List of Abbreviations
Adipose-derived stem cells
Autologous hematopoietic stem-cell transplantation
Chronic antibiotic-refractory pouchitis
The Food and Drug Administration
Familial Mediterranean fever
Fecal microbiota transplantation
Granulocyte colony-stimulating factor
Genome-wide association studies
Inflammatory bowel disease
Inflammatory bowel disease–unclassified
Inflammatory bowel disease–variant
Ileal pouch-anal anastomosis
Mycobacterium avium paratuberculosis
NOD2-associated autoinflammatory disease
Nucleotide-binding oligomerization domain containing
Orthotopic liver transplantation
Primary sclerosing cholangitis
Tumor necrosis factor
Dr. Bo Shen is supported by the Ed and Joey Story Endowed Chair.
Classic inflammatory bowel disease (IBD) consists of Crohn’s disease (CD) and ulcerative colitis (UC), which run chronic diseases with relapsing and remitting clinical courses. The diagnosis of IBD is made with histologic evidence of inflammatory and structural changes, with or without acute inflammation, combined with supporting clinical, endoscopic, and radiological features. The exact “triggers” for IBD are not entirely clear but the diseases are believed to be caused by a combination of genetic predisposition, abnormal immunity, and environmental exposures. Generally speaking, classic IBD has been considered to be idiopathic.
As it is apparent in clinical practice, the phenotype IBD is far from uniform. Whereas one patient may have severe Crohn’s colitis with debilitating arthropathy, another patient may have mild UC with pyoderma gangrenosum (PG). The diagnosis may be unclear in 10%–15% of patients who will carry a disease entity of IBD-unclassified (IBD-U). To confuse the picture even further, infectious mimics to IBD such as intestinal tuberculosis (ITB) are frequently difficult to differentiate from CD.
Inflammatory bowel disease can be associated with traditionally defined extraintestinal manifestation (EIM) such as primary sclerosing cholangitis (PSC) and PG. However, various autoimmune and autoinflammatory diseases occur concomitantly with IBD but are not considered to be classic IBD-associated EIMs. For example, IBD patients may have concurrent psoriasis, autoimmune hepatitis, or celiac disease. Those immune-mediated diseases have normally viewed as separate, distinct entities from IBD. However, those immune-mediated disorders, including classic, idiopathic IBD, may share part(s) of common pathways.
Aside from the Montreal classification for CD and UC, multiple different classifications have been proposed. Vermeire et al. advocated for a molecular reclassification of IBD, noting that there is frequently a poor correlation between genetic-based subgroups and clinical phenotypes. Shen et al. proposed a systemic overlap syndrome of the gut (IBD) and liver. Cleynen et al. reported genetic evidence supporting a location-focused classification of IBD, for example, ileitis, colitis, ileocolitis, and proctitis. Levine et al. elaborated on the Montreal classification, advocating for pediatric IBD to be subdivided into diagnosis before or after age 10. We propose an adjustment from the traditional concept of IBD as a stand-alone field to IBD as a part of a spectrum of immune-mediated intestinal diseases and overlap syndromes. IBD can be further delineated into primary (“classic”) and secondary IBD. Variables such as the genome, exposome, microbiome, and immunome contribute to the variation in disease presentations, disease course, and outcomes.
Conventional Theories of Pathogenesis of IBD
The conventional theory in the pathogenesis of UC and CD begins with a dysbiotic shift followed by dysregulated innate and adaptive immunity in genetically susceptible hosts. Noticeable shifts in bacterial flora are apparent in IBD patients, with an overrepresentation of select bacterial taxa and decrease in overall diversity of the bacterial community. This dysbiosis leads to an inflammatory response. Genetic variants such as nucleotide-binding oligomerization domain containing 2 (NOD2), lead to alterations in host innate immunity such as defective sensing of bacteria and decreased production of antimicrobial peptides. Defects in innate immunity result in dysregulated T-cells of the adaptive immune system, chronic inflammation, and enterocyte apoptosis. The etiopathogenesis of UC and CD is also described in Chapter 1 . Genetic factors, such as interleukin-10 (IL-10)/IL-10R mutations may play a more important role in pathogenesis of infant or pediatric onset of IBD.
