bowel movements in addition to a CRP > 45 mg/L, or the product of number of stools multiplied by 0.14 CRP equaling ≥ 8 on day 3 of steroids is suggestive of a 75 to 85% chance of acute progression to colectomy. 212 , 213 If the result is less than a 40% reduction in the bowel movements on day 5, the conclusion is that the patient did not respond to steroids. 210 Colonic dilatation greater than 5.5 cm, mucosal islands, and/or small bowel ileus noted on radiography each predict a 73 to 75% chance of same-admission colectomy. 211 , 214 The presence of severe ulceration noted on endoscopy predicts progression to colectomy in 93% of patients. 215 These dire numbers should be used to set the stage for the patient’s decision regarding salvage therapy versus procession directly to surgery.
Refractory Proctitis and Distal Colitis
A category of disease that warrants separate discussion is refractory proctitis and distal colitis. By definition, these patients have failed topical mesalamine and/or topical steroids as well as oral mesalamine and steroids. These patients suffer from the most distressing symptoms (urgency, tenesmus, incontinence), but may not progress to fulminant disease or even develop the dangerous signs of fevers and abdominal pain and distention. 102 As such, the decision to proceed with surgery is never forced and the patient is left in limbo, suffering from disease, medication side effects, and the tormenting decision regarding whether or not to proceed with “elective” surgery.
In these situations, it is important to exclude medication noncompliance, alternative diagnoses, and proximal constipation/dysmotility (which may decrease the concentration of medication delivered to the site of inflammation). If these alternative explanations are not identified, additional treatments may be considered. Remaining therapeutic options include oral and/or rectal calcineurin inhibitors and biologics. 216 , 217 , 218 , 219 The overall efficacy of these salvage therapies specifically for isolated distal disease has not been established; however, they are not likely any more efficacious than in more extensive colitis. A considerable number of these patients will therefore fail salvage therapy.
Alternative therapies are not validated; however, small trials have suggested some conferred benefit with fatty acid enemas, lidocaine enemas, acetarsol suppositories, epidermal growth factor (EGF) enemas, and transdermal nicotine patches. 220 , 221 , 222 , 223 , 224 , 225 Cohort studies also suggest that up to 90% of patients with ulcerative proctitis may have significant improvement in disease activity after appendectomy, suggesting a possible etiologic link if not a potential therapeutic intervention. 226
Salvage Therapy
Cyclosporine has shown efficacy as salvage monotherapy in steroid-refractory severe acute UC, with 76 to 85% of patients avoiding colectomy on that admission. Median time to improvement in pooled series is approximately 4 days, which allows for timely colectomy in those who do not respond. 102 , 227 , 228 , 229 , 230 , 231 Response rates are no different between 2 and 4 mg/kg/d, 102 , 227 , 228 , 229 , 230 , 231 and considering the narrow therapeutic index, significant side effect profile, and 3 to 4% risk of death with cyclosporine, the 2 mg/kg/d dosing has become the most commonly used regimen. 102 Responders should be transitioned to oral cyclosporine at a dose of 5 mg/kg/d divided twice daily. This regimen is generally continued for 3 to 4 months, and is given in conjunction with a standard maintenance therapy for steroid-refractory patients such as azathioprine (see section on maintenance therapy). 232
While relatively effective for avoidance of colectomy in the short term, the long-term impact of cyclosporine on colon salvage is questionable. Follow-up data suggest that between 58 and 88% of patients who required cyclosporine rescue underwent a colectomy within 7 years. 230 , 233 Patients who were able to maintain some degree of remission on a thiopurine, or who were thiopurine naïve at baseline, were much more likely to avoid colectomy over the long term after cyclosporine salvage. 230 , 234 , 235
Tacrolimus has a similar mechanism of action as cyclosporine (calcineurin inhibitor) but binds a different receptor. Both have multiple downstream immunologic effects, with inhibition of IL-2 likely the most significant. Several trials have shown a benefit in terms of induction of remission, symptomatic improvement, and decreased steroid requirements. Ogata et al performed a randomized controlled trial to identify dose efficacy of oral tacrolimus for salvage therapy. Results of treatment with high trough concentrations (10–15 ng/mL), low trough concentrations (5–10 ng/mL), and placebo were compared at 2 weeks. Sixty-eight percent of patients in the high-trough group experienced symptomatic improvement, with 20% entering clinical remission and 79% showing evidence of mucosal healing. These marked a substantial improvement over placebo (10% improvement) and low trough (18% improvement), although the study was underpowered to show a significant difference between the high- and low-trough strategies. An open label extension of that study showed the majority of patients who were improved at week 2 remained improved over an extended period, with 55% of patients demonstrating an improvement in disease activity index at week 10. Prednisolone doses were weaned from 19.7 mg/d at study entry to 7.8 mg/d at week 10. 236 A more recent prospective observational study showed clinical remission at 4 weeks in 76% of steroid refractory patients treated with tacrolimus. 237 The results of this latter study, while encouraging, are likely confounded by the inclusion of patients with moderate disease. Other case series have shown mixed outcomes, with clinical improvement rates ranging from 47 to 90%. 238 , 239 , 240 Long-term colectomy-free survival has been reported as 57% at 44 months; however, this included patients with both moderate and severe diseases. 238
Infliximab has demonstrated efficacy for rescue therapy in patients with steroid-refractory severe acute UC. Case series report 20 to 75% colectomy rates in steroid refractory severe colitis treated with infliximab. 241 , 242 , 243 , 244 , 245 Järnerot et al 136 performed a randomized controlled trial comparing colectomy rates in patients with steroid-refractory severe and moderately severe UC who received a 5 mg/kg dose of infliximab versus those who received placebo. Colectomy rates at 3 months and 3 years were significantly lower in those who received infliximab (29 vs. 67%; p = 0.017 and 50% vs. 76%; p = 0.02, respectively). 246 The Active Ulcerative Colitis Trial 1 (ACT-1) and ACT-2 randomized controlled trials evaluated infliximab induction and maintenance therapy in ambulatory patients with moderate to severe active UC. Clinical remission occurred significantly more in the infliximab group relative to placebo (≥ 30 vs. 15%). The studies also identified significantly decreased rates of colectomy at 54 weeks in the infliximab group (10 vs. 17%; p = 0.02). 137 It should be noted that the ACT studies looked at ambulatory patients specifically and were not evaluating true salvage therapy per se, which likely explains the lower overall colectomy rate in its population when compared to inpatient studies.
