Concepts in Inflammatory Bowel Disease Management

Genetics

The familial aggregation of IBD was an important early clue to the genetic basis of IBD. Indeed, the greatest risk factor for IBD is a positive family history, with first-degree relatives of IBD patients having an approximately 15-fold higher risk for IBD than the general population. Additional evidence suggesting a genetic susceptibility came from observed ethnic differences in IBD incidence and prevalence, the association of IBD with other disorders with recognized genetic susceptibility (such as ankylosing spondylitis and psoriasis), and the existence of genetic syndromes (such as glycogen storage disease type1b and Hermansky-Pudlak syndrome) with phenotypes resembling IBD. Although familial and ethnic aggregation may reflect shared genetic factors, environmental factors may be shared as well. Powerful support for the contribution of both genes and the environment in IBD came from twin concordance studies. Concordance for CD among monozygotic twins ranged from 35% to 58%, whereas dizygotic twin concordance ranged from 0% to 4%. In UC, monozygotic and dizygotic concordances were 16% to 19% and 0% to 5%, respectively. The twin concordance studies confirmed (1) the contribution of both genes and the environment in IBD and (2) the stronger genetic influence in CD compared with UC.

A number of approaches were taken to identify IBD susceptibility genes. Linkage studies identify regions of the human genome associated with disease susceptibility by testing a series of marker alleles for cosegregation (linkage) with disease status across a number of families. In 1996 the first CD linkage study mapped a susceptibility gene to chromosome 16. In 2001 two studies identified three low-frequency risk variants within the NOD2 (nucleotide-binding oligomerization domain containing 2) gene on chromosome 16. NOD2 is an intracellular receptor that binds muramyl dipeptide (MDP), a bacterial wall peptide. The three NOD2 variants individually conferred odds ratios (ORs) of 2 to 4 in heterozygotes and 20 to 40 for homozygotes. At least one mutation was present in 30% to 40% of CD cases compared with 6% to 7% in European controls. Notably, these NOD2 variants were not observed in CD patients of Japanese or Chinese ancestry. Owing to the fact that linkage studies can easily detect only highly penetrant risk loci, subsequent studies replicated only a few additional loci (6p21 [human leukocyte antigen (HLA)], 5q31, and 19p).

Genome-wide association studies (GWASs) constituted the next milestone in IBD gene discovery. These studies became feasible after technologic advances allowed the mapping of the human genome using a relatively limited number of selected single nucleotide polymorphisms (SNPs) (<1 million SNPs for individuals of European ancestry), which comprehensively assayed common genetic variations across the genome. Compared with linkage studies, GWASs have greater statistical power to detect loci of small to moderate effect sizes. The first GWAS, conducted in a Japanese population and published in 2005, found that polymorphisms in TNFSF15 (tumor necrosis factor superfamily, member 15) conferred susceptibility to CD. TNFSF15 is involved in Th1-mediated inflammatory responses in the gut. Next, a GWAS published in 2006 identified IL23R as a susceptibility gene for both UC and CD (17068223), thereby implicating the interleukin-23 (IL-23)/IL-17–producing T-helper cell (Th17) axis in intestinal inflammation. Subsequent GWASs identified genes involved in autophagy ( ATG16L1 and IRGM ), innate immune responses (TLR4, CARD9, IL23R, STAT3), and the adaptive immune system (HLA, IRF5, PTPN22).

Further insights were provided by meta-analyses of GWAS. The largest meta-analysis, performed in cohorts of European descent (75,000 cases and controls), identified a total of 163 IBD loci, including 30 new loci. There were 110 loci conferring risk to both IBD subtypes, 30 loci unique to CD, and 23 loci unique to UC. This degree of sharing of genetic risk suggests that nearly all of the biologic mechanisms involved in one disease have some role in the other. With the exception of NOD2 (OR 1.5 for CD) and IL23R (OR 1.5 for IBD), all other IBD loci conferred low risks (1 < OR < 1.5). Among the UC-specific loci, HLA conferred the highest risk (OR, 1.15). The loci identified by the meta-analysis explained only 13.6% of CD and 7.5% of UC total disease variances, suggesting that other factors, such as rarer genetic variations not captured by GWAS or environmental exposures, make substantial contributions to pathogenesis. Many of the identified IBD loci were implicated in other immune-mediated disorders, most notably ankylosing spondylitis and psoriasis. The IBD loci identified were also markedly enriched in genes involved in the primary immunodeficiencies, such as genes that correlate with reduced levels of T-cell subsets, such as Th17 cells (STAT3), memory T cells (SP110), and regulatory T cells (STAT5B). Finally, the IBD loci were enriched in genes leading to mendelian susceptibility to mycobacterial disease ( IL12B, IFNGR2, STAT1, IRF8, TYK2, and STAT3 ). Additional analyses were conducted to classify the loci according to immune pathways. These analyses identified genes involved in cytokine production ( interferon-gamma [IFN-γ], IL12, tumor necrosis factor-alpha [TNF-α], IL10 ); activation of T, B, and NK cells; response to molecules of bacterial origin; and the IL-17/IL-23 signaling pathway ( IL23R, IL12B, JAK2, TYK2, and STAT3 ). There was no evidence of CD or UC specificity in any of these pathways.