Features of Classic IBD
The definition of IBD has been traditionally limited to CD and UC. In some cases the diagnosis is unclear, despite clinical, endoscopic, radiographic, and gross pathologic evaluations. The label of IBD-indeterminate or indeterminate colitis (IC) is usually given in these cases. The Montreal classification is commonly used to categorize CD in terms of age, behavior (inflammatory, stricturing, penetrating), and location (ileal, colonic, ileocolonic, upper gastrointestinal (GI) tract, and perianal). In UC, patients are categorized based on the extent of disease (proctitis, left-sided colitis, and extensive colitis) and disease severity.
The conventional discriminating features have been used to differentiate UC from CD for decades. CD may involve any part(s) of the GI tract, whereas UC is confined to the colon, rectum, and in some cases, distal ileum (backwash ileitis). Inflammation in CD (with or without granulomas) may extend transmurally, leading to a diverse array of complications including stricture, fistula, and abscess, whereas inflammation in UC is generally limited to the mucosa, muscularis, and up to the superficial submucosa ( Fig. 2.1 ). In addition, a segmental distribution with skip lesions and rectal sparing are characteristics of CD, whereas the rectum is always affected in UC at presentation. Of interest, the disease location and extent of CD remain relatively stable, even after ileocolonic resection, whereas the disease extent of UC often migrates proximately. These phenomena suggest that etiopathogenetic pathways of CD and UC do not completely overlap.
The treatment of IBD involves use of antiinflammatory agents such as mesalamines, corticosteroids, immunomodulators (e.g., thiopurines and methotrexate), anti-tumor necrosis factor α (TNFα; e.g., infliximab, adalimumab, golimumab, and certolizumab pegol), anti-integrins (e.g., vedolizumab and natalizumab), anti-ILs (e.g., ustekinumab), and pathway-targeted small molecules (e.g., ozanimod). The newer agents have been developed to target various etiopathogenetic pathways involved in IBD.
Conventional Classification of IBD
Inflammatory bowel disease has traditionally been classified into UC and CD. The term of IC has been used by the GI pathologist to describe transmural inflammation in severely inflamed colon, which precludes the diagnosis of UC or CD. Patients with clinical, endoscopic, and histologic features, which do not completely fit the diagnosis of UC or CD, may be labeled as IBD-U.
Both UC and CD are then further subclassified based on the age of onset, disease location/extent, and disease phenotype and severity. The most commonly used is the Montreal classification (see Chapter 1 ).
However, the current classification systems are not able to cover the spectrum of immune-mediated IBD and IBD-like conditions, with a wide range of etiopathogenesis, disease course, phenotypes, and histopathologic features.
Genetic Contributions in Development of IBD
The classic boundary between UC and CD is frequently obscured. Up to 9% of patients diagnosed with either UC or CD may have their diagnosis changed within the first 2 years after diagnosis. Approximately 23%–35% of patients with CD have disease limited to the large bowel, that is, Crohn’s colitis. The distinction between Crohn’s colitis and UC can be challenging, especially if the colon has severe inflammation, leading to a histopathologic diagnosis of IC ( Fig. 2.2 ). It has also been reported that some patients with UC may have duodenal involvement, especially in those with concurrent PSC. These ill-defined “gray zones” have posed a great challenge for clinicians in the diagnosis and management of IBD.
Attempts have been made to further define these “gray zones” based on genetics. It is known that of the 163 confirmed IBD susceptibility loci, many are dually associated with both UC and CD, making the majority of IBD polygenic. Moreover, a recent genetic-phenotype profiling study redefined IBD into three groups, ileal CD, colonic CD, and UC, noting that disease location has a strong association with genetics. In contrast, traditional distinctions (UC and CD) or disease behavior (penetrating, stricturing, inflammatory) do not match up well with predictive models of genetic risk.
