The Natural History of Inflammatory Bowel Disease

1. Approximately one third of patients has chronically active disease or fluctuating disease and hence this is the cohort most likely to be treated by more aggressive immunomodulating therapy. However, treating beyond symptom remission to achieve mucosal healing as well, will very likely increase the use of immunomodulating medications

2. Mortality is increased in Crohn’s disease compared to the general population. Surgery, especially the first 30 days from surgery, is associated with mortality

3. In general, at least from North America, there seems to be a recent trend toward reduced hospitalization rates. Hospitalization rates are dependent on the prevalence of disease, the approaches to settling disease in an outpatient setting and also the nature of the health care system and availability of access to inpatient therapy

4. In the pre biologic era, from population based studies, approximately 40–50 % of subjects will have undergone intestinal surgery within 10 years from diagnosis and the risk of postoperative recurrence may be about 50 % by 10 years. Surgery rates though are clearly falling. The drop in surgery rates began even before the introduction of biological therapy

5. In Crohn’s disease the location of disease is mostly stable over time with 15–20 % having a change in disease location when followed over time. While ileitis has been a hallmark of Crohn’s disease, in population based studies the colon (either alone or in combination with small bowel disease) is involved more often at a rate of up to 75 %. Data from some countries suggest that ileal disease location is more likely to be associated with complicated disease such as development of fibrostenosing or penetrating disease while data from other countries suggest that colonic disease is more likely to be associated with complicated disease. At diagnosis up to one third have evidence of fibrostenosing or penetrating disease but by 20 years half of all patients had these complications. Since there is a lack of uniformity in examining for disease behavior at all sites at time of diagnosis, the true incidence of evolution over time to complicated behavior is unknown

6. The main difference in pediatric onset disease versus adult onset disease is a higher prevalence of upper gastrointestinal disease in children; however, much of this may be related to the fact that pediatricians are more likely to pursue upper endoscopy in newly diagnosed children with Crohn’s disease compared to the use of upper endoscopy in newly diagnosed adults. While it has been shown that disease onset prior to age 40 is more aggressive than after age 40 years it seems that pediatric onset disease (age less than 17 years of age at diagnosis) is in fact not more aggressive than adults presenting prior to age 40 years

7. Mucosal healing seems to be a desired effect of therapy and its presence may reduce some negative outcomes. Whether it can ultimately be associated with long term reduction in complications (strictures or fistulas), surgery, and other comorbidities remains to be proven. Symptoms may not always be associated with active inflammation; hence in the setting of increased symptoms it is important to document the presence of active inflammation either with endoscopy or surrogate markers before increasing anti-inflammatory or immunomodulating therapy

Table 27.2

Summary statements for disease course in UC

1. Mortality rate is not increased in UC compared to the general population, but surgery is associated with mortality in UC

2. Hospitalization in a hospital with low volume rates for colectomy in UC are associated with higher mortality rates

3. Surgery rates at 10 years from diagnosis are approximately 10 % from recent studies which is much lower than reported colectomy rates in studies completed prior to 1990

4. Short term colectomy rates in severe hospitalized UC have remained stable at 27 % for several years

5. Pancolitis is a risk factor over lesser disease extent for requiring surgery

6. Children seem to have higher rates of extensive colitis at diagnosis than adults. There also seems to be higher rates of colectomy in children than for adults (i.e., at least 20 % at 10 years) and perhaps this reflects a higher rate of extensive disease

7. While mucosal healing has become a study endpoint for clinical trials in UC it is unclear if treating an asymptomatic patient until mucosal healing is achieved will change prognosis and it would require treating a large number of persons who otherwise had no symptoms. However, many patients have symptoms that are unrelated to having active inflammation. Hence, when a patient presents with symptoms it is important to document active inflammation either with endoscopy or surrogate markers before increasing anti-inflammatory or immunomodulating therapy

Mortality

In ulcerative colitis (UC) mortality rates in the middle of the twentieth century were as high as 22 % but fell steadily to approximately 5 % in the 1960s. Rates had been the highest for those with a severe first attack, those operated on emergently compared to those operated on electively and among those over age 60 [2–4]. The magnitude of the rates of mortality have dramatically changed but the variables that pose risks for adverse outcomes have remained the same over the past 50 years. Being early in the disease course, undergoing surgery, particularly emergent surgery, and older age were all shown to be mortality risk factors in a recent population based analysis of mortality rates in Manitoba, Canada [5].