The first multiethnic association study of IBD (87,000 Europeans and 10,000 individuals of East-Asian, Indian, or Iranian descent) added 38 new IBD loci (4 for UC, 7 for CD, and 27 common loci) to the 163 loci already identified by the GWAS meta-analysis, thus bringing the total of confirmed IBD loci to more than 200. For most loci, the direction and magnitude of effect were consistent in European and non-European cohorts. However, there was genetic heterogeneity at several established risk loci, driven by a combination of differences in allele frequencies (NOD2) , effect sizes (TNFSF15, ATG16L1), or a combination of both (IL23R, IRGM) . It is clear that the genetic architecture of IBD varies across diverse populations.

Altogether, the identified IBD gene and genetic loci can be grouped into several pathways crucial for intestinal homeostasis, including barrier function, epithelial restitution, autophagy and other types of antimicrobial defense, regulation of innate and adaptive immunity, cytokine signaling, generation of reactive oxygen species, endoplasmic reticulum stress, and metabolic pathways associated with cellular homeostasis. For all populations, HLA is the major risk factor for UC. Among white people, NOD2 mutations confer by far the highest risk for CD.

It appears that, in European populations, linkage studies and GWAS have uncovered all variants with large effects (OR > 3) and frequencies greater than 1%. In addition, it is probable that nearly all variants with frequencies greater than 5% and ORs greater than 1.2 have also been identified. The remaining genetic contribution is expected to arise from a combination of common variants with ever-smaller effect sizes and rare variants. Whole exome sequencing (WES) has recently become a practical and cost-effective tool to identify rare variants not detected by GWAS. WES was first used in a boy who presented at 15 months with intractable CD, detecting a rare mutation affecting the regulatory function of the X-linked inhibitor of apoptosis (XIAP) gene. XIAP had not previously been associated with CD but has a central role in the proinflammatory response and bacterial sensing through the NOD signaling pathway. WES has since been used to identify other rare variants implicated in the pathogenesis of very early onset IBD (VEOIBD), a nonclassic form of IBD that occurs in children less than 6 years of age. VEOIBD tends to be resistant to standard medical and surgical therapies and is commonly caused by single gene mutations ( IL-10, IL-10R, NCF2, LRBA, and TTC7 , among many other genes).

The identification of IBD loci using linkage, genome-wide association, and sequencing studies has constituted only a first step in our understanding of IBD. A summary of the key points on the genetics of IBD is provided in Table 161.1 . Fine-mapping studies and experimental work will be required to identify the causal gene variants and elucidate the disease mechanisms.

Epidemiology

IBD is a disease of the modern world. Isolated reports in the late 19th century and the first half of the 20th century gave way to more frequent diagnosis in the second half of the 20th century. A recent, systematic review of more than 200 population-based studies has crystallized the global IBD geography. Both the prevalence (up to 0.5%) and incidence of IBD (10 to 30/100,000 per year) are highest in the Western world, namely Europe, North America, and Oceania. Within Europe, the incidence of IBD is highest in Western Europe, low in countries adjacent to the Mediterranean, and varies from low to high in Eastern Europe. Population data outside the Western world are limited to studies from Japan, China, South Korea, South America, and South Africa. In all of these countries, the incidence of IBD is significantly lower than in Western countries. Interestingly, studies of people emigrating from countries of low IBD prevalence to developed countries found that the incidence of IBD increased in successive generations, highlighting the importance of environmental factors in the development of disease.