Single-gene mutations, rather than gene profiling, can also define phenotype in pediatric IBD. Monogenic mutations in IL-10RA and IL-10RB determine phenotype in a very early-onset IBD. In very early-onset IBD, infants develop symptoms of perianal fistulas, diarrhea, oral ulcers, and folliculitis within the first year of life. The defect in IL-10 signaling in this special form of IBD has a Mendelian inheritance pattern with complete penetrance. Owing to the unique genetic defect leading to disease, patients with very early-onset IBD do not usually respond to conventional IBD therapy and require alternative treatments such as allogeneic stem-cell transplants. Thus in many forms of IBD, there are genetically driven (monogenic and polygenic) disease phenotypes.
IBD With EIMs
Inflammatory Bowel Disease frequently presents with EIMs involving the skin, eyes, joints, liver, lungs, or pancreas. The classic EIMs include erythema nodosum, PG, uveitis, episcleritis, iritis, ankylosing spondylitis, sacroiliitis, and PSC. The gut disease activity of IBD may or may not be associated with the presence of and severity of those EIMs. The treatment of the underlying IBD is a key to controlling many of these EIMs. In contrast the severity of some EIMs is not driven by underlying intestinal inflammation. For example, the disease courses PG, ankylosing spondylitis, and PSC are independent of bowel inflammation. This is particularly apparent in the disease course of PSC in postcolectomy UC patients.
It is unclear why certain diseases have been labeled as EIMs of IBD, whereas other commonly IBD-associated disorders are referred to as separate disease entities. For instance, ankylosing spondylitis is considered an EIM of IBD, whereas rheumatoid arthritis is regarded as a concurrent autoimmune disorder (AimD) ( Table 2.1 ). In fact, patients with IBD were shown to have comparably high odds ratios of having ankylosing spondylitis (odds ratio = 5.1) or rheumatoid arthritis (odds ratio = 3.5). As another example, unlike erythema nodosum or PG, psoriasis is not considered a dermatologic EIM of IBD, despite the known association between psoriasis and IBD. In fact, ustekinumab, an agent for psoriasis, was recently approved for the treatment of CD in the United States. Other immune-mediated diseases, such as autoimmune thyroiditis and autoimmune hepatitis, which occur concomitantly with IBD, are considered as concurrent AimDs but as classic EIMs of IBD. Up to now, we have taken this traditional classification system at face value, which has created confusion in clinical practice. This is now leading to our proposal for reclassification of IBD and its associated disorders ( Table 2.2 ).
|Classic Extraintestinal Manifestations of Inflammatory Bowel Disease||Examples of “Concurrent” Autoimmune Disorders of Inflammatory Bowel Disease|
|Skin||Pyoderma gangrenosum, erythema nodosum||Psoriasis, Hashimoto thyroiditis, celiac disease|
|Liver||Primary sclerosing cholangitis||Primary biliary cirrhosis, autoimmune hepatitis|
|Joint||Ankylosing spondylitis||Rheumatoid arthritis|
|Disease location, extent and depth +/− granulomas||Ulcerative colitis||Classic ulcerative colitis|
|Crohn’s disease||Classic Crohn’s disease|
|Age of onset||Very early onset||Age 0||IL-10/ILR mutations|
|Early onset||Age 0–10 years|
|Age 10–17 years|
|Regular onset||Age 17–40 years|
|Late onset||Age >50 years|
|Phenotype||Inflammatory||Inflammatory Crohn’s disease; classic ulcerative colitis|
|Stricturing||Stricturing Crohn’s disease; ulcerative colitis with stricture|
|Penetrating||Fistulizing Crohn’s disease|
|Extraintestinal||Metastatic Crohn’s disease of the skin, lung, liver|
|Concurrent or immune-mediated disorders||IBD||Isolated ulcerative colitis or Crohn’s disease of the gut|
|IBD-variant||IBD +||IBD + classic extraintestinal manifestations||Ulcerative colitis with concurrent primary sclerosing cholangitis|