The decision to proceed with a calcineurin inhibitor versus infliximab in acute, steroid refractory colitis is not straightforward. Pooling of available nonrandomized data suggests superiority of infliximab over cyclosporine, with therapeutic response rates of 75 and 55%, respectively. Similarly, nonrandomized 3- and 12-month colectomy rates were lower in patients treated with infliximab (24 vs. 43% and 21 vs. 37%, respectively), although the difference at 3 months was not statistically significant. 247 However, recent randomized controlled trials have refuted the superiority of infliximab, and meta-analysis of the three studies suggests similar clinical response rates of 43.8% for infliximab and 41.7% for cyclosporine. The 3- and 12-month colectomy rates of 26.5 versus 26.4 and 34 versus 43%, respectively, were also similar. 247 , 248 , 249 , 250
In the absence of definitive data showing superiority of one rescue therapy over another, the treatment decided upon should be patient-centric. With the widespread use of infliximab for maintenance therapy, the decision regarding salvage therapy is often made by default, as the inflammation is already refractory to infliximab. Cyclosporine is therefore preferable in these patients if an attempt at rescue therapy is desired. Another theoretical benefit of cyclosporine over infliximab is the shorter half-life (8 hours vs. 8–10 days, respectively), which may be a factor for those who believe that infliximab imparts an increased risk for postoperative infections. Infliximab, however, has the desirable feature of being able to be continued for maintenance therapy in responders. In patients who are naïve to both therapies and do not have a contraindication to either, there is some evidence to suggest the safety of sequential therapy if one fails. 251 Nevertheless, current recommendations discourage this strategy in favor of early colectomy based on prior studies suggesting an adverse risk-to-benefit ratio. 102 , 252 , 253
An area of ongoing research is the ability to predict response to infliximab. Jürgens et al 254 correlated inflammatory markers and IBD-associated genetic variations to infliximab response in patients with moderate to severe UC. High CRP levels, ANCA seronegativity, and homozygosity for IBD-associated variants of the IL23R gene were associated with improved clinical responses. In the future, such predictive models will hopefully be able to guide our decision making as it pertains to escalation of medical care, timing of surgery, and even the recommendation for pouch reconstruction versus stoma.
Management in the Ambulatory Setting
Patients who respond to initial management may stratify into a number of disease courses. These categories include relapsing disease, steroid dependence, refractory, and long-term remission. The ultimate goal of medical management is long-term, steroid-free remission. Achievement of that goal depends on a symbiosis between disease phenotype, patient compliance, and selection of the appropriate therapeutic agent.
Relapsing Disease
Early relapsing disease (< 3 months) should be treated with induction therapy as before, depending on the severity and extent of disease. Additionally, azathioprine or mercaptopurine should be started in order to minimize the risk of ongoing relapses or smoldering disease. Repeat endoscopy is unnecessary unless it will change management (i.e., proximal extension necessitating the addition of oral mesalamine). Relapses greater than 3 months apart should be treated as they were previously, without the addition of immune modulators. 102
Steroid Dependence
Steroid dependence is defined as the inability to taper glucocorticoids to less than 10 mg/d within 3 months of starting steroids without recurrent disease. 255 Steroids should not be indefinitely continued due to the substantial side effect profile. Ardizzone et al 256 compared azathioprine with mesalamine for induction and maintenance of steroid-free remission. Patients on 40 mg/d of prednisolone were randomized to receive azathioprine 2 mg/kg/d or oral mesalamine 3.2 g/d. After 6 months, 53% of patients in the azathioprine group were in steroid-free remission compared to 21% in the mesalamine group (odds ratio [OR], 4.78; 95% confidence interval [CI], 1.6–14.5). Longer-term follow-up in an uncontrolled study looking at steroid-dependent patients suggested a similar benefit of azathioprine, with 12-, 24-, and 36-month steroid-free remission rates of 55, 45, and 42%, respectively. 257 As such, azathioprine should be the mainstay therapy in all steroid-dependent ambulatory patients without a contraindication.