The increase in IBD incidence in Western countries in the second half of the 20th century has long been evident and was confirmed by time-trend analysis in a systematic review. Several additional temporal trends are becoming apparent. The previously observed North-South gradient in Europe appears to be fading. IBD incidence is increasing in urbanized societies in East Asia (South Korea, Japan, China, and Hong Kong) and in developing countries. Finally, evidence points to a disproportionate increase in childhood IBD (as compared with adult IBD) over the past few decades.

The sequential rise of IBD in the West and then in the westernizing countries of the developing world may reflect the adoption of the high–animal fat/low-fiber Western diet. Observational studies have linked IBD with consumption of greater amounts of meat and fats, particularly polyunsaturated fatty acids (PUFAs) and omega-6 (n-6) fatty acids. Conversely, diets high in fiber, fruits, and vegetables seem to be protective. The results of the European Prospective Investigation into Cancer and Nutrition (EPIC) study and the Nurses' Health Study are particularly noteworthy because of their prospective design. EPIC found an association between a higher consumption of linoleic acid (an n-6 PUFA present in high concentrations in red meat, cooking oils, and margarine) with a higher incidence of UC. In contrast, people who consumed higher levels of omega-3 (n-3) PUFA docosahexaenoic acid were less likely to be diagnosed with UC. In the Nurses' Health Study, a greater consumption of long-chain n-3 PUFAs and a higher ratio of n-3:n-6 PUFAs was again protective against the development of UC. A higher consumption of fiber, particularly fruits, was associated with a lower risk of CD, but not UC.

In addition to the Western diet, a number of other environmental factors have been implicated in IBD pathogenesis. Cigarette smoking has been the most extensively studied environmental factor. Multiple studies have demonstrated that smoking is associated with an increased risk of CD and more aggressive CD, namely stricturing and fistulizing complications, requirement for immunosuppressive therapy and surgery, earlier postoperative recurrence, and need for reoperation. Smoking cessation is associated with decreased CD activity and a lower risk of postoperative recurrence. Remarkably, smoking has opposite effects in UC. Smokers have a lower risk of a UC reactivation, as well as a lower risk of colectomy. The mechanisms behind the divergent effects of smoking on CD and UC are not well understood. Appendectomy before age 20 for appendicitis (but not for nonspecific abdominal pain) is associated with a lower incidence of UC. Vitamin D deficiency and use of nonsteroidal antiinflammatory drugs have been associated with an increased risk of both UC and CD.

In the United States, IBD represents a significant burden for the health care system. A study published in 2007 used insurance claim data to estimate that 993,300 Americans were affected with IBD in 2005. The prevalence of both UC and CD was lower in the South compared with the Northeast, Midwest, and West. Interestingly, IBD appeared to be more common in commercially insured individuals, compared with those insured by Medicaid. Using insurance data from 2003 and 2004, one study estimated mean annual costs per patient at $8265 for CD and at $5066 for UC. In both diseases, hospitalizations, outpatient care, and medications each accounted for approximately one-third of direct costs. Annual direct costs in the United States were estimated at $6.3 billion ($3.6 billion for CD, $2.7 billion for UC). A follow-up study using the same claims database estimated that 1,171,000 Americans suffered from IBD in 2008 to 2009. Time-trend analysis showed an increase in IBD prevalence over three successive periods (2004 to 2005, 2006 to 2007, and 2008 to 2009), likely attributable to the absence of a cure and the low disease mortality. A population study from Olmsted County, Minnesota, found that the incidence of CD and UC remained stable between 1970 and 2000. In contrast, a population study from Kaiser Permanente Northern California found that, between 1996 and 2006, the incidence of UC increased from 1.8 to 4.9 per 100,000 per year ( P < .001). The annual incidence of CD increased from 2.2 to 4.3 per 100,000 per year, albeit without reaching statistical significance ( P = .09). A summary of key points on IBD epidemiology is provided in Table 161.2 .