|IBD ++||IBD + autoimmune and/or autoinflammatory disorders ± classic extraintestinal manifestations||IBD with concurrent microscopic colitis, celiac disease, hidradenitis suppurativa|
|IBD +/−||Diseases sharing clinical features and possible etiopathogenetic pathways with classic IBD ± classic extraintestinal manifestations of IBD, autoimmune disorders or autoinflammatory disorders||Lymphocytic colitis, collagenous colitis; Behcet’s disease, cryptogenic multifocal ulcerous stenosing enteritis, ulcerative jejunitis|
|Etiology of IBD||Primary or idiopathic||Monogenic||IL-10, IL-10RA, IL-10RB mutations |
Very early–onset IBD
|Polygenic||Classic ulcerative colitis; classic Crohn’s disease|
|Secondary||Identifiable pathogens||Mycobacterium avium paratuberculosis|
|Medication-induced||Mycophenolate-associated colitis; Ipilimumab-associated colitis|
|Organ transplantation-induced||Post-solid organ transplant IBD-like conditions, cord colitis syndrome;|
|Surgery-induced||Pouchitis, Crohn’s disease-like conditions of the pouch, postcolectomy enteritis, bariatric surgery-associated IBD|
|Genetic etiology||Monogenic||IL-10/IL-R mutations, familial Mediterranean fever|
|Polygenic||Classic Crohn’s disease and classic ulcerative colitis|
|Disease spread process||Intrinsic (“inside-out”)||Starting from the lymphatic system or mesentery, spreading to gut mucosa||Subset of obese Crohn’s disease patients; subset of sclerosing mesenteritis or lymphangitis|
|Extrinsic (“outside-in”)||External trigger (e.g. bacteria) leading to mucosal inflammation||Fulminant ulcerative colitis: from mucosal disease to transmural inflammation|
Overlap of IBD and AimDs
Inflammatory bowel disease and other AimDs are known to occur together in a subset of patients. The list of AimDs overlapping with IBD is extensive, including celiac disease, microscopic colitis, Hashimoto thyroiditis, psoriasis, rheumatoid arthritis, and autoimmune hepatitis. For example, two common immune-mediated bowel diseases, celiac disease and IBD, can occur concurrently. CD and UC have a reported prevalence of 4% and 3.2%, respectively, in patients with celiac disease, which is higher than the prevalence of 1%–2% in the general population. Interestingly, patients with coexisting celiac disease and UC were more likely to have extensive colitis. This more extensive clinical phenotype suggests that overlap of celiac disease and UC may represent a distinct phenotype of IBD. In addition, lymphocytic colitis and collagenous colitis can also be present in IBD patients. In the majority of reported cases, older patients with microscopic colitis, usually collagenous colitis, later on may develop UC. One of explanations for the overlap between IBD and those AimDs is from genetic association. AimDs, such as psoriasis, systemic lupus erythematosus, type 1 diabetes mellitus, multiple sclerosis, and vitiligo, have been clinically as well as genetically linked with IBD in genome-wide association studies (GWAS). GWAS highlight the genetic contribution to IBD overlap with autoinflammatory disorders (AinDs). The genetic studies are the first step for the molecular classification of the spectrum of immune-mediated disorders.
Overlap of IBD and AinDs
Autoinflammatory disorders are diseases with episodes of unprovoked inflammation without known high-titer antibodies or antigen-specific T-cells. The overlap between CD and other AinD is, in part, being elucidated with genetic sequencing studies. The association between CD and NOD2 gene was discovered in 2001, paving the way for the discovery of other AimD associated with NOD2 variants such as NOD2-associated autoinflammatory disease (NAID), Blau syndrome, and familial Mediterranean fever (FMF). For example, a patient with CD has been reported to have FMF and chronic idiopathic urticarial with angioedema without identified NOD2 mutations.
