While not as common as colorectal cancer, the prevalence of Crohn's disease is around 300 per 100,000 persons in the West1. Crohn's disease is a chronic inflammatory condition of the gastrointestinal tract that relapses and remits despite medical therapy. Ileocolic disease is the most common location and occurs in approximately one-third of patients2,3. Treatment often involves medical therapy with steroids to induce remission, followed by immunomodulators with or without advanced agents such as infliximab to maintain remission. Surgery is typically reserved for disease refractory to medical therapy or for complications such as stricturing or fistulation. Unfortunately, Crohn's disease cannot currently be cured medically and nearly 80% of patients eventually require surgery4. Even with surgery, there is a substantial recurrence rate which may be influenced by many factors. Some of these factors such as genotype and phenotype of disease cannot be influenced by the clinician. The patient may be able to reduce recurrence by stopping smoking and the physician may influence postoperative recurrence with more aggressive postoperative medical surveillance and treatment. It is appealing to the surgeon to feel that they may also influence recurrence rates with their surgical management. Such management and surgical techniques are the focus of this article.

Before considering how the surgeon can influence postoperative recurrence, defining recurrence is necessary. However, this can be surprisingly challenging since recurrence may be clinical, endoscopic, surgical, or radiographic. Clinical recurrence, the primary outcome in many studies on postoperative recurrence, may simply be due to symptoms induced by the resection itself. Current scoring systems, such as the Crohn's Disease Activity Index, fail to differentiate symptoms related to altered anatomy from those caused by recurrent inflammation. This ambiguity probably explains the variation in clinical recurrence rates in the literature which range from 4-44% in one year5,6.

Endoscopic recurrence is in some ways easier to define. Rutgeerts et al. were the first to attempt a definition and noted two factors7; (a) endoscopic recurrence usually occurs at the ileocolic anastomosis and the ileum just proximal to the anastomosis, (b) endoscopic recurrence predates clinical recurrence by several months. Based on these observations, the Rutgeert's score was designed to enable prediction of clinical recurrence and the need for further surgery. A score of i2 or higher predicted surgical recurrence. However, this score has significant limitations including low inter-observer reproducibility8,9 which can lead to incorrect therapeutic decisions in up to 10% of patients. Additionally, some have suggested that lesions confined to the anastomosis are post-surgical and not predictive of subsequent clinical recurrence10. A “modified Rutgeerts' score” accommodates for this11 but the predictive value is still questioned12. Last, the score was designed based on observations of end-to-end or end-to-side anastomoses only. There has been a trend in the last two decades toward side-to-side anastomoses leading to configurations that are not easy to score. These factors together with the heterogeneity of disease characteristics and variable postoperative management of patients in subsequent studies make it difficult to be certain of the usefulness of endoscopic recurrence in guiding treatment. Nevertheless, the POCER trial13, which incorporates the Rutgeerts' score from a postoperative ileocolonoscopy, has influenced practice recommendations for managing patients after surgery4,13,14 (see later).

Perhaps to augment or reduce the frequency of endoscopic surveillance, there is an increasing interest in non-invasive techniques for detecting recurrence. These laboratory tests include fecal calprotectin which correlates with endoscopic recurrence, and C-Reactive Protein which does not. Some data exist regarding bowel ultrasound, magnetic resonance enterography and capsule endoscopy17, 18, 19. However, none of these techniques has been evaluated sufficiently to replace endoscopy as yet.

Surgical recurrence is defined as the need for repeat surgery at the site of anastomosis. While perhaps the most robust definition of recurrence, not all surgical re-interventions are due to recurrent disease. Missed strictures at the time of index surgery or technical errors may account up to 20% of early re-interventions15. Additionally, repeat surgery may be for non-Crohn's related problems such as adhesive obstruction or incisional hernias. Anastomotic strictures may be treated endoscopically, avoiding the need for surgery in some patients. In addition to these caveats, understanding the incidence of surgical recurrence requires prolonged observation. One large Scottish population-based cohort suggests a surgical recurrence rate of 8% at 1 year, 25% at 5 years and 29% at 10 years16.

Four main areas have been investigated regarding pathophysiology that are the basis of recent surgical innovations.

The microbiota and fecal stasis

One of the first studies to suggest an element of the fecal stream was responsible for recurrence was a very simple observational experiment in three patients with an anastomosis defunctioned proximally. Microscopic inflammation was triggered only after infusion of the anastomosis with fecal content20. Potential causatives in the fecal stream include the microbiota, bile acids and/or dietary components but none have been so far proven definitively. Modulation of the intestinal microbiome is currently being explored (NCT03943446). The fecal stream has also been implicated in recurrence, with relative stasis at the anastomosis and consequent bacterial overgrowth as potential contributory factors21.