Etiopathogenesis

A large number of animal IBD models were developed in the 1990s and early 2000s. Models of aberrant immunoregulation included genetically engineered mice that underexpressed ( IL2 ^−/− , IL10 ^−/− , TCR ^−/− ) or overexpressed ( TNF1 ^ΔARE , STAT4 ) immune regulatory molecules, as well as models of regulatory T-cell transfer (CD45RB ^high → SCID). In other models, IBD was induced by epithelial barrier defects (dominant-negative N-cadherin, mdr1a ^−/− , Muc2 ^−/− ) and toxins (TNBS-induced colitis). Most animal models were characterized by heightened effector cell responses or inadequate regulatory cell responses, ultimately producing either a Th1 inflammatory profile (excessive IL-12/IFN-γ/ TNF-α secretion), or a Th2 profile (increased IL-4/IL-5/IL-13 secretion). CD was viewed as a Th1 disorder, whereas UC was considered an atypical Th2 disorder (characterized by enhanced production of IL-5 and IL-13 but low levels of IL-4). In the early 2000s, the Th1/Th2 paradigm was revised with the discovery of the proinflammatory Th17 cells, which are stimulated by IL-23. IL-17 production was found to be increased in some animal models, as well as in IBD, especially CD. Intriguingly, a GWAS published in 2006 identified IL23R as an IBD gene.

Genetically engineered animal models had inherent limitations. In contrast to the genetic heterogeneity seen in human IBD, the animal models were characterized by single genetic defects. In addition, most models failed to recapitulate small bowel disease and its complications, such as strictures and fistulas. The greatest contribution of the early disease models came from the observation that susceptible animals did not develop IBD when raised under germ-free conditions, thus demonstrating the critical role of commensal intestinal flora in IBD pathogenesis.

Gene discovery refocused basic research in IBD. NOD2 , the first gene definitively linked to CD, is an intracellular receptor that is expressed in several cell types, particularly intestinal monocytes and Paneth cells. NOD2 binds MDP, a bacterial wall peptide, producing a proinflammatory immune response. The discovery of NOD2 thus reframed IBD as a disorder partly caused by an aberrant innate immune response to common bacterial motifs. More than 15 years after the identification of NOD2 as a risk gene, there is still uncertainty as to how loss-of-function mutations of this “proinflammatory” protein could lead to inflammation. Decreases in the antibacterial activities of Paneth cells may play a role. In addition, NOD2 plays a key role in autophagy, a critical innate immune response against intracellular microbes. For example, NOD2 -deficient (NOD2 ^−/− ) mice, but not wild-type mice, were colonized by Bacteroides vulgatus , suggesting that NOD2 may prevent colonization by specific commensal bacteria. Finally, tantalizing clues came from a study showing that NOD2 may be tolerogenic (i.e., antiinflammatory) under certain conditions, such as chronic stimulation by MDP. Hence loss-of-function mutations may abrogate the tolerogenic function of NOD2.

Besides NOD2 , other autophagy genes confirmed to increase CD risk include ATG16L1 and IRGM . In vitro, animal and human studies are examining the role of impaired autophagy in IBD pathogenesis. NOD2 and ATG16L1 risk alleles were associated with impaired autophagy of Salmonella in vitro. Monocytes from CD patients homozygous for the ATG16L1 risk allele were less effective at killing adherent-invasive Escherichia coli when compared with monocytes from patients who were homozygous for the ATG16L1 protective allele. ATG16L1-deficient mice displayed impaired autophagy in the ileum; decreased, aberrant, and disorganized granules within Paneth cells; defective granule exocytosis by Paneth cells; and altered expression of genes involved in regulating response to cell injury. Strikingly, CD patients homozygous for the ATG16L1 risk allele displayed Paneth cell abnormalities similar to those seen in ATG16L1-deficient mice.

The intestinal microbiome represents the newest frontier in IBD research. The human intestine houses several trillions of microbial cells that express almost 10 million genes. More than a billion years of mammalian-microbial coevolution has led to interdependency. Not surprisingly, the intestinal microbiota contribute to a wide array of homeostatic functions, including the maturation and continued education of the host immune response, protection against pathogens, and elimination of toxins. Imbalances in the composition of the intestinal microbiome (dysbiosis) have long been noted in IBD. Dysbiosis encompasses alterations in the relative abundances of certain bacterial taxa, as well as a decrease in the diversity of the community. Emerging research is uncovering the interactions between impaired autophagy and intestinal dysbiosis. For example, inflamed ileal tissue from patients homozygous for the ATG16L1 risk allele contained increased numbers of Fusobacteriaceae, whereas inflamed ileal tissue of patients homozygous for the ATG16L1 protective allele showed decreased numbers of Bacteroidaceae and Enterobacteriaceae and increased Lachnospiraceae. The ATG16L1 allele did not affect the bacterial composition in noninflamed ileal tissue. In addition to alterations in microbial composition, dysbiosis also encompasses alterations in the normal functions of the microbiome. Intriguingly, changes in the intestinal virome were described in patients with IBD. It must be emphasized that dysbiosis is not a proven cause of intestinal inflammation and that altered bacterial composition and function may be sequelae of inflammation. Moreover, there may be bidirectional interactions between intestinal bacteria and inflammation.