In this study assessing mortality in all persons in Manitoba with IBD between 1984 and 2012 Crohn’s disease (CD) was associated with an increased risk for death [hazard ratio, HR of 1.26 (95 %, CI, 1.16–1.38)] but the HR for UC was 1.04 (95 % CI, 0.96–1.12). The lack of an increased death rate among persons with UC compared to controls suggests that either the disease is being diagnosed earlier, in less ill persons, or that it is being managed better, or both. Variables that increased the risk for death significantly compared to controls included the first year from diagnosis in both CD and UC, and the greatest risk for death in both CD and UC was within the first 30 days following GI surgery. Among IBD cases alone male sex, and older age at diagnosis, as well as increasing comorbidities significantly impacted on mortality in both CD and UC. The increase in mortality in CD compared to controls and the lack of increase in UC relative to controls generally mirror that reported elsewhere in the literature (reviewed in ref. [5]); however, there is heterogeneity by jurisdiction.

Surgery

While surgery can lead to prolonged disease free states for many with CD [6] and can enhance the quality of life for many with UC, it is, nonetheless, a disfiguring procedure accompanied by risks, a marked impact on daily activities and quality of life and as noted above, even in our modern era, occasionally death. So even while a well-timed surgery can be a critically important positive intervention to get a patient’s life back on track, most of the time surgery is considered a marker of failed therapy-failure to treat soon enough, or aggressively enough, or simply a failure of the available medical approaches in their effectiveness. Sometimes the person’s disease just doesn’t respond regardless of how timely and comprehensive the medical approach is that is taken.

CD: Surgery rates prior to 1970 were nearly 75 % [7, 8] which markedly influenced the long held view that surgery was a near inevitability at some time in the course of a patient’s life. Even in a later era but prior to the availability of biologic therapy, approximately 40–50 % of subjects in population-based cohorts underwent intestinal surgery within 10 years from diagnosis. Reoperation rates were up to 50 % by 10 years [9]. The most recent data from Canada and the UK suggest that surgical rates were falling prior to the advent of biologic therapy, and continue to fall. For instance 5-year surgery rates by years of CD incidence were 59 % (1986–1991), 37 % (1992–1997), and 25 % (1998–2003) from Wales [10] and 30 % (1988–1995), 22 % (1996–2000), and 18 % (2001–2008) from Canada [11]. In Canada infliximab became available only in 2001 for CD and surgery rates were already falling before its availability.

Small bowel disease, perineal disease, penetrating disease, and fibrostenosing disease all increase the likelihood of requiring surgery. Reoperation rates are also impacted by disease phenotype, but studies extended out over years and that cover decades of biological therapy use may change some of these paradigms. As an example, it was always considered that persons with isolated CD colitis would have lower reoperation rates in those who have their entire colons removed and have an end ileostomy. However, when survival curves are extended far enough out, rates of reoperation among proctocolectomy patients may approach that of those treated with segmental resection [12].

UC: Surgery rates in UC were as high as high as 30 % at 10 years disease duration prior to 1990; however, in most modern studies at 10 years from diagnosis surgery rates are approximately 10 % [13]. The lower rates of colectomy in recent times might reflect improved therapy but might also reflect greater access to colonoscopy and greater diagnoses of subjects with milder disease. Interestingly, a population based study on colectomy rates in UC from Manitoba Canada found no reduction in early colectomy (within 90 days of diagnosis) by the era of diagnosis (different than overall colectomy rates which have been falling by era of diagnosis) [14]. One implication is that advances in medical therapy of acute severe colitis have not been widely experienced or have simply not been sufficiently effective. A recently published review of short term colectomy rates in severe hospitalized UC suggested that the short term colectomy rate has remained stable at approximately 27 % [15]. Acute severe colitis in patients with UC still represents a condition with a high early colectomy rate and a measurable mortality rate. However, overall colectomy rates have been declining by era in Manitoba. Ten-year colectomy rates decreased significantly over time (12.2 % [era of incident cases 1987–1991], 11.2 % [1992–1996], 9.3 % [1997–2001], p = 0.014) [14].