Oral Steroid Refractory
Ambulatory patients with UC refractory to oral steroids already stratify into a challenging disease phenotype. This group is the subject of intense pharmacologic research, and really is the target population for the use of “biologic” therapy. While these medications have generated the majority of publicity in recent years, there may be a role for thiopurines and calcineurin inhibitors as well.
Biologic Therapy
The drive behind ongoing research on biologic therapies in autoimmunity is based on the inadequacy of disease control with the use of conventional therapies. Currently, there are six biologic agents used in UC: infliximab, adalimumab, golimumab, vedolizumab, etrolizumab, and tofacitinib, all of which are superior to placebo for both induction and maintenance of remission. 258 While no direct comparative studies have been performed, a recent meta-analysis including available randomized controlled data indirectly compared infliximab to adalimumab, and found that infliximab is more likely to induce a clinical response (OR, 2.36; 95% CI, 1.22–4.63) and mucosal healing (OR, 2.02; 95% CI, 1.13–3.59).
Tumor Necrosis Factor-α Inhibitors
Infliximab: The approval of infliximab for UC in 2004 opened new avenues and offered medically refractory patients hope for remission. The ACT-1 and ACT-2 randomized controlled trials evaluated the efficacy of infliximab (TNF-α inhibitor) for induction of remission in such patients. A 5-mg dose was as effective as was a 10-mg dose at inducing remission, which was observed in 21% of patients at 30 weeks (compared to 7% of controls/placebo). The studies also identified significantly decreased rates of colectomy at 54 weeks in the infliximab group (10 vs. 17%; p = 0.02). 137 A more recent placebo-controlled trial looking at immune-modulator-naïve patients with steroid-refractory UC suggests that the combination of infliximab and azathioprine has an additive, if not synergistic, effect on remission induction. Sixteen-week steroid-free remission was achieved in 24, 22, and 40% of patients taking azathioprine monotherapy, infliximab monotherapy, and combination therapy, respectively. 259 It is therefore recommended that infliximab be administered in combination with azathioprine.
Adalimumab: Another TNF-α inhibitor that has shown promise in induction and maintenance of remission is adalimumab. The recommended initial dose of adalimumab is 160 mg, followed by a second dose 2 weeks later of 80 mg, and a maintenance dose of 40 mg every other week, thereafter. A randomized controlled trial comparing adalimumab to placebo in steroid refractory, moderately severe UC showed improved 8-week steroid-free remission rates in the adalimumab group (18.5 vs. 9.2%). At 8 weeks, there was no substantial difference in remission rates between the subgroup who was on steroid therapy at the time of adalimumab induction and those who were not (16.9 vs. 9%). 260 In contrast, another randomized controlled trial looking at the efficacy of adalimumab for long-term steroid-free remission suggested that patients who were on steroids when adalimumab was started had significantly higher 54-week steroid-free remission rates than those who were not (13.3 vs. 5.7%). 261 This likely indicates improved efficacy for adalimumab in steroid dependency as opposed to steroid refractory, although it had a benefit in both groups. Interestingly, this study included patients who had previously been on infliximab, suggesting that failure of one TNF-α inhibitor may not translate to a failure of all.
Golimumab: Golimumab is another recently investigated TNF-α inhibitor in refractory moderate to severe UC. While the same volume of data do not exist, it appears from randomized controlled trials that its efficacy for inducing and maintaining remission is similar to infliximab and adalimumab. Sandborn et al studied 6-week clinical response rates for two different induction doses of golimumab (200 mg and then 100 mg, or 400 mg and then 200 mg, 2 weeks apart) and found 51 and 55% responses for 200/100 mg and 400/200 mg doses, respectively. These were in comparison to a 30% improvement rate for patients randomized to placebo. Fifty-four-week follow-up using maintenance doses of golimumab found a persistent clinical response rate of 47% for patients receiving a 50-mg dose every 4 weeks and 50% of patients receiving a 100-mg dose every 4 weeks. Remission and mucosal healing rates were higher among patients receiving the 100-mg maintenance dose (27.8 and 42.4%) than those receiving the 50-mg maintenance dose (23.2 and 41.7%) or placebo (15.6 and 26.6%; p < 0.004). 262 Impact of maintenance therapy was only analyzed in patients who had a response to the induction dose.
Sequential TNF-α Inhibitors
Recent studies have looked at the impact of sequential TNF-α inhibitor use in ambulatory refractory UC. 263 Contrary to intuition, the data seem to suggest that failure of two consecutive TNF-α inhibitors does not predict a nonresponse to a third. De Silva et al 263 found that over 50% of patients trialing a third anti-TNF agent after failure of two previous anti-TNF therapies were able to remain on the third agent for over a year. Three year follow-up data are less encouraging, with only 25% remaining on the same therapy by that time. Prior primary nonresponders to the first anti-TNF agent and persistent disease activity at 3 months after commencement of a third anti-TNF predicted poorer response.