Introduction

In recent years, several medical societies have published authoritative consensus guidelines concerning the management of inflammatory bowel disease (IBD). These papers agree on the majority of topics, but several discrepancies can be noticed between guidelines from different societies. The present consensus on Crohn’s disease (CD) and another on ulcerative colitis (UC) [1] represent a think tank of Italian colorectal surgeons affiliated with the Italian society of colorectal surgery (SICCR) (http://www.siccr.org/index.php?lang=ENG&pagid=home). These are not intended to replace or implement available guidelines. The aim of this project, promoted by SICCR, was to identify the agreement of surgeons affiliated with SICCR who treat patients with IBD with the evidence in the literature, in order to identify areas where there is room for improvement and points that could be clarified by collaborative studies.

The project was articulated in three steps with iterative revision of the output (Delphi method):

  1. 1.

    Review and identification of relevant questions Literature review and identification of critical points from a surgeon’s perspective, drafting of “items” and supporting text, and circulation among members of the project committee. Experts joining the consensus defined relevant questions concerning surgery for IBD based on the literature data and personal experience. The junior consensus participants had the task of systematically reviewing the literature (assigning priority to papers published within the last 15 years) for the defined questions, and to write the first draft of the text. The level of evidence (EL) was graded according to the “The Oxford Levels of Evidence 2” of the Oxford Centre for Evidence-Based Medicine (OCEBM) (http://www.cebm.net/index.aspx?o=5653). The panel Chairs reviewed the text according to the literature data and wrote provisional items that were circulated among the committee along with the supporting text, which was modified according to members’ comments. The text with items was then submitted to the SICCR Governing Board. Items were kept if they achieved enough priority and agreement; otherwise, they were removed or changed.

  2. 2.

    Broad circulation of “items” After revisions and approval by the Governing Board, items with EL were circulated among SICCR members dealing with IBD through two Web-based surveys (CD and UC). Members were asked to rate items anonymously (“Agree”, “Partially agree”, “Do not agree”, “Not able to respond”), and to make any comments they considered important.

  3. 3.

    Collection of results and manuscript finalization The results of the surveys were collected and analysed. Agreement is reported after each item. Items that achieved less than 70 % of agreement are discussed in detail in a dedicated section. The text was revised according to comments, if any, and the section including the output of the surveys was added to the manuscripts. The revised text was approved by members of the committee and was submitted to the SICCR Governing Board.

Section I: Items and supporting text

Abdominal CD

Classification

The Montreal revision (Table 1) [2] added isolated upper gastrointestinal (GI) location (L4) and introduced “perianal disease” (“p”), which can overlap with CD patterns defined by the Vienna classification (“non-stricturing non-penetrating”, “stricturing”, “penetrating”). The use of a classification system is recommended at diagnosis as a shift towards more aggressive patterns can be observed over time, and several CD-related factors may predict disease evolution [3, 4].

Table 1 Vienna classification and Montreal revision for Crohn’s disease [2]
Full size table

A review of population-based studies reported that several factors are commonly associated with an aggressive disease course, with more frequent clinical recurrences needing treatment [5]. Specifically, early disease onset, CD localized in the terminal ileum, and penetrating/stricturing disease were associated with clinical recurrence. An US population-based study suggested that ileocolonic (HR, 3.3), small bowel (HR, 3.4), and upper GI (HR, 4.0) extent, and penetrating disease behaviour (HR, 2.7) were strongly associated with the need for major surgery, whereas a slightly higher risk was observed in active smokers (HR, 1.7), male patients (HR, 1.6), and patients who began corticosteroid treatment early (HR, 1.6) [6]. The European Crohn’s and Colitis Organization (ECCO) recommended proactive medical management of patients with two or more predictors of an aggressive disease course [7].

Preoperative assessment

A complete serological assessment is mandatory in CD patients scheduled for surgery. The use of biomarkers is nowadays accepted as more reliable than clinical assessment alone, and the utility of tools such as the CD activity index (CDAI) alone is debated and should be questioned in assessing response to treatment and when planning clinical trials [8]. C-reactive protein (CRP) is a valid tool, which can be used in assessing disease as well as in guiding treatment [9] and post-operative management.

Most CD patients need surgery for disease involving the terminal ileum, which is easily reached by ileocolonoscopy. Small bowel enteroscopy is not routinely recommended, and capsule endoscopy is associated with the risk of obstruction due to potential impaction of the capsule in a stricture. Endoscopy allows for adequate assessment of the mucosa, but it should be implemented with extramural assessment by means of cross-sectional imaging. CT and magnetic resonance (MR) enterography and enteroclysis are equally accurate in assessing disease severity and in detecting extraluminal complications [10]. The latter achieves better bowel distension. MR enterography is particularly useful in planning surgical treatment [11] and should be considered the ideal tool for young patients [12]. However, it is expensive and time-consuming, and an experienced operator is required.

Transabdominal contrast-enhanced US has recently been suggested as a reliable, non-invasive tool for assessing stricturing CD [13, 14], but a high level of expertise is required and the usefulness of the examination can be limited by patient- and disease-related factors.

Perioperative considerations

Observational studies support a role of steroids in increasing the rate of post-operative complications and agree in indentifying a cut-off of 20 mg for more than 6 weeks [15, 16]. Data on 8260 CD patients obtained from the American College of Surgeons National Surgical Quality Improvement Program® (ACS-NSQIP®) (http://site.acsnsqip.org/) confirmed that intra-abdominal septic complications (15.2 vs. 12.9 %, p = 0.004) and venous thromboembolism [odds ratio (OR) 1.66; 95 % confidence interval (CI) 1.17–2.35] after an abdominal surgery [17].

