Article Text

Download PDFPDF

Pathogenesis of postoperative recurrence in Crohn's disease
  1. Tasneem Ahmed1,
  2. Florian Rieder2,
  3. Claudio Fiocchi1,2,
  4. Jean-Paul Achkar1,2
  1. 1Department of Gastroenterology, Digestive Diseases Institute, Cleveland Clinic, Cleveland, Ohio, USA
  2. 2Department of Pathobiology, Lerner Research Institute, Cleveland Clinic, Cleveland, Ohio, USA
  1. Correspondence to Jean-Paul Achkar, Cleveland Clinic, Desk A31, 9500 Euclid Avenue, Cleveland, OH 44195, USA; achkarj{at}


The majority of patients with Crohn's disease (CD) require surgery during the course of their disease, but such surgery is typically not curative. Although some studies suggest that the disease state is theoretically reset to its earliest phase following surgery, disease phenotype and natural history of CD do not change significantly after surgery, leading to high rates of recurrence. Factors predisposing to this recurrence are not well defined, so there is a need for and a unique opportunity to develop a better understanding of the pathogenesis of recurrent inflammation and associated risk factors after an ileocolic resection. This paper reviews the postoperative disease outcome and evolution based on defining the combination of the patient's microbial flora, environmental exposure history, immune response and genetic make-up.

  • Crohn's
  • postoperative
  • recurrence
  • pathogenesis
  • risk factors
  • bacterial interactions
  • crohn's disease
  • genetics
  • immune response

Statistics from


The majority of patients with Crohn's disease (CD) will require surgery at some point during the course of their disease, but such surgery is typically not curative.1 After ileocolonic resection and anastomosis, endoscopic recurrence of disease arises in the neoterminal ileum in 30% of patients after 3 months and in up to 80% of patients after 1 year.2 3 Clinical recurrence may be difficult to detect early after surgery as symptoms may be due to postoperative changes such as choleretic diarrhoea. Endoscopic examination is therefore useful for both diagnosis and prognostication as the severity and extent of endoscopic recurrence has been shown to predict the length of time to clinical recurrence.3 With this in mind, clinical recurrence rates have been reported to be as high as 20–30% at 1 year with a 10% increase in each of the subsequent years.1 The need for repeat surgery occurs in 15–45% of patients at 3 years, 26–65% at 10 years and 33–82% at 15 years.1

The presence of granulomas, perforating disease phenotype, ileal or ileocaecal resection with ileocolic anastomosis and smoking have all been suggested as potential risk factors predisposing to disease recurrence.1 4 5 To date, however, the presence of an ileocolic anastomosis and smoking are the only well-replicated predictors of postoperative recurrence, making prevention a challenging problem.1 4 5 Despite multiple studies, the need for medical prophylaxis and the choice of the best preventive regimen if therapy is pursued still remain unclear. This uncertainty about postoperative disease management stems in part from a need for a better understanding of the pathogenesis of recurrent inflammation and the associated risk factors after an ileocolic resection.

CD is a heterogeneous disorder caused by a complex interplay of microbial, environmental, immunological and genetic variables. It seems probable that the same factors which underlie the pathogenesis of CD at its early stages also contribute to disease recurrence in the postoperative setting (figure 1). In fact, the postoperative state in which an ileocolic resection with anastomosis is performed for intent of disease remission is an ideal setting in which to study these different variables, because one may argue that removal of the involved segment of bowel in part resets the disease to its earliest phases, providing a window to understanding factors which predispose to disease recurrence. This is especially important if one considers that prior studies have indicated that the original disease phenotype remains consistent postoperatively.6 Furthermore, the high rates of disease recurrence may lend further credence to the belief that the natural history of CD truly does not change after surgery. As an example, in figure 2, similar endoscopic changes of the terminal ileum in the same patient can be seen before first ileocolic resection (figure 2A) and 1 year postoperatively when disease had recurred (figure 2B). Identifying risk factors in the postoperative setting may therefore help to shed light on these same disease-inducing factors and the overall pathogenesis of CD. Moreover, a better understanding of the pathogenesis of CD in the postoperative setting is needed in order to tailor appropriate medical therapy.

Figure 1

Pathogenesis of postoperative recurrence of Crohn's disease.

Figure 2

Endoscopic changes of the terminal ileum in the same patient (A) before first ileocolic resection and (B) 1 year postoperatively when disease had recurred.

If primary and recurrent CD represent the same condition, one may predict disease outcome and evolution based on defining the combination of the patient's microbial flora, environmental exposure history, immune response and genetic make-up. This article will review how the pathogenesis of CD relates to disease recurrence with the goal of better understanding the postoperative CD state. A review of the current knowledge and future directions is shown in box 1.

Box 1

Postoperative recurrence of Crohn's disease: current knowledge and future directions

  • Endoscopic and surgical postoperative CD recurrence rates are high (80% at 1 year and 15–45% at 3 years).