NOD2-associated autoinflammatory disease is a recently described disease with a constellation of inflammatory symptoms including periodic fever, dermatitis, arthritis, abdominal pain, non-bloody diarrhea, lower extremity swelling, and sicca-like symptoms. Yao et al. described a cohort of 54 adults with identified NOD2 variants and low titers of antinuclear antibody (ANA) without coexisting AimD, AinD, and classic IBD. Although NAID and IBD may represent separate disease entities, there are shared clinical manifestations such as arthritis and colitis. The colitis in NAID usually is not severe and does not have histologic changes of classic IBD. As with UC and CD, genetic association may be partially responsible for the overlap seen between IBD and AinD.
Spectrum of Immune-Mediated Disorders of the Gut
Diverging from the Montreal classification of CD and UC, the different clinical phenotypes of IBD could be organized into alternative classifications based on available, emerging data ( Table 2.2 ). The spectrum of immune-mediated disorders can be categorized according to the clinicopathological features.
Pattern of Disease Extent: Small and Large Bowels
Disorders of the gut could be described as a range of diseases involving the colon to the small bowel ( Fig. 2.3 ). Typical disease phenotypes vary from UC proctitis to ileal CD. Cleynen et al. recently published data on genetic risk scores for corresponding phenotypes in 35,000 patients with IBD. Comparing the Montreal classifications, patient samples were genotyped using the Immunochip array (Illumina, San Diego, CA) to evaluate for 195,806 polymorphisms associated with IBD. The authors noted incongruence between genetic-based subgroups and clinical subphenotypes. Rather than finding correlations with the disease classifications, CD versus UC, the location of disease was found to have the strongest genetic association. The genetic substructure of IBD was divided into UC, colonic IBD-U, colonic CD, ileocolonic CD, and ileal CD. Stricturing CD, a behavioral classification, did not have a strong genetic association. Similar findings were described when examining gene expression and regulation whereby CD could be clearly divided into colon and ileal subclasses. These findings point toward location as a fundamental biological aspect of disease over behavior, which is likely a marker of disease progression.
There are variants for the classified pattern of disease extent. UC with duodenitis is an example.
Pattern of Disease Spread: Extrinsic Versus Intrinsic
Immune-mediated gut disorders have a wide range of histopathologic features. Classic mucosal disease, such as lymphocytic colitis, collagenous colitis, and celiac disease, has disease involvement limited to the epithelium, glands, and lamina propria. UC can involve the mucosa, muscularis, and superficial submucosa. However, severe or fulminant UC can present with deep fissure-like ulcers and transmural inflammation with lymphoid aggregate ( Fig. 2.2 ). These diseases are classic examples for the extrinsic or “outside-in” theory of immune-mediated gut disorders, that is, from the mucosa to deeper layers of the bowel wall ( Fig. 2.4 ). The theory of dysbiosis as a trigger for IBD fits the clinical and histopathologic picture for these disease processes. Logically speaking, the change in gut microbiome may be the primary event in spread of disease from the outside to inside. Supporting this theory, patients with CD or UC may have a preceded viral or bacterial gastroenteritis and some patients with CD may respond to luminally active antimicrobial agents.
In contrast, the “inside-out” theory or the intrinsic pathway of IBD implies that the disease process starts from within the mesentery, serosa, or deep bowel wall, subsequently outward to the mucosa. In these diseases, inflammation may not always follow the route from the lumen to serosa; mucosal ulceration may be a terminal event rather than the first event . CD and Behcet’s disease are classic examples of transmural, segmental, inflammatory disorders of the gut. CD-associated gene mutations in NOD2, ATG16L1, and IRGM lead to defective immune response to intracellular bacteria, the bacteria not invading from the gut lumen. In our clinical practice, we have noticed that mucosa-active agents, such as mesalamines, are not effective to treating patients with transmural CD. Furthermore, the mesentery is hypothesized to harbor the source for chronic intestinal inflammation, which may play a key role in the development of intrinsic IBD.