A meta-analysis of comparative studies supported the safety profile of thiopurines in CD patients undergoing surgery [18]. Conversely, septic complications were increased by steroids and anti-TNF drugs [18]. Concerning biologics, data on overall post-operative complications have been conflicting. However, when assessing only septic complications and removing confounders, recent systematic reviews with meta-analysis found increased likelihood of complications with biologics [19, 20].

A close interaction between IBD team members and the patient is important to optimize surgical timing.

Indications for surgery

  • Ineffective medical treatment

Symptomatic patients with mild to moderate disease should be approached with first-line medical treatment (i.e. aminosalicylates, antibiotics, and steroids), and failure to obtain a response defines severe CD [21]. Intravenous (IV) steroids, immunosuppressant drugs, and biologics should be considered. Surgical treatment is warranted in case of failure of first- and second-line therapy, as well as in patients with severe or steroid-dependent CD of limited extent and in those with contraindications or adverse effects to medical treatment [22].

Surgery should be considered early in patients with CD colitis with signs or symptoms suggestive of free perforation or failure to improve after 48–96 h since rescue medical treatment was initiated [23, 24].

  • Complications

  1. a.

    Intra-abdominal abscess and perforation

There is no consensus as to what constitutes the ideal management of intra-abdominal abscess. A recent review suggested that percutaneous drainage with antibiotics, assisted by imaging tools, should be the first-line approach to these patients [25], reserving surgery for unsuitable patients or if prior approach fails [23, 24]. An option is to use a two-stage approach, meaning that percutaneous drainage is suited to drain acute sepsis, delaying surgery until the clinical condition of the patient has been optimized [25]. Immediate resection offers less chance of subsequent complications compared with simple suturing. After resection of a perforated segment of small bowel, an end stoma, diverted and non-diverted anastomosis are all viable options and are based on intra-operative findings and the general status of the patient. A non-diverted anastomosis can be performed [23, 24]. Similarly, for colonic perforation, a proximal diversion and mucous fistula, a Hartmann’s procedure, and diverted anastomosis are adequate options [23, 24, 26].

Asymptomatic entero-enteric fistulas do not require surgery, but surgical intervention is mandatory when signs of abdominal or systemic sepsis occur [25, 27]. Resection is the recommended approach.

  1. b.

    Obstruction

Fibrotic strictures are a common complication in CD patients. Fibrosis results from complex, immune-mediated interactions between the patient and environmental factors, and no medical treatment is able to reverse bowel fibrosis [28, 29]. When strictures are within the reach of an endoscope, and if the procedure is performed in centres with experienced surgeons, endoscopic dilation can be an option for short strictures (<4 cm) [30]. However, there is a risk of subsequent surgery even if the procedure is performed in specialized centres, with more than 30 % patients needing surgery within 3 years [31].

  1. c.

    Haemorrhage

Haemorrhage is an exceptional complication of CD and can result from a penetrating ulcer eroding into the vessels of the submucosa. Treatment depends on patient status. When the patient is stable, endoscopy is recommended, as it may identify the source of bleeding and endoscopic hemostasis can be implemented [32, 33]. Angiography with embolization is an alternative in stable patients. When these measures fail, or if the patients are hemodynamically unstable, surgical exploration must be performed, with bowel resection and/or intra-operative endoscopy [32, 33].

  1. d.

    Growth retardation

CD patients in childhood or developmental age should be monitored from adequate height and weight gain with growth charts. Failure to thrive with signs of skeletal alterations and delayed puberty despite adequate medical treatment could justify surgery in prepubertal CD patients [34].

  • Neoplasia

The 3rd Scientific Workshop of the ECCO focused on the risk of malignancy in IBD and concluded that CD patients are at increased risk of adenocarcinoma arising in bowel segments with long-standing, severely active disease via an inflammation-driven pathway [35, 36]. Some cases of carcinomas arising at the site of stricturoplasty have been reported; hence, it is recommended to perform biopsies in segments suited for bowel sparing techniques in patients with long-standing disease [36].

The risk of malignant transformation of large bowel strictures is reported to be as high as 3.6 and 4.9 % at 5-year and 10-year follow-up in a European population-based study [37]. The risk of developing colorectal cancer in patients with stricturing phenotype is 5.5 % after 5 years and 7.5 % after 10 years, compared with 0.4 % of other patterns [37]. Hence, strictures of the large bowel should be carefully surveyed or patients should be offered resection.

There is still debate about how to assess dysplasia in IBD patients [38]. Dysplasia predicts concomitant or future development of aggressive carcinoma, and it is prudent to recommend resection in case of DALM, high-grade dysplasia, and multifocal low-grade dysplasia located in the large bowel, after biopsies with pathological confirmation [39, 40].

In CD patients, colorectal cancer should be suspected early in order to avoid diagnostic delay that can lead to poorer outcomes [4, 35, 41]. A study on more than 2843 IBD patients (1201 CD) from a tertiary US centre observed over 30 years showed that CD and UC patients undergoing surgery for colorectal cancer had similar short-term morbidity, but CD patients had significantly worse 5-year survival (41 vs. 29 %, p = 0.04) [41]. These data suggest that CD patients more frequently have advanced disease when they are diagnosed.

Perianal CD

Classification

Parks et al. [42] defined fistulas according to the relationship of the track with the external sphincter, whereas