  • The presence of an ileocolic anastomosis and smoking are the only well-replicated risk factors to date.

  • Bacterial flora, environmental factors, immune response and genetic variants are thought to be key factors in both disease pathogenesis and postoperative recurrence.

  • The natural history of CD after resection appears to remain stable.

  • Future directions in CD relate to prediction of disease outcome and evolution based on an individual's bacterial flora, environmental exposure history, immune response and genetic make-up.

Bacterial flora

The intestinal microbiota is recognised as a central factor in the pathogenesis of CD with several lines of evidence suggesting that bacteria play a role in the onset as well as perpetuation of intestinal inflammation. The composition of the intestinal microflora is complex and its role is subtly intertwined with genetic factors and the host's immune response. Evidence linking bacterial flora and intestinal inflammation includes the observation of relatively high concentrations of bacteria (>1012 organisms/g) in the normal distal ileum and colon, which is thought partly to explain the predominance of CD in these segments of the gastrointestinal tract.7 Furthermore, studies in patients with CD have demonstrated various changes in the luminal flora with a possible link to local inflammation.8 9 Increased numbers of Clostridia spp., Bacteroides vulgatus and Escherichia coli have been observed in mucosal biopsies of active inflammatory bowel disease (IBD).10 11 Likewise, other studies have reported the loss of commensal species such as Clostridium leptum, Eubacterium, Lactobacillus spp. and Bifidobacteria in patients with active IBD.10 11 Ott et al reported a loss in biodiversity by 30–50% when comparing healthy subjects with those with IBD.9

In addition, studies in a variety of animal models of IBD have shown that experimental colitis fails to develop under germ-free conditions, thus underscoring the pivotal role played by the gut microbial flora in the initiation of inflammation.12 Furthermore, animal models have also shown failure of activation of the immune system in the absence of commensal bacteria, again highlighting the relationship between the immune response and gut microbiota.12 13

Based on these lines of evidence, the bacterial flora is thought to play an important role in the pathogenesis of CD.14 However, one must keep in mind the extreme difficulty in analysing this microbial flora in a consistent fashion owing to the large numbers and diversity of bacteria which greatly contribute to the obstacles of clearly elucidating its exact role in disease pathogenesis.

The role of bacterial flora has been traditionally viewed as one where commensals incite a proinflammatory response that eventually evolves into full-blown CD. A recent alternative view proposes that some commensal bacteria play a protective role and may actually prevent gut inflammation. Segmented filamentous bacteria are an example of commensal species that can skew the composition of T helper cells and modulate the mucosal immune response.15 Polysaccharide A produced by Bacteroides fragilis can prevent or treat experimental colitis, and oral administration of live Faecalibacterium prausnitzii or its culture supernatants reduces inflammation in animals with colitis.16–18 These effects are presumably reproducible in humans, underpinning the concept that the gut microbiota shapes the intestinal immune response in both health and disease.18

Bacterial flora and postoperative recurrence

In trying to link the pathogenesis of postoperative CD recurrence with changes in the microbial flora, we must first recognise the importance of the faecal stream. In an important study, Rutgeerts et al demonstrated that five patients with CD who had undergone resection with a diverting loop ileostomy proximal to an ileocolic anastomosis had no endoscopic evidence of recurrence in the neoterminal ileum 6 months after surgery.19 However, after bowel continuity was restored, essentially all patients had rapid endoscopic recurrence in the neoterminal ileum. These findings suggest that the diversion of bowel contents protects the neoterminal ileum from recurrent inflammation and that bacterial flora and/or other components of the faecal stream play a key role in disease initiation. This hypothesis was further supported by a follow-up study in which three patients who had undergone ileocolonic resection with ileocolonic anastomosis and diverting ileostomy underwent infusion of their ileal effluent into the efferent limb of the ileostomy over an 8-day period and then underwent endoscopy of the ileum. The investigators found histological evidence of rapid recurrence of microscopic inflammation in the mucosa of the excluded ileum.7

In patients who have undergone an ileocaecal resection, loss of the ileocaecal valve leads to changes in the luminal environment. Animal models have shown that the ileocaecal valve prevents microbial reflux of colonic contents into the terminal ileum. A study by Neut et al in 61 patients who underwent ileocaecectomy showed that the neoterminal ileum became heavily colonised by colonic bacteria.8 Not all these patients developed endoscopic recurrence of disease, but the authors found that an increase in specific bacteria such as E coli, Enterococci, Bacteroides and Fusobacteria was associated with early disease recurrence.8 20 This finding may be explained by the theory that, with the loss of the ileocaecal valve, the neoterminal ileum becomes exposed to colonic bacterial species which are then recognised as ‘foreign’ by the local immune cells and trigger the host's immune response.