Key predictors of colectomy in UC include age and disease extent [13]. Colectomy rates are higher in children than adults (i.e., at least 20 % at 10 years) perhaps reflecting a higher rate of extensive disease in children. Hospital volume has an impact on surgical outcomes. In a National Inpatient Survey study from the US hospitals with low volume colectomy rates were associated with an increased risk of death (adjusted OR relative to high-volume, 2.42; 95 % CI, 1.26–4.63), similar to medium-volume hospitals (OR, 2.02; 95 % CI, 1.02–4.01) [16]. Hence, centralizing IBD care in specialty centers may not only reduce surgery rates but also surgical mortality rates. It will also be of interest to determine how the introduction of therapy with monoclonal antibodies to tumor necrosis factor for moderate to severely active disease impacts on colectomy rates over the next 10 years.

Hospitalization

CD : Trends in hospitalization rates for CD from Europe and North America have been inconsistent, likely owing as much to differences in health systems as in disease incidence. For instance, in Copenhagen County, Denmark between 1962 and 1987, 83 % of patients with CD were admitted at least once within the first year after diagnosis, and then admissions decreased over the next 5 years to a steady rate of about 20 % per year [17]. However, the local treatment policy in effect at the time, encouraged hospitalization for more expeditious diagnosis and management, and so hospitalization in that era had a different implication than of a more recent era. In a European multi-country referral center prospective follow-up study over 10 years from 1991 (the EC-IBD), the cumulative risk of overall hospitalization was 52.7 % at 10 years from diagnosis, but with considerable differences between countries [18].

In a recent population-based study from In Manitoba, Canada where there is universal health care coverage for all residents for persons with CD diagnosed between 1988 and 2008, the highest hospitalization rates were within the first year of diagnosis and there were no differences among those diagnosed during 1988–1995, 1996–2000, or those diagnosed after 2001 [11].

In a population-based inception cohort from Olmsted County, Minnesota (diagnosed 1970–2004 and followed through mid-2009), with a median follow-up of 11.8 years, 71 % of CD patients were hospitalized at least once, and the cumulative risk of hospitalization was 62 % at 5 years and 71 % at 10 years after diagnosis [19]. Factors associated with time to first hospitalization included ileocolonic disease, small bowel, or upper gastrointestinal disease (relative to colonic only), and fibrostenotic or penetrating complications at baseline. Sonnenberg and colleagues analyzed the US Veterans Administration (VA) database for CD hospitalization patterns among US military veterans over a 32-year period [20]. The hospitalization rate peaked in the late 1980s, decreased through the 1990s, and remained relatively stable over the last 6 years of the study period. Trends in VA data are very dependent on American military activity and availability of heath care to veterans outside of the VA system.

In contrast to these studies, data from Kaiser Permanente Northern California (a health maintenance organization) suggested that hospitalization rates for CD decreased by 33 % between 1998 and 2005 [21]. The decrease in hospitalization rates in the Kaiser study and the stable rates in Manitoba (where disease prevalence rates are rising) support the possibility that more aggressive medical therapy accounts for these findings, or simply that better health care access for subjects in those health systems maintains their disease in a better state with a reduction in need for hospitalizations . These trends also may reflect health systems with greater incentives to capitate costs and control inpatient management. Hence, examining trends in hospitalizations requires an understanding of the individual health system, the prevalence of the disease within the population and management approaches (i.e., availability of different medical therapies and aggressiveness of using them).

UC : In a Manitoba population-based cohort of newly diagnosed persons with UC in 1987 followed for 15 years, 31 % were admitted at least once to hospital for an IBD-specific diagnosis, and, of those admitted, 51 % were readmitted at some point [22]. In the Kaiser Permanente of Northern California cohort 20 %, underwent colectomy during their initial hospitalization [23]. By 1 year after initial hospitalization, 29 % of those who had not undergone colectomy at first hospitalization, were rehospitalized for UC (most of which occurred within the first 3 months post discharge) and an additional 10 % required colectomy. By 5 years after initial hospitalization, 39 % of those who had not undergone colectomy after initial presentation, were rehospitalized for UC and an additional 15 % required colectomy. Hence, for patients with UC that is sufficiently severe such that they get hospitalized, for those that do not undergo colectomy there is nearly a one in two chance of being hospitalized again.

Superimposed infections remain a risk factor for hospitalizations and colectomy in UC patients. In a study from the Cleveland Clinic, patients with Clostridium difficile infection had significantly more UC-related emergency room visits in the year following initial infection (37.8 % vs. 4 %), and significantly higher rates of colectomy 1 year following the index infection admission (35.6 % vs. 9.9 %), than those without the infection [24]. Similar findings were reported from Mount Sinai, New York (2004–2005) where subsequent UC-related hospitalizations and colectomy rates 1 year after initial hospitalization were higher in patients with Clostridium difficile infection than those without infection [25].