Animal models and pathological specimens suggest recurrence after surgery is initiated at the mesenteric border of the surgical anastomosis22,23. This border is supplied by end arteries whereas the antimesenteric border has collateral supply. The collateral supply means that disease in the mesentery would tend to disrupt the blood supply to the mesenteric border of the bowel before the antimesenteric border and cause the pattern of ischaemic type changes (essentially ulceration) that are seen in post operative recurrence24. Further evidence for this theory comes from studies that have examined strictureplasty. This intervention does not resect bowel; instead, the antimesenteric border of the bowel is opened longitudinally and closed transversely, thereby increasing the luminal diameter. A systematic review of this procedure suggests a very low site-specific surgical recurrence rate despite leaving diseased bowel and mesentery behind25.

Creeping fat is pathognomonic of Crohn's disease and consists of hyperplastic mesenteric adipose tissue that wraps around the bowel wall, mirroring the distribution of inflammation present in the bowel. Recent evidence suggests that the mesentery may play a central role in the onset and maintenance of inflammation and could be considered the main trigger of the inflammatory response in Crohn's disease26,27. Indeed, the amount of creeping fat correlates with the severity of local inflammation in imaging studies and surgical specimens28.

Recent studies on T cell clonality, predominantly the CD8 T cell compartment, suggest corelation with endoscopic postoperative recurrence29. Such clonality may predominate or be absent in the mesentery and act as a trigger for reigniting disease after surgery.

Several risk factors for postoperative recurrence have been recognised and provide a basis for the surgeon to influence recurrence. These include clinical and histological factors.

Clinically, patients with penetrating disease, early age of onset and previous resection have been recognised in current guidelines as high-risk factors for recurrence30. However, they remain controversial. Several studies have suggested no correlation31, 32, 33 possibly associated with the fact that these patients often receive aggressive postoperative medical therapy. One factor that remains strongly correlated is smoking34. This possibly relates to the influence of smoking on T cell clonality35. Postoperative complications may also be an independent risk factor for endoscopic recurrence. One cohort of 237 patients suggested the risk of surgical recurrence within 1 year was 6.77 times higher if there had been a postoperative intraabdominal septic complication36.

From a histologic standpoint, a randomized controlled trial published by Fazio et al., compared limited versus extended resections of macroscopic Crohn's disease37. No differences were seen between groups, initiating the concept of minimal surgical resection for Crohns. It should be noted, however, that subsequent studies have suggested an association between histological inflammation at the resection margins and an increased risk of endoscopic postoperative recurrence38, 39, 40, 41.

The ileal digestive wall is innervated by a rich enteric nervous system composed of both submucosal and myenteric plexi. Inflammatory cells contiguous to or within an enteric nerve bundle define plexitis. These myenteric plexi has been observed in Crohn's patients for many years42. Whilst a recent meta-analysis found a positive association between plexitis and endoscopic postoperative recurrence43 the finding could simply reflect microscopic inflammation at the resection margin. Granulomas in mesenteric lymph nodes may be associated to endoscopic and surgical postoperative recurrence and can also be a marker of more severe disease44. Both the presence of granulomas and plexitis may guide postoperative therapy rather than influence surgical technique.

Our understanding of the pathophysiology and the clinical and histological risk factors associated with post-operative recurrence have stimulated changes in surgical practice. Novel techniques to potentially improve the efficacy of surgical intervention have been developed based on this understanding. However, the limitations also mentioned above, especially regarding the definition and detection of recurrence, mean many of these innovations remain unproven.

Preoperative intervention

The evidence that short term complications can influence long term recurrence means that every effort should be made optimise a patient before surgery. Whilst smoking does not per se appear to increase the risk of anastomotic leak it does influence recurrence. Former smokers appear to have a similar rate of recurrence as non-smokers, so the benefit is clear45.

Other factors that more directly increase the risk of complications include nutrition, sepsis and immunosuppressive medication, particularly steroids46.

Nutritional assessment and treatment before surgery is essential4,14. This may be in the form of enteral or parenteral nutrition if necessary. There is a recent interest in the use of exclusive enteral nutrition in the weeks before surgery, even in patients who are assessed as nutritionally competent. Studies are ongoing47.

Percutaneous drainage of any abscesses and treatment with antibiotics with correction of coexisting metabolic imbalances before surgical resection reduces the length of bowel requiring resection, the need for a stoma and the potential for complications48. The data on steroid use before surgery is conflicting probably related to the heterogeneity of the populations studied. However, current guidelines universally recommend reduction to low levels, preferably zero, before surgery is contemplated, particularly if an anastomosis is planned4,14.