Another possibility is that bacteria associated with postoperative recurrence are more pathogenic with respect to increased adherence and penetrance. One such example is a strain of adherent invasive E coli (AIEC) which could act as a pathogen in CD through a number of different mechanisms including increased mucosal colonisation, adherence, replication and induction of tumor necrosis factor α (TNFα) secretion.21 In one study AIEC were recovered from 100% of biopsies of endoscopic lesions in postoperative patients.22 Additionally, AIEC were found in higher concentrations at 3 months after surgery compared with 1 year, suggesting a role in initiating postoperative disease recurrence.8 23 Alternatively, recurrence of postoperative CD may not be simply due to neocolonisation by pathogenic bacteria but may instead be related to a lack of protective commensal species. Supporting this second possibility is a report showing that a reduction of a major member of the Firmicutes, Faecalibacterium prausnitzii, is associated with a higher risk of postoperative recurrence of ileal CD.17

Given the above evidence regarding the role of intestinal bacteria in the pathogenesis of postoperative CD, antibiotic trials have been used to determine whether the risk of postoperative recurrence could be reduced. In a randomised double-blind placebo controlled trial, 60 patients were randomised to either metronidazole (20 mg/kg daily) or placebo for 3 months with treatment started within 1 week of a curative ileocolic resection.24 At 12 weeks, 12/28 patients (75%) in the placebo group had recurrent lesions in the neoterminal ileum compared with 12/23 (52%) in the metronidazole group (p=0.09). Although the overall endoscopic recurrence rates were not significantly different between the two groups, the metronidazole-treated group had significantly less severe endoscopic lesions. In addition, metronidazole was found to reduce clinical recurrence rates at 1 year (4% vs 25%), but reductions at 2 and 3 years were not significantly different.24 In a subsequent randomised placebo controlled double-blind trial, ornidazole (another nitroimidazole antibiotic) was also shown to reduce postoperative recurrence. Eighty patients were randomised to ornidazole 1 g/day or placebo for 1 year starting within 1 week of ileocolic resection. Ornidazole significantly reduced clinical recurrence rates at 12 months (38% of patients in the placebo group vs 8% of patients in the ornidazole group, p=0.0046) and endoscopic recurrence rates at 12 months were significantly reduced (79% of patients in the placebo group vs 54% in the treatment group, p=0.037). Significantly more patients discontinued treatment in the ornidazole group (12/38) compared with the placebo group (5/40), with side effects including nausea, vomiting, paraesthesias and neuropathy.25 The therapeutic effects of metronidazole and ornidazole seen in these studies suggest that anaerobic bacteria are important targets in disease pathogenesis, and also lend further credence to the role of bacterial flora in postoperative recurrence.

Given previous data suggesting a loss in commensal probiotic bacteria such as Lactobacillus and Bifidobacteria species contributing to the dysbiosis in the intestinal flora, probiotics have also been studied for prophylaxis in postoperative disease recurrence. Although probiotics have shown efficacy in patients with ulcerative colitis (UC) and pouchitis, this has not been replicated in CD.26 27 Three randomised placebo controlled trials comparing use of Lactobacillus with placebo in patients who had undergone curative surgical resection for CD did not demonstrate any effect on prevention of endoscopic recurrence at 3, 6 or 12 months or in reduction in the severity of lesions.28–30 However, a single blinded study reported only in abstract form used combination treatment of rifaximin for 3 months followed by VSL-3 (a probiotic combination of eight species of Lactobacillus, Bifidobacteria and Streptococcus) for 12 months compared with mesalamine and showed some benefit in prevention of endoscopic recurrence.31

While the intestinal microbiota plays an integral role in the pathogenesis of CD, further studies are needed to better understand this role and its association with the immune response. In the postoperative setting this is even more relevant as a better understanding of the changes in the bacterial flora which occur after an ileocolic resection may allow for better tailored medical treatment to prevent recurrence of CD.

The relationship between the bacterial flora and postoperative recurrence of CD is shown in box 2.

Box 2

Bacterial flora and postoperative recurrence

  • In the postoperative setting the microbial flora is altered through loss of the ileocaecal valve, altered integrity of the mucosal barrier and perhaps an imbalance between commensal and pathogenic bacteria.

  • The Human Microbiome Project will allow for better mapping of the intestinal microflora and its relationship with the immune response in individual patients and CD subgroups.

  • Further antimicrobial therapies can be developed to specifically target potentially pathogenic bacteria associated with recurrence.

  • Biological therapy could also be expanded based on specific immune responses associated with bacterial species implicated in recurrence.