Phenotype and Its Impact on Disease Course

Disease Localization

There is ample evidence that more extensive colon involvement in UC is associated with more aggressive disease and worse outcomes in terms of likelihood for developing colon cancer or having colectomy [26–31]. Further defining disease extent in UC is not a difficult ask especially with the widespread availability of colonoscopy. It is in CD where there is more discussion as to ways to define disease extent and its implications. The localization of disease in CD is divided as per the widely used Montreal Classification into ileal, colonic, ileocolonic and upper gastrointestinal disease [32]. The relative distributions of disease location vary by jurisdiction; however, in most studies ileocolonic is the most common site for disease. Whatever is reported as the relative distribution of disease by location remains relatively stable over time. The IBSEN study from Norway reported on changes in location over 5 years in 14 % [33]. This is not dissimilar from what Louis et al. reported in a referral population from Belgium where location changed in only 16 % by 10 years and 20 % changed overall [34]. Another oft reported finding is that of a high prevalence of upper gastrointestinal tract disease in pediatric CD. In a French population based cohort of children diagnosed between 1988 and 2002, at diagnosis, 63 % had ileocolonic disease and 36 % had upper gastrointestinal disease [35]. This is such a high prevalence of upper gastrointestinal tract disease that it suggests that perhaps they were considering any histologic inflammation as being pathological. The European guidelines on diagnostic evaluation for pediatric IBD includes an upper endoscopy for all which is much different than the approach in adults [36]. Has this practice biased what is perceived to be a higher rate of upper GI tract disease in children than adults?

Disease Behavior

In the Norwegian IBSEN cohort, at diagnosis, 28 % had fibrostenosing disease and 12 % had penetrating disease, leaving 61 % with inflammatory disease [33] with a change in behavior at 5 year follow-up of 14 %. Complicated disease (both fibrostenosing and penetrating disease ) occurred in 86 % of those with isolated ileal disease, 30 % of those with colonic disease and in 60 % of those with ileocolonic disease. In the EC-IBD study at diagnosis 16 % had fibrostenosing disease and 8 % had penetrating disease and 2 % had both. Hence, 74 % had inflammatory disease, a higher proportion than the Norwegian cohort [37].

In a referral population from Belgium it was reported that 90 % of patients present with inflammatory disease but 69 % and 88 % had fibrostenosing or penetrating disease by 10 and 25 years, respectively [34]. In this study 46 % changed behavior over 10 years. Penetrating disease was more common from a colonic only or ileocolonic location than from an ileal only location [34] (the exact opposite from the IBSEN data). In a French referral cohort the 20 year actuarial rates of having inflammatory disease, fibrostenosing disease and penetrating disease were 12 %, 18 % and 70 %, respectively [38]. Fibrostenosing disease was associated with ileal disease (HR 2.5, 95 % CI 1.9–3.3) or jejunal disease (HR 3.2, 95 % CI 2.2–4.7) and penetrating disease with colonic disease on univariate but not multivariate analysis. In the Olmsted County cohort diagnosed between 1970 and 1993, 50 % had fistulas by 20 years [39] and fistulas were more likely to be seen in those with ileocolonic disease than in patients with disease at other sites. So how can we reconcile that studies from Belgium and France suggest that colonic disease location is the key to having penetrating disease and from Norway suggests that small bowel disease is associated with penetrating disease? To further complicate the issue the American study suggests that it is an ileocolonic site most associated with penetrating disease.

There is a notion that being diagnosed with CD in childhood is predictive of a more aggressive course of disease. In a Danish study of childhood onset Crohn’s disease the mean yearly operation rate was 13 % with a cumulative probability of surgery at 20 years of disease of 47 %, which is not dissimilar to what is seen in adults [40]. In a French pediatric study complicated behavior occurred in 29 % at diagnosis (including 4 % with penetrating disease) and 59 % at follow-up (including 15 % with penetrating disease). What then has driven the notion that children have more aggressive disease and worse outcomes than persons diagnosed as adults?