The recognition of intestinal dysbiosis in IBD refocused attention to diet as a cause of IBD. Dietary preferences are an established determinant of the intestinal microbiome. A seminal study found striking differences between the microbiota of European children and those of children from a rural village in Burkina Faso. Similarly, a later study found pronounced differences in gut bacterial composition and bacterial functional gene repertoires of individuals living in the United States compared with individuals living in Malawi and the Amazonas State of Venezuela. The dietary intake of animal protein, animal fat, and fiber modulates the gut microbiome. Separately, consumption of fat and fiber may influence the development of IBD. Although it is tempting to speculate that the Western diet may predispose to IBD via effects on the gut microbiome, other mechanisms are probably operable, including effects of the diet on the epithelial barrier and production of inflammatory mediators. In turn, the gut microbiome is influenced not only by diet but also other factors, including mode of birth, age, infections, antibiotic use, gut inflammation, genetics and, possibly, cigarette smoking.

In summary, over the past decade, we have seen tremendous progress in our understanding of IBD pathogenesis, particularly in the areas of abnormal immune regulation (IL-17/IL-23 signaling pathway), autophagy, and intestinal dysbiosis. In parallel, more complex animal models of IBD are being developed, namely models that incorporate abnormalities in autophagy, interactions between host and specific commensal bacteria and viruses, and/or dietary interventions.

Fig. 161.1 summarizes our current understanding of IBD pathogenesis. Genetic susceptibility and proinflammatory environmental factors are necessary, but not sufficient, for the development of IBD. Intestinal dysbiosis (i.e., alterations in the composition and function of intestinal bacteria) is suspected to be a major proinflammatory factor. In turn, the Western diet may select for commensal bacteria that promote the development of intestinal inflammation.

Phenotypic Variability of Inflammatory Bowel Disease

In the first detailed characterization of CD, Crohn, Ginsburg and Oppenheimer described 14 young adults (mean age, 32 years) with disease limited to the terminal ileum. It was soon recognized that, although the disease process typically affected the terminal ileum, the more proximal small bowel and/or the colon could be involved as well. Over time, investigators recognized phenotypic variability in the age of onset and the risk of stricturing, perforating, and perianal complications. In parallel, researchers described the variable phenotype of UC. Some patients presented with nonprogressive proctitis or proctosigmoiditis, whereas others presented with extensive colitis or pancolitis, or experienced proximal extension of initially distal disease. Younger age at diagnosis of UC was associated with a greater risk of colectomy. In 2005 a consensus group codified the phenotypic variability of IBD in the Montreal classification ( Table 161.3 ). This scheme categorized IBD into three types: CD, UC, and IBD, type unclassified (IBDU). IBDU was defined as isolated colitis (i.e., no small bowel involvement) but without definitive histologic or other evidence to favor either CD or UC. CD was characterized according to age of onset, location, and luminal (stricturing or penetrating) and perianal complications. UC was characterized by disease extent and severity of individual relapses. The consensus group felt there was insufficient information to classify UC by age of diagnosis. The most important criticism of the Montreal classification concerns the absence of classification according to long-term risk. Nonetheless, the development of standardized definitions was a significant step toward bringing consistency across clinical studies.

The phenotypic heterogeneity of IBD is not surprising given the multiple genetic, immunologic, microbiomic, and environmental factors involved in IBD pathogenesis. A number of consistent correlations have been observed between certain genetic and environmental factors and disease phenotype. In early studies, NOD2 mutations correlated with small bowel CD and stricturing complications, whereas the HLA-DRB1*0103 allele was associated with severe and extensive UC. A recent study of almost 30,000 IBD patients assessed the association between genotype and phenotype, including age at diagnosis and time to surgery (all IBD), disease location and behavior in CD, and disease extent in UC. Three loci ( NOD2 , MHC, and MST1 3p21 ) were associated with IBD phenotype. NOD2 was associated with CD location, behavior, and age at diagnosis. After adjusting for the other phenotypes, the association of NOD2 with behavior was shown to derive almost entirely from its correlation with location and age at diagnosis. HLA was associated with age of onset, CD disease location and behavior, UC extent, and surgery. MST1 was associated with age of IBD onset. The investigators found that little to no genetic association with disease behavior remained after conditioning on disease location and age at onset. It may be hypothesized that for a disorder like IBD with numerous low-risk alleles, the total burden of alleles may influence disease susceptibility and phenotype. Genetic risk scores that combined information from the 163 known IBD gene associations were generated and were strongly associated with disease subphenotype, even after exclusion of NOD2 , HLA , and MST1 . The genetic scores separated IBD into three distinctive subgroups, namely ileal CD, colonic CD, and UC (rather than the currently defined IBD types of CD and UC). The investigators concluded that disease location is in part genetically determined and that disease location (rather than genotype) is the major determinant of disease behavior. A major caveat is that current studies have examined only phenotype associations with known IBD genes. It is quite plausible that genes that do not predispose to IBD per se may actually influence IBD behavior.

The only consistent environment-phenotype correlation involves cigarette smoking. As reviewed previously, cigarette smoking is associated with an increased risk of complicated CD and a decreased risk of severe UC requiring colectomy. Research correlating the intestinal microbiome with IBD type and behavior is at its infancy. In a study of children with CD, Ruminococcus was implicated in stricturing complications and Veillonella in penetrating complications. These results will need to be replicated in other cohorts. Moreover, the microbiome may interact with genotype to produce distinct disease phenotypes.

The Montreal classification, having been developed in 2005, could have been only clinical. Since then, we have achieved a more sophisticated understanding of IBD pathogenesis. Computational methods have been developed that combine genetic, microbiomic, and inflammatory pathway markers and their interactions. These methods will hopefully generate classification schemes that will integrate the system biology of IBD, provide risk stratification, and predict response to different therapeutic approaches.

Variation in Care and Quality Improvement in Inflammatory Bowel Disease

Variation in care may result in differences in health care quality, outcomes, and costs and could therefore be the target of quality improvement programs. Examples of variation in care include management of anemia, vaccination practices, prophylaxis against venous thromboembolism (VTE), and testing for Clostridium difficile . Colectomy rates for UC vary according to geographic location in the United States, with rates in the West and Midwest regions threefold higher than in the Northeast. There is even variation in the care provided by IBD experts. A study from 2007 analyzed the use of five medication classes among 311 children with newly diagnosed CD followed at 10 pediatric gastroenterology centers in North America. Median use was 56% for immunomodulators (range, 29% to 97%), 78% for prednisone (32% to 88%), 29% for antibiotics (11% to 68%), 64% for aminosalicylates (18% to 92%), and 8% for infliximab (3% to 21%). There was statistically significant intercenter variation in use of all therapies, which remained significant even after adjustment for demographic and clinical factors.

Studies demonstrating disparities in care according to race and socioeconomic status (SES) provide additional examples of variation in care. A systematic review reported race- and SES-based disparities in the content of medical and surgical health care, use of inpatient and ambulatory medical care, adherence to medical therapy, and disease perceptions and knowledge. For example, most studies found lower rates of immunomodulators and infliximab use among minority patients. Among African Americans, there was lower utilization of specialist care and lower adherence to medical therapy. In addition, several studies identified race- and SES-based disparities in outcomes for IBD, including in-hospital mortality rates and health-related quality of life. Minority status and non-private health insurance were associated with lower rates of resection for CD, whereas higher median income was associated with lower mortality from surgery for CD. In UC, African-American race and Hispanic ethnicity were associated with lower odds for colectomy, and Medicaid (as opposed to private) insurance was associated with higher mortality from colectomy.

Finally, differences at the health care system level may also be associated with divergent patterns and outcomes of care. Low hospital volume has been associated with increased mortality following colectomy for UC (adjusted OR of 2.42; 95% confidence interval [CI], 1.26 to 4.63). In a Canadian study, UC patients hospitalized under a nongastroenterologist had higher in-hospital mortality compared with those admitted to a GI service (1.1% vs. 0.2%, P < .0001).

The need for improvement in the quality of IBD care cannot be overemphasized. UC and CD are common diseases that result in significant morbidity and costs. Disease algorithms have become even more complex with the approval of new agents, yet studies of comparative effectiveness are lacking. Practitioners vary in their familiarity with best practices. Finally, access to care varies according to race and SES. Recognizing the opportunities for quality improvement, the American Gastroenterological Association (AGA), the Crohn's and Colitis Foundation of America (CCFA), and a Canadian collaborative group developed indicators to measure processes and outcomes of IBD care. Common themes in these measures are summarized in Table 161.4 .

Variation in IBD care may be decreasing. In a prospective study of 2690 adult IBD patients followed at seven high-volume centers between 2012 and 2015, investigators found no variation among centers in the treatment of UC patients with immunomodulators alone, anti-TNF-α agents alone, or combination therapy. There was significant variation in the treatment of CD with immunomodulators and combination immunomodulator and anti-TNF-α therapy but not with anti-TNF-α agents alone. In a study from the Allegheny General Hospital in Pittsburgh, Pennsylvania, IBD specialists performed better than general gastroenterologists on a set of IBD quality measure (73.9 vs. 66.3, P = .001), but both groups scored above the recommended score of 60. The best example of IBD quality improvement in action has come from the Improve Care Now Network, a network of pediatric IBD centers. The collaborating centers measure and share data on quality of care and then develop and assess protocols to achieve selected goals. Using this model, the network achieved significant increases in the proportions of CD and UC patients with quiescent disease, as well as a significant increase in the proportion of CD patients not taking prednisone.

Chronic Disease Management in Inflammatory Bowel Disease

IBD is a lifelong, dynamic, multifaceted disease that requires comprehensive care. There is a plethora of treatment options and a multitude of disease- and treatment-related complications. Moreover, the disease has profound effects on the physical and emotional health of patients and on their personal aspirations. For these reasons, IBD management can be approached only through the framework of chronic disease management.

Patient empowerment and education are central to the chronic disease model. Caregivers must provide education, support, and open lines of communication. Patient trust in the physician is associated with improved adherence to IBD therapy, which is a surrogate for more prolonged remission. Understanding the patient's concerns, preferences, and expectations allows for a tailored approach and deepens the therapeutic relationship. Educational resources, in the electronic web (such as the CCFA website; www.ccfa.org ) or in other formats, may supplement office discussions and improve outcomes and cost of care. Two randomized controlled trials (RCTs) in UC have shown that, in comparison with standard management, patient education and self-management were associated with more prompt control of flares, fewer doctor and hospital visits, and cost savings.

A multidisciplinary approach is critical in addressing all aspects of the patient's illness. Recognizing the importance of the multidisciplinary approach, various experts have advocated the development of IBD comprehensive care units and have proposed criteria for such units. Involving the primary care physician is integral to overall coordination of care and may increase the implementation of preventive measures, such as vaccinations, assessment of bone health, and screening and intervention for cigarette smoking. Transitioning from pediatric to adult care is associated with nonadherence, increased emergency department use, and increased hospitalizations and should be approached in a gradual, well-planned manner. Close collaboration between the gastroenterologist and the colorectal surgeon is critical in optimizing patient status before surgery and planning for medical therapy to reduce the risk of postoperative recurrence of CD. Developing a network of consultants with an interest in IBD is important in managing the extraintestinal manifestations of the disease. Nutrition experts assist in the management of malnutrition and deficiencies of vitamins and micronutrients and educate patients on proper diets according to disease state (such as active or quiescent disease, or after extensive resection). Stoma nurses provide education and support to patients and their families. As with other chronic diseases, stress, depression, and anxiety are common in IBD patients. Psychiatric comorbidity appears to be associated with poorer clinical outcomes and greater health care costs in patients with IBD. Treatment of these disorders may improve general and emotional well-being and may even improve IBD outcomes.

An emerging trend in the care of IBD patients involves the increased use of mid-level providers. Given the need to improve care while controlling costs, specialist nurses, nurse practitioners, and physician assistants will inevitably assume greater roles in the management of patients and may even direct care in some domains. These providers can triage patients (reducing unnecessary emergency room or office visits), facilitate access to medical therapies, and make appropriate referrals to primary providers, collaborating specialists, and mental health professionals. Studies are beginning to examine the role of IBD-specialist nurses. A study from Norway reported that, in comparison with conventional follow-up, nurse-led follow-up produced similar outcomes in terms of hospitalizations, surgery, sick leave, performance of endoscopic procedures, and number of additional telephone consultations. Moreover, nurse-led follow-up was associated with a significantly faster treatment upon relapse.

Principles of Medical Therapy

UC and CD are lifelong diseases with a relapsing and remitting course. Consequently, induction of clinical remission followed by maintenance of remission emerged as the main goals of therapy. Additional goals include improved quality of life, prevention of disease- and treatment-related complications (including opioid dependence and excessive exposure to diagnostic radiation), restoration and maintenance of optimal nutritional status, optimization of preoperative status, and, in the case of CD, prevention of postoperative recurrence.

Traditionally, the selection of inductive therapy has been based on symptoms and laboratory indicators. This approach is endorsed by all existing guidelines. In clinical practice, disease activity is classified qualitatively. In UC for example, activity is classified as mild, moderate or severe, according to the Truelove-Witts criteria (see Table 161.3 ). Similarly, the American College of Gastroenterology developed operational definitions for CD activity. Activity is categorized into four states: (1) remission: asymptomatic disease off steroids; (2) mild-moderate disease: mild symptoms without manifestations of dehydration, systemic toxicity, abdominal tenderness, painful mass, intestinal obstruction, or greater than 10% weight loss; (3) moderate-severe disease: prominent symptoms of fever, significant weight loss, abdominal pain or tenderness, intermittent nausea or vomiting but without obstruction, or significant anemia; and (4) severe-fulminant disease: persistent symptoms despite the introduction of conventional corticosteroids or biologic agents as outpatients, or high fevers, persistent emesis, cachexia, significant peritoneal signs, or evidence of obstruction or abscess.

These classification schemes were inadequate for regulatory purposes. Therefore disease activity indices had to be developed and validated for therapeutic trials. These indices incorporated symptoms, signs, laboratory results, and, sometimes, endoscopic activity. The Crohn's Disease Activity Index (CDAI) was for many years the preferred instrument to assess CD activity in the context of clinical trials. However, it has long been known that the CDAI correlates poorly with endoscopic activity. Endoscopy offers the most objective assessment of UC severity as well. As a result, there is now broad consensus that assessment of drug efficacy in clinical trials must incorporate endoscopic response.

The lack of correlation between CDAI and endoscopic severity is not surprising to clinicians who frequently see coexistent irritable bowel syndrome masquerading as active CD. Consequently, endoscopy is necessary not only in clinical trials but also in daily practice to objectively assess the inflammatory burden. The situation is somewhat different for UC, where symptoms generally correlate well with endoscopic activity. Hence treating most patients based on symptoms (i.e., without endoscopic assessment) seems appropriate in clinical practice.

Numerous studies have shown that, regardless of the agent used, endoscopic healing is associated with improved long-term outcomes in both UC and CD, including sustained clinical remission, steroid-free remission, and reduced hospitalizations and surgeries. However, even with potent therapies, such as the combination of an immunomodulator and an anti-TNF-α agent, endoscopic healing occurs only in a minority of patients. The achievability of healing may also be limited by the nature of the disease process. For example, CD of longer duration responds less well to anti-TNF-α therapy than newly diagnosed disease. Finally, even if achievable, endoscopic healing would be desirable only insofar as it was cost effective. Notwithstanding these issues, there is growing consensus that endoscopic healing, besides being a trial endpoint, should also be a treatment goal in standard clinical practice.

Traditionally, the selection of therapy has been based solely on the severity of symptoms in real time. This approach does not take into account long-term risks, such as stricturing and fistulizing complications in CD and colectomy in UC. To this effect, the AGA has added long-term risk in its clinical decision tools for both CD and UC. For example, immunomodulators and/or biologics are recommended in patients with severe endoscopic disease activity (regardless of symptoms). Similarly, the European Crohn's and Colitis Organization (ECCO) has incorporated risk in some recommendations. For example, ECCO advocates early immunomodulator or anti-TNF therapy in patients with markers of poor prognosis, such as young age at diagnosis, extensive disease, requirement for steroids at diagnosis, or perianal disease at diagnosis. Future guidelines will likely incorporate endoscopic severity (a predictor of both short- and long-term risk) and other predictors of long-term risk in decision making. Current and evolving principles of medical therapy are summarized in Table 161.5 .

Classes of Medical Therapies

An ever-expanding number of agents are available for the therapy of UC and CD. Some agents are used for both UC and CD, whereas others are effective for only one of the two diseases. For each medication, it is important to understand its relative positioning in the induction and maintenance phases of therapy, as well as to recognize potential adverse effects ( Tables 161.6 to 161.8 ).

Medical Therapies