Role of antibodies directed against bacterial peptides

In conjunction with the theory that an aberrant immune response to gut microbial flora is a key factor in the pathogenesis of CD, a variety of antimicrobial products have been shown to induce a systemic immune response in patients with CD.32–40 Serological markers linked to CD such as anti-Saccharomyces cerevisiae (ASCA), anti-outer membrane porin C (OmpC) of E coli, anti-Pseudomonas-associated sequence I2 (Anti-I2) and antibacterial flagellin cBir1 (Anti-cBir1) have been extensively studied in their role, both as diagnostic markers and prognostic indicators.32–40 Most recently, newer markers have been identified as antibodies to glycan epitopes found in the cell walls of bacteria in patients with CD: anti-mannobioside carbohydrate antibodies (AMCA), anti-laminaribioside carbohydrate antibodies (ALCA), anti-chitobioside carbohydrate antibodies (ACCA), anti-laminarin carbohydrate antibody (Anti-L) and anti-chitin carbohydrate antibody (Anti-C).32–40 The discovery of these antibodies supports the theory of an innate immune defect leading to loss of tolerance to certain luminal flora. Of note, these antibodies have been shown to be positive in up to 44% of patients in whom ASCA was negative.41

Several independent cross-sectional and prospective cohort studies have demonstrated a high discriminatory capacity for serological markers in distinguishing CD from UC as well as an association with and prediction of complicated CD behaviour defined as the occurrence of fistulae and strictures or need for surgery.32–40 This appears to be true for both a qualitative (marker positivity) and a quantitative (marker levels) immune response.32–40 In other words, the higher the amount of positive antibodies in the serum of a patient with CD or the higher the cumulative levels of the antibodies combined, the stronger the likelihood for and the earlier the occurrence of complicated CD behaviour and CD-related surgery.32–40 Likewise, studies have reported an association between Anti-I2 and fibrostenotic CD, Anti-OmpC and penetrating CD disease and Anti-cBir1 and both phenotypes combined.42 43 In addition, studies have reported an association between ASCA and ileal disease location.33–37 42 Based on these data, it appears that a strong immune response to microbial components confers a more severe disease course.

The exact role of antibody markers in the postoperative setting is not yet known. One study by Eser et al found no correlation between preoperative ASCA levels and risk of surgical recurrence.44 This study is reported in abstract form only at this time and no information is provided on endoscopic or clinical postoperative recurrence. Desir et al reported that a cohort of pediatric patients with CD positive for ASCA were more than twice as likely to experience a relapse during follow-up compared with ASCA-negative patients (OR 2.9, p<0.05).45 Although this study did not look at postoperative patients, it does suggest that serum antibodies could be used to predict postoperative recurrence as they have been shown to be associated with disease relapse. Moreover, serum antibody titres do not change with disease activity and, even though slight increases in titres over time have been reported,34 they are generally widely stable throughout the disease course.

As noted above, after an ileocaecal resection the neoterminal ileum is exposed to colonic microbial flora which induces an immune response. This immune response could result in the formation of new antibodies or a specific pattern of markers predictive of recurrence after the first surgery. One can postulate that a distinct marker pattern before the first surgery may indicate a higher risk for postoperative recurrence.

The relationship between serological bacterial antibodies and postoperative recurrence of CD is shown in box 3.

Box 3

Serological bacterial antibodies and postoperative recurrence

  • Studies have found associations between serological markers and disease location and complicated disease behaviour.

  • Use of serological markers for disease prediction is especially attractive as antibody titres do not change with disease activity and remain stable over time.

  • Production of these antibodies has not been studied at the tissue level, but measurement of mucosal antibody production may better reflect the actual immune response to the bacterial flora implicated in recurrence.

  • The new serological markers AMCA, ALCA, ACCA, anti-L and anti-C support the involvement of gut commensal flora in the postoperative recurrence of CD.

  • More studies are needed to define the value of serological markers in the postoperative period.

  • The association of serological with genetic markers may better predict the risk of postoperative recurrence than either marker alone.

Environmental factors

A variety of environmental factors such as smoking, hygiene, drugs and diet have been postulated to be associated with disease pathogenesis and flares in patients with CD.3 46 Here we discuss the potential effects of these factors in the postoperative setting.

The rise in the prevalence of IBD in Europe and North America in the second half of the 20th century and the lower incidence of both CD and UC in developing countries compared with industrialised nations implies that there may be strong environmental influences on the development of IBD.47 Changed dietary practices and increased hygiene are two factors thought to be involved in altering the intestinal microbial milieu from one with a high microbial load to one with a low microbial load.48 At birth the intestinal lumen is largely sterile followed by rapid colonisation in the immediate neonatal period. The hygiene hypothesis proposes that lack of exposure to select microbial agents in childhood increases the risk of developing certain autoimmune and inflammatory diseases because the exposure to pathogens or parasites, especially early in life, stimulates protective immunity that prevents later aggressive immunological challenges.47 49 Thus, the increased hygiene in developed countries combined with the delayed or weak exposure to bacteria is postulated to have led to the rising incidence of CD and other autoimmune disorders such as rheumatoid arthritis and type I diabetes in Western populations.20 47 Also in support of this theory is the observation from a large database in Manitoba that patients with CD are less likely to have lived on a farm or to have drunk unpasteurised milk and to have had a pet before the age of 5 years than controls without IBD.50