The need for cessation of biological therapy is more controversial. The recently published PUCCINI trail suggests there is no association between the use of tumour necrosis factor inhibitors within 12 weeks of surgery and postoperative complications49. Nevertheless, it makes sense to reduce immunosuppressing medication as much as possible. The aim should be to operate in the window where sepsis is resolved, the patient is nutritionally competent and the medications are minimised but the patient does not have a flare of disease. Emergency Crohn's resection should be avoided as far as possible but the timing of elective resection in this small window is crucial to achieve good outcomes in the short term and potentially reduce recurrence in the long term.

Early Bowel Resection

Whilst most clinicians and patients prefer a treatment algorithm consisting of escalation of medical therapy before surgery, there is a school of thought that considers surgery should not be the ‘last resort’. Instead bowel resection earlier in the treatment algorithm may actually reduce the risk of recurrence50. A seminal study in this respect was the LIRIC trial which clearly showed that a laparoscopic ileocolonic resection is a reasonable alternative to infliximab therapy as a first approach in patients with inflammatory ileocolic Crohns disease51. In the long term the need for surgical intervention/reintervention and escalation of medical therapy is reduced. Although many Crohn's patients persist with pharmacotherapy as long as possible, when questioned after they had failed medical treatment and undergone surgery, over three quarters wish they had had surgery earlier52.

The mechanism by which earlier surgery reduces recurrence is unclear. It may be that removal of the mesenteric creeping fat (which may contain pro-inflammatory mediators) together with removal of the affected bowel that contains the compromised mucosal barrier, reduces the potential for chronic inflammation and cumulative bowel damage53, 54, 55. A sort of ‘top-down’ surgical therapy analogous to the ‘top down’ medical approach for treatment that is the current vogue56. Current guidelines suggest any patient with Crohn's Disease who is being considered for escalation of medical therapy should have surgery discussed as a treatment option and the relative merits and disadvantages of each intervention explained, preferably by both a surgeon and gastroenterologist together4,14,57.

Minimal bowel resection

The current trend of bowel preservation remains surgical doctrine for Crohn's surgery, allowing preservation of bowel and reduced chance of short bowel syndrome37. The potential need for further surgery with Crohn's disease means this trend is likely to continue, particularly as extended bowel resection confers little advantage. The trials that suggest increased recurrence if there is microscopic resection margin positivity or presence of plexitis are unlikely to change this vogue41, 42, 43, 44 and the potential for frozen section analysis in deciding resection margins is also unlikely to become mainstream practice. However, Coffey highlights the need to identify the proximal mesenteric and mucosal transition zone, where the macroscopic appearances of thickened bowel and mesentery change to a more normal appearance, as the marker for placement of the resection. If carefully placed, the proximal margin is clear microscopically in most cases58.

Extended Mesenteric Excision

The current concept of bowel preservation with Crohn's surgery often extends to the mesentery. Traditionally, there has been no requirement to radically resect the mesentery as one would do with cancer surgery. While some advocate complete preservation of the mesentery to protect the neurovascular supply to the remaining ends of the bowel (see Kono-S) most surgeons currently resect along ‘the plane of least resistance’ where the thickened bowel becomes thinner and more pliable. Here resection is easier and bleeding sequelae less likely to occur.

Recently operative strategies aimed at a more radical ‘cancer like’ resection have been advocated to reduce recurrence. Supporters suggest the mesentery is the focus of the disease, with high visceral fat content, increased lymphatic vascular density at the resection margin and granulomata in the mesenteric lymph nodes associated with recurrence59, 60, 61. Failure to remove this sump of proinflammatory mediators leads to reignition of bowel inflammation. Two comparative trials on extended mesenteric resection have suggested a large reduction in surgical recurrence compared with historical controls. In one trial the reduction was spectacular with a 5-year recurrence rate of 2.9% compared with 40% with standard excision58. The other trial showed a threefold reduction from 30% to 10.6%. However, both trials are methodologically poor and prone to significant bias62.

One criticism of the original description of this radical technique is the need for ligation of the ileocolic vessel close to its origin resulting in devascularisation of a larger portion of the colon and extended bowel resection. A recent modification of the technique which preserves the ileocolic trunk has been advocated63. Given the poor quality of evidence for extended mesenteric resection currently, further research is required. Some large studies are ongoing (NCT04538638, NCT04539665 and64).

Anastomotic configuration

Many surgeons feel that the anastomotic configuration and method of construction influences the rate of post operative recurrence. Some profess the value of a hand sewn technique, others the need for a side to side or side to end or end to side construction. Nevertheless, the evidence is conflicting despite being investigated extensively. Several studies have focused on this topic including four RCTs65, 66, 67, 68. Most studies have compared the configuration, typically end-to-end versus side-to-side. Whilst end to end anastomoses tend to be handsewn and side to side stapled, one RCT focused specifically on the hand-sewn versus stapled component66. The results favoured a stapled technique, but recurrence was more likely with a stapled technique if only an ileocolic anastomosis was considered (albeit with the caveats of post-hoc subgroup analysis). Also, when it comes to comparing stapled versus handsewn anastomoses the risk of over-estimating endoscopic recurrence in stapled anastomoses must be accounted for. Some ulcerations may be associated with ischemia or a reaction against the stapler rather than an aphthous ulcer and a recurrence of Crohn's11.