The major flaw in all of these reports regarding phenotype regardless of the population being studied is a lack of uniformity in assessing the location and behavior of disease at diagnosis in most centers. Some patients have early surgery where previously undiagnosed fistulas are identified and hence early on are labeled as having penetrating disease whereas their diagnostic imaging studies may not have identified this. The lack of uniformity in methods used to establish the extent and behavior of disease limits the reliability of phenotype data established at time of diagnosis. The variation of disease behavior from even a single center is best reflected in a large French cohort of 2008 patients evaluated within 3 months of diagnosis between 1978 and 2002 [41]. In each 5 year period the prevalence of fibrostenosing disease ranged from 5–26 % with a high of 26 % in 1978–1982 and a low of 5 % in 1998–2002, likely accounted for because some with fibrostenosing disease either had surgery for stricture segment removal or evolved to penetrating disease. The prevalence of penetrating disease ranged from 29–54 %. Another example is an Olmsted County cohort where complicated disease behavior (fibrostenosing or penetrating disease ) at diagnosis was noted for 18.6 % [42]. Within the first 90 days of follow-up an additional 4.6 % developed fibrostenosing complications and an additional 14.1 % developed penetrating complications. Hence, 37.3 % of the cohort were documented with complicated behavior at 90 days. There was no information in this study on which diagnostic tests were utilized at diagnosis or within this 90-day period of follow-up. It is likely that the doubling of complicated disease within 90 days reflected delayed testing that brought these complications to light. It is unlikely that advancement of phenotype occurred within such a short period.

With this in mind, Israeli et al. undertook a study where all charts were reviewed from a single referral center practice at a median disease duration of 11.1 years and an analysis was undertaken for the location and behavior patterns of disease at diagnosis and over time in children and adults [43]. Within 1 year of diagnosis, the proportion of patients with upper gastrointestinal involvement and ileocolonic location was higher in those diagnosed before age 17 (18.4 %) than those diagnosed between 17 and 40 years (10.2 %) or those diagnosed after age 40 (2.7 %). The rate of patients that underwent imaging with an upper endoscopy in the youngest group (37.4 %) was more than twice as high when compared to the other groups (those diagnosed over age 17, 14.3 %, p < 0.01, and those diagnosed over age 40, 16.9 %, p < 0.01). After adjusting for imaging testing, the difference in likelihood of having upper gastrointestinal disease between those diagnosed under age 17 and those diagnosed age 17–40 was no longer statistically significant (but regardless of imaging children had significantly more upper gastrointestinal disease than those diagnosed over age 40 years). Children had more extensive imaging than adults at the time of diagnosis. Despite more extensive imaging, complicated disease behavior (fibrostenosing or penetrating disease) was less prevalent in children particularly. At the last follow-up complicated disease behavior was similar regardless of age at diagnosis being under 17, under 40 or over 40 years. To further dispel the notion that being diagnosed in childhood carries a worse prognosis at final follow-up IBD-related abdominal surgery rates were significantly lower for children than those diagnosed after age 17 (OR = 0.63, 0.41–0.98) but not compared to those diagnosed after age 40 years (OR = 0.71, 0.40–1.27). The conclusions of this study were that studying the phenotype of CD among different cohorts has to account for the differing patterns of diagnostic imaging investigations. Further, while children are at increased risk of pan-enteric disease, this does not lead to them being more likely to have more complicated disease or surgery.

Can the Disease Phenotype Predict Outcomes?

In a large French cohort ileal location was associated with an increased risk for surgery (HR = 2.78, 95 % CI 2.19–3.15) [41] and absence of rectal involvement was associated with a decreased risk for surgery in this cohort (HR = 0.34, 95 % CI 0.27–0.43). In the Norwegian IBSEN cohort being less than 40 years at diagnosis, having fibrostenosing or penetrating disease and terminal ileum disease were associated with higher risk for surgery [44]. It was also shown that age less than 40 years at diagnosis and need for corticosteroids during the first presentation in IBSEN and age less than 40 years at diagnosis in the EC-IBD were associated with higher relapse rates [37, 44]. While it is useful to know that persons less than 40 years at diagnosis have more aggressive disease than older persons at presentation, since the peak age of incidence is the third decade and the majority of CD patients have presented by age 40, more refined predictors than age less than 40 years would be required to help identify those requiring earlier, more aggressive therapy. No studies have reported that sex impacts on disease outcome except for a Swedish study that suggested that females have a higher likelihood of postoperative recurrence [45].

Only gold members can continue reading. Log In or Register to continue