Diet is another key environmental factor that has been suggested to potentially contribute to the pathogenesis of IBD.20 Decreased consumption of fibre, increased consumption of sugars and meat and increased use of refrigeration in food preparation and storage have been hypothesised to alter the composition and virulence of some bacteria, thus contributing to the development of IBD.20 This again is best exemplified by the rise of IBD in Western countries in the second half of the 20th century and, more recently, in Eastern European and Asian countries as they adopt more Western-like dietary practices. Studies seeking to link diet and IBD directly have been generally inconclusive; however, one study by Persson et al suggested that frequent fast food intake confers a 3–4-fold greater risk of IBD.51 Some studies have suggested an elemental diet to decrease inflammation in IBD and decrease the risk of surgery; however, these results are inconclusive.52 Currently, there have been no proven dietary approaches to reduce the risk of IBD.

Given the changes in microbial flora seen following surgery, as discussed above, it is intriguing to consider possible further effects of hygiene and diet in the pathogenesis of postoperative recurrence, especially if one considers that removal of the involved segment of bowel resets the disease to time zero. However, these issues have not been studied in the postoperative setting.

Medications have also been suggested to be associated with IBD. Oral contraceptives and non-steroidal anti-inflammatory drugs (NSAIDs) have been studied and found to be associated with disease relapse.46 50 While a direct causal relationship between oral contraceptives and IBD has not been established, a meta-analysis by Godet et al showed a modest increase in the RR of CD in women taking oral contraceptives compared with controls.53 With regard to NSAIDs, an association with IBD has been established in patients in remission who relapse with the use of these anti-inflammatory drugs.46 48

Smoking and postoperative recurrence

Based on solid epidemiological data and observational studies, the best defined and most important environmental factor in CD is smoking. Smoking is strongly associated with an increased risk for CD and with a poorer prognosis, as defined by higher relapse rates, a more aggressive disease course and increased risk of postoperative recurrence.50 Cottone et al reported that 6 years after surgery 60% of non-smokers, 41% of ex-smokers and 27% of active smokers were free of clinical recurrence.54 In addition, the investigators found smoking to be the only significant predictor of surgical recurrence (HR 2.0; 95% CI 1.2 to 2.3).54 Sutherland et al reported that repeat surgery at 5 years was required in 20% of non-smokers compared with 36% in smokers; these figures rose to 41% and 70%, respectively, at 10 years.55 Lastly, a recent meta-analysis by Reese et al demonstrated a twofold increase in clinical relapse after surgery in smokers compared with non-smokers as well as a 2.5-fold increase in the risk of further surgery by 10 years.56 While studies have repeatedly shown that smoking is associated with higher relapse rates postoperatively, the exact mechanism underlying this association has not been identified but several factors have been postulated. For example, it is thought that smoking may have immunosuppressive effects influencing cellular and humoral immunity.57 Another plausible explanation is that the vascular changes caused by components of tobacco smoke (such as oxidising chemicals) may have prothrombotic effects which in turn could exacerbate ischaemia of the bowel wall.54 57 In addition, some have speculated that smoking may have a role in the alteration of the composition of mucus in the gut.54 These potential explanations almost certainly apply to both primary disease and recurrent disease.

The relationship between environmental factors and postoperative recurrence of CD is shown in box 4.

Box 4

Environmental factors and postoperative recurrence

  • Other than smoking, there is scant evidence that environmental factors such as hygiene, diet or medications (NSAIDs and oral contraceptives) impact on postoperative recurrence.

  • If bacterial colonisation with different species depends on environmental conditions, manipulation of the intestinal microflora through changes in environmental factors could prevent or eliminate disease recurrence.

  • Studies are needed to determine if altering the patient's geographical location (ie, relocating from a colder to a warmer climate or vice versa), adopting different hygiene practices, changing dietary habits or avoiding certain medications affects postoperative CD recurrence rates.

Immune response

In the healthy gut the mucosal immune system maintains a balance between pro- and anti-inflammatory factors, thereby allowing an effective defence against luminal pathogens while simultaneously preventing an overactive immune response to dietary and commensal microbial agents. In IBD this balance is altered with a shift towards a proinflammatory state owing to a dysregulated immune system with a complex interplay between effector and regulatory T cells and immunosuppressive and proinflammatory cytokines.58 CD is clearly associated with the T helper 1 (Th1) pathway characterised by increased production of interferon γ (IFNγ) and TNFα.58 CD has recently also been associated with the Th17 pathway characterised by an increased production of interleukin 17 (IL-17) which is conditioned by increased levels of IL-23.59 IL-17 is a potent amplifier of inflammation due to its induction of additional proinflammatory cytokines such as TNFα, IL-6 and IL-1β.58 60 Further underscoring the importance of the Th17/IL-23 signalling pathway in CD, recent genome-wide association studies and a meta-analysis have found that polymorphisms within several genes in the Th17 pathway are associated with CD.49 58 In IBD, anti-inflammatory cytokines such as IL-10 and transforming growth factor β (TGFβ) also play an important role in intestinal inflammation. For example, IL-10 knockout mice develop colitis through increased production of IFNγ.58 Thus, insufficient anti-inflammatory activity due to inadequate production or action of immunosuppressive cytokines like IL-10 and TGFβ is an additional crucial factor in triggering or maintaining intestinal inflammation.

Studies have sought to determine whether there is an association between increased cytokine production and disease relapse in patients who are in remission with medical treatment. Schreiber et al found that increased mucosal secretion of TNFα and IL-1β was predictive for relapse at 1 year in a cohort of 137 patients with CD in steroid-induced remission.61 A similar study by Arnott et al demonstrated that elevated levels of IL-1β (p<0.004) and IL-8 (p<0.02) in whole gut lavage of 54 clinically inactive patients with CD were associated with a greater chance of relapse.62 Thus, elevated levels of some cytokines may translate to subclinical mucosal inflammation and help identify those patients at increased risk for relapse. Although these studies did not include patients who underwent surgical resection for disease, they clearly demonstrated an association between sustained proinflammatory cytokine production and an increased risk of disease recurrence in subjects with CD in clinical remission.

Immune response and postoperative recurrence

In patients who experience recurrence of CD after undergoing an ileocolic resection, heightened production of inflammatory mediators in macroscopically normal intestinal tissue probably plays a key role in promoting recurrence of disease. In a study by Yamamoto et al, 36 patients who were in remission after surgical resection for CD were followed for 1 year and monitored for evidence of relapse (defined as a Crohn's disease activity index (CDAI) score of >150 or an increase in the CDAI from baseline by >40 points).63 Mucosal IL-1β, IL-6 and TNFα levels in the ileum at the time of remission postoperatively were significantly higher in the group that subsequently developed recurrence (median time to relapse was 9 months) than in the group who remained in remission at 1 year (IL-1β, p=0.02; IL-6, p=0.003; TNFα, p=0.03) on univariate analysis. However, on multivariate analysis, only elevated mucosal IL-6 levels were shown to be an independent risk factor for postoperative disease recurrence. This study also highlighted the absence of a significant correlation between plasma cytokine and mucosal cytokine levels, suggesting that mucosal inflammatory activity is a more sensitive and more robust predictor of CD recurrence than systemic inflammatory changes.63 Meresse et al measured IL-10, an anti-inflammatory cytokine, in ileal biopsies preoperatively to assess if this could serve as a marker for endoscopic recurrence in the neoterminal ileum at 3 months.64 These investigators found higher rates of endoscopic recurrence in patients with low IL-10 levels than in those with high IL-10 mRNA levels (80% vs 40%, p=0.02). Of note, recombinant human IL-10 has been studied in a clinical trial as possible postoperative prophylactic therapy but did not show evidence for prevention of endoscopic recurrence or in the degree of severity of endoscopic findings at 12 weeks, although the study was not adequately powered.65 In addition, an important question that arises from this negative study is whether recombinant IL-10 would instead have shown a benefit if patients with low levels of mucosal IL-10 at the time of surgery had been specifically targeted for this intervention. Other cytokines that have been evaluated in the postoperative setting relate to factors involved in wound healing and promotion of fibrosis, such as TGF-β1.66 67 Scarpa et al measured TGF-β1 in ileal mucosal samples of patients with ileocolonic CD at the time of surgery and followed these patients for 26 months postoperatively to assess whether there was a relationship between TFG-β1 and clinical recurrence.66 Increased TGF-β1 production in uninvolved bowel suggested a correlation with clinical recurrence of postoperative CD (r=0.43, p=0.04). Patients who expressed high TGF-β1 mRNA transcripts in uninvolved intestinal mucosa had a higher cumulative postoperative recurrence rate than those who expressed low TGF-β1 mRNA levels (20% vs 0%, p=0.02). These findings suggest a role for cytokines as markers of predictors for postoperative recurrence, as well as lending importance to their role in the pathogenesis of CD.

In addition, medical treatment has already suggested that manipulation of cytokines can successfully maintain remission of postoperative CD. Regueiro et al showed infliximab, an antibody to TNFα, to be effective in preventing endoscopic recurrence following surgery.68 In this randomised placebo controlled trial, 24 patients were assigned to 1 year of treatment with either placebo or infliximab (5 mg/kg) within 4 weeks of ileocolic resection. The authors found a significant benefit for infliximab therapy 1 year postoperatively, with lower endoscopic (9% in the infliximab group vs 85% in the placebo group, p=0.0006) and histological recurrence rates (27% in the infliximab group vs 85% in the placebo group, p=0.01). In addition, a higher proportion of patients in the infliximab group were found to be in clinical remission (CDAI score <150), but this secondary outcome was non-significant (p=0.38).68

The relationship between the immune response and postoperative recurrence of CD is shown in box 5.

Box 5

Immune response and postoperative recurrence

  • Baseline levels of certain immune parameters such as cytokines may predict an individual's risk of relapse after medically or surgically-induced remission.

  • Local mucosal inflammatory activity may be a more sensitive and earlier predictor of CD recurrence than systemic inflammatory activity.

  • Identifying reliable immune markers of postoperative recurrence would allow clinicians to better select more targeted medical prophylactic therapy from the growing armamentarium of available biological agents.

  • The use of anti-TNF antibodies could be specifically selected for patients in whom high TNF levels would be associated with an increased risk of recurrence or the use of recombinant human IL-10 could be re-evaluated in a select group of patients with low mucosal IL-10 levels at the time of surgery.


Epidemiological studies have long suggested a genetic basis for IBD based on higher rates of IBD among patients of Caucasian and Jewish ethnicity, familial aggregation of IBD and higher concordance rates among monozygotic twins compared with dizygotic twins.69 In 2001, three independent mutations within the nucleotide-binding oligomerisation domain containing 2/caspase recruitment domain family, member 15 (NOD2/CARD15) gene on chromosome 16 were found to be associated with CD.69 The product of NOD2/CARD15 is an intracellular receptor involved in the recognition of the bacterial cell wall component muramyl dipeptide.70 The mutations in NOD2/CARD15 lead to inappropriate NFκB signalling.10 NOD2/CARD15 mutations have also been found possibly to affect the function of Paneth cells and expression of defensins, natural antimicrobial peptides which represent another line of defence against bacteria.59

More recently, genome-wide association studies have further demonstrated strong evidence for an association between CD and more than 30 additional genes or loci, most notably autophagy-related 16-like 1 gene (ATG16L1), immunity-related GTPase family MTop of Form (IRGM) and interleukin 23 receptor (IL23R), all of which highlight specific pathways of disease pathogenesis involving the intestinal flora and the immune response.71 The ATG16L1 gene is involved in the intracellular clearance of bacteria, with studies demonstrating that the T300A variant of this gene results in defective microbial killing with increased exposure to commensal bacteria.72 The IRGM gene has also been implicated in CD susceptibility with its role in autophagy, most notably controlling elimination of Toxoplasma gondii, Listeria monocytogenes and Mycobacterium tuberculosis.73 The IL23R gene is involved in the immune response through activation of the Th17 pathway leading to activation of a number of proinflammatory cytokines.69 The identification of these genes has further underscored the role for a defective immune response and implicated defects in the intracellular handling of bacteria.

Previous genotype–phenotype correlation studies have demonstrated associations between NOD2/CARD15 and ATG16L1 variants and ileal involvement.74 Disease behaviour—namely, fibrostenotic or fistulising complications—have also been found to be associated with NOD2/CARD15 variants as well as a younger age at diagnosis in patients with CD.74 A recent study by Henckaerts et al demonstrated homozygosity in a CD gene variant, rs1363670 G allele in the AK097548 gene, not only to be independently associated with stricturing disease (OR 5.48, p=0.07) but also with a shorter time to stricture (p=0.01).75 Furthermore, some studies have suggested an increased need for surgery in patients with NOD2/CARD15 variants.76 77 This growing body of literature in CD genetics suggests that susceptibility-associated genetic markers could also serve to predict disease course.