The plethora of studies investigating anastomoses in Crohn's resection have allowed several meta-analyses of the data, which appear contradictory in their conclusions. In the most recent, Feng et al simultaneously compared several types of anastomoses69. They concluded that the side-to-side anastomosis was more likely to decrease the risk of clinical and surgical post operative recurrence. However, the noted the heterogeneity in definitions of recurrence and the length of follow-up are caveats for drawing firm conclusions and despite many current guidelines recommending a wide lumen stapled side to side anastomosis4,14, the evidence is still unclear.

One observational study on anastomotic configuration included an animal model of an antiperistaltic side to side anastomosis and concluded that that perpendicular transection of the intestinal circular muscle layers and intestinal reconstruction in an anti-peristaltic orientation disrupts bowel motility, with consequent stasis of the enteric content70. Fecal stasis has long been implicated in anastomotic recurrence21 leading to the belief that, whatever anastomosis is constructed, there should be a wide anastomotic lumen.

The Kono-S anastomosis

A great deal of interest has been accumulating about the Kono-S anastomosis. This configuration considers much of our current understanding of the pathophysiology of recurrence. A concept of mesenteric exclusion with a completely antimesenteric anastomosis is compatible with the theory of the mesentery as the driver of recurrence. The observation made in animal models that recurrence starts at mesenteric border of the bowel also lends credence to the Kono-S principle22,23. Any recurrence will not present clinically until markedly advanced. The wide-open lumen without a blind stump minimizes the risk of fecal stasis and bacterial overgrowth. A supporting column may prevent distortion from fibrosis and delay clinical symptoms of recurrence if it should occur.

The literature on Kono-S is summarised in two systematic reviews71,72. There is only one RCT73, which found a significant reduction in endoscopic recurrence at 6 months and less clinical recurrence at 24 months with the Kono-S anastomosis compared with a standard stapled anastomosis (18% versus 30%, p=0.04, OR 3.47). The previously mentioned issues of overscoring recurrence due to ulceration related to a staple line and inherent bias in the assessment (endoscopists were not blinded to the intervention) are potential issues with this trial and may explain the high rate of endoscopic recurrence seen in the comparator group (>60% at 6 months). Further quality research is required to confirm these findings and several studies are ongoing ([74] and NCT03256240).

In the original description of the technique by Kono, close bowel dissection and mesenteric preservation were postulated as critical elements for limiting recurrence by preserving vascular and nervous supply to the remaining bowel75. This contrasts with the advocates of extended mesenteric excision. Despite the different techniques, there may be commonalities that explain why both are effective. Both aim to isolate the anastomosis from diseased mesentery. Kono-S does this by a totally antimesenteric anastomosis as far away as possible from the mesentery; radical mesenteric resection does this by removing the theoretical disease driver (the mesentery). A combination of techniques is possible and may increase efficacy. This combined approach has been described by Holubar et al. who showed that the approach was feasible and safe76. The effect of mesenteric excision and exclusion (Kono-S anastomosis) on the prevention of postoperative recurrence, will be further investigated in the MEErKAT trial74. A brief comparison of the anastomotic techniques is in Figure 1.

There are 2 considerations when it comes to postoperative care that are specific to recurrence. The consideration of the microbiome as a causal agent with postoperative antibiotics as a potential modifier and the need for careful endoscopic follow up.

Metronidazole

The microbiome may be a cause of post-operative recurrence and modification of this microbiome has been postulated as beneficial for years. Indeed, in a Cochrane meta-analysis, nitroimidazole antibiotics were associated with a reduced risk of clinical and endoscopic recurrence relative to placebo77. These agents are associated with higher risk of serious adverse events and poor tolerance and compliance. A lower dose for a 3 month postoperative period may reduce these side effects whilst maintaining efficacy78 and could be considered.

Endoscopic follow up

The POCER trial compared a group of patients undergoing step up therapy based on endoscopic assessment 6 months after ileocolic resection with those having standard follow up and clinically guided postoperative therapy13. Using the Rutgeert's score those with lesions of I2 or more after endoscopy had this step-up therapy. Around half of the intervention group had endoscopic postoperative recurrence at 18 months versus 67% who did not undergo this active management (p=0.03). This confirms the benefit of an active management strategy based on an early post-operative ileocolonoscopy and international guidelines now recommend an ileocolonoscopy within one-year post-surgery4,14.

As with preoperative decision making and patient management, close communication with the gastroenterologist is essential to ensure the optimal care for the patient in the post operative setting.