Spanning from the duodenum to rectum, the mesentery comprises a continuous network of lymphatics, nerves, vasculature, connective tissue, and adipose tissue. Owing to its anatomic complexity and functional role in regulating local and systemic homeostasis, Coffey et al. postulated that the mesentery is a distinct organ, playing an important role in pathogenesis in various diseases. Mesenteric lymph and adipose tissue have been implicated in the early pathogenesis of CD, with bacterial invasion into lymph nodes leading to lymphangitis, lymphangiectasia, and mesenteric adenitis development before any apparent mucosal pathology. The deep, linear ulcers’ characteristics of CD are seen on the mesentery side of the bowel, supporting a model of mesenteric adenitis and lymphangitis in mucosal inflammation. Indeed, the mesentery, including lymphatic, vascular, and neural systems, may help explain the patchy, “skip-lesion” distribution of inflammation in CD .
In addition the available body of evidence suggests that the adipose tissue in the mesentery plays an important role in CD and maybe IBD in general. Mesenteric adipose tissue produces proinflammatory cytokines, including TNF. Increased mesenteric fat hypertrophy is seen in obesity and CD. The characteristic intestinal adipose tissue proliferation known as “fat wrapping” of inflamed bowel is a hallmark of CD. In postoperative CD patients, increased visceral fat area is associated with endoscopic recurrence of CD . If this is true, the prevention of postoperative CD recurrence may depend upon surgical handling of the mesentery ( Fig. 2.4 ).
In clinical practice, we have noticed a growing number of obese patients with CD, in stark contrast to the more typical, underweight, malnourished CD patients. Obese CD patients with increased mesenteric fat, particularly African Americans or Hispanic Americans, often present with severe fistulizing perianal disease. Obese CD patients have been shown to have a higher risk for postoperative complications and postoperative disease recurrence. Malnourished CD patients often have severe mucosal inflammation, along with strictures or fistulae, whereas obese CD patients may have severe enteroenteric or perianal fistulae with minimal mucosal disease ( Fig. 2.4 ). Given that increased visceral adipose mass is also associated with insulin resistance and the metabolic syndrome, it is possible that as the obesity epidemic continues, there may be a unique phenotype of obese CD patients with more severe inflammation that is driven by the mesentery outward to the mucosa rather than by mucosal inflammation extending inward to deeper layers of bowel wall. CD may even be considered as part of metabolic syndrome.
Thus, two models of CD spread exist: extrinsic or “outside-in” and intrinsic or “inside-out.” The currently accepted “outside-in” model does not explain the clinical presentation of all CD patients. The “inside-out” theory suggests that signals from mesenteric adipose tissues, lymphatic system, and possible vascular and neural system contribute to the initiation of (the “inside-out” theory) or exacerbation of CD .
Overlap Syndrome of Immune-Mediated Disorders
Overlap syndromes can be defined as multiple disease states occurring together to produce a unique clinical phenotype and disease behavior, which extends beyond classic IBD and classic EIM. Analogous to the local overlap syndrome of GI system is that in the hepatopancreaticobiliary system. PSC may exist as an isolated disease but may occur concurrently with primary biliary cholangitis and/or autoimmune hepatitis. In fact, PSC per se represents a spectrum of diseases ranging from classic PSC to immunoglobulin G4 (IgG4)-related sclerosing cholangitis to autoimmune sclerosing cholangitis with varying clinical presentations, natural history, and disease course.
At times, the diagnosis is unclear with overlap of clinical, histologic, and immunologic features of multiple diseases. The establishment of the correct diagnoses is important because treatment may be targeted to multiple diseases or a common “denominator.” The mechanisms for the development of overlap syndromes are not well defined, hence optimal treatments for overlap syndromes are also not well defined. The understanding of systemic overlap syndromes involving IBD may lead to more efficacious treatment options for these unique phenotypes.
Overlap Syndrome of Gut
Similar to the liver diseases, a subset of IBD patients will develop an overlap syndrome of immune-mediated disorders in the gut. A classic example is concurrent UC or CD and microscopic colitis, celiac disease, or autoimmune enteritis ( Fig. 2.5 ).