Genetics and postoperative recurrence

To date, the role of genetic factors in postoperative recurrence of CD has not been clearly defined. There have been four studies examining the association between genetic variants and surgical recurrence in patients with CD. All four of these studies have focused on NOD2/CARD15 variants, with two suggesting an association between these variants and the need for reoperation while two did not find such an association. Buning et al investigated the prevalence of the three common NOD2/CARD15 variants (Arg 702Trp, Gly908Arg, 3020 insC) and their association with disease phenotypes in 180 patients with CD. The investigators confirmed previously described associations between these NOD2/CARD15 mutations and certain disease phenotypes: younger age at diagnosis (p=0.03), ileal disease location (p=0.01), ileocaecal resections (p=0.0002). However, they also found a higher risk of postoperative relapse and reoperation, with 12 of 29 patients (41.4%) with NOD2/CARD15 mutations requiring a second operation compared with only 2 of 12 patients (16.7%) without NOD2/CARD15 mutations.74 A later study by Alvarez-Lobos et al showed that patients with NOD2/CARD15 variants underwent surgery more frequently (OR 3.63; 95% CI 1.42 to 9.27) and required surgery at an earlier time (p=0.004).78 In addition, in this cohort of 170 patients who were followed for a mean of 7.4 years for surgical recurrence, it was observed that approximately one-third of patients required reoperation after initial surgery for stricturing disease and that NOD2/CARD15 variants were an independent predictive factor (OR 3.29; 95% CI 1.13 to 9.56). These patients with NOD2/CARD15 variants also experienced earlier postoperative surgical recurrence (p=0.03). The association between NOD2/CARD15 variants and stricturing disease is based on the hypothesis that these genetic variants shift the immune balance towards increased TGF-β production and increased collagen deposition by smooth muscle cells and fibroblasts, thus resulting in stricture formation and an earlier need for surgery.78 In contrast to these two studies, Ahmad et al found no association between NOD2/CARD15 variants and the need for a second operation in a cohort of 244 patients with CD,79 and a study by Maconi et al revealed no correlation between NOD2/CARD15 variants and the need for reoperation in a cohort of 253 patients with CD.80 These conflicting findings highlight the limitations in our current understanding of the role of genetic variants in the pathogenesis of IBD and its recurrence in the postoperative period.

The promise of genetic testing in the postoperative setting relates to the development of predictive models to help determine those at highest risk of disease recurrence. One advantage of genetic factors over other potential predictors is that they remain stable over time. Such predictive models are already being applied to predict the disease course in CD. For example, Henckaerts et al studied 50 single nucleotide polymorphisms (SNPs) associated with CD and developed predictive models of CD complications using a multifactor dimensionality reduction technique.75 In this study the authors identified a model that predicted a need for first surgery that involved two of the studied SNPs. It would be of great importance to determine if this same model or other models could be applied to the prediction of postoperative recurrence.

The relationship between genetics and postoperative recurrence of CD is shown in box 6.

Box 6

Genetics and postoperative recurrence

  • Genome-wide association studies have allowed the identification of genetic variations underlying CD and have highlighted certain components of the immune system and microbial recognition pathways in the pathogenesis of CD.

  • Identification of NOD2/CARD15 variants and multiple other gene variants may be clinically useful in identifying subgroups of patients with a severe disease course who are at high risk for postoperative recurrence of CD.

  • Given the large number of genes associated with CD and the relatively small effect of each individual gene, predictive models that identify combined genetic signatures hold great promise in stratifying the risk of postoperative recurrence.

  • The association of genetic with serological bacterial markers may better predict the risk of postoperative recurrence than either marker alone.


The prevailing theory regarding the pathogenesis of CD is that this disease is the result of a dysregulated immune response to intestinal microflora in the setting of a genetic predisposition and conditioning environmental factors.12 In patients who have undergone ileocolic resection for the purpose of surgically-induced remission, there is a unique opportunity to gain further insights into the underlying pathogenesis of CD. We suggest that, although the disease state is thought to be theoretically reset to its earliest phases following surgery, the disease phenotype and natural history of CD do not change significantly postoperatively thus leading to high rates of recurrence. Moreover, most factors predisposing to CD remain stable postoperatively including genetic make-up and altered immune response, which also supports the belief that the natural history of CD remains essentially unchanged after surgery. On the other hand, changes in environmental factors may certainly alter disease recurrence, considering studies demonstrating a 2–2.5-fold increase in clinical and surgical recurrence in smokers. While it may be difficult if not impossible to permanently change one's intestinal bacterial flora, the combination of modifying environmental factors such as diet, certain medications and postoperative changes such as loss of the ileocaecal valve and altered intestinal permeability may in turn affect recurrence of disease. Although it is difficult to analyse the gut microflora, studies documenting the efficacy of antibiotics in the prophylaxis of postoperative recurrence of CD suggest an important role for bacteria in disease recurrence. As discussed previously, although one's genetic make-up cannot be altered, the identification of CD genotype–phenotype associations may help to predict one's disease course early on and allow physicians to better tailor postoperative medical therapy sooner. Building on these findings, smoking cessation, antibiotic therapy and, more recently, prophylactic anti-TNFα therapy may all serve to decrease the risk of postoperative recurrence (see box 7).

Box 7


  • The pathogenesis of CD centres on the following key factors: dysregulated immune response to intestinal microflora, genetic predisposition and environmental conditions.

  • Given there is no evidence that the natural history of CD truly changes after surgery or over time, the pathogenesis of postoperative recurrence also centres on a better understanding of these factors.

  • While it is difficult to change these factors, studies with antibiotics and immunomodulators/biological agents in the prophylaxis of postoperative CD recurrence have suggested that the natural history may perhaps be changed after surgery.

  • More importantly, ongoing genotype–phenotype studies may predict one's disease course and allow physicians to better tailor treatment to prevent disease complications and surgery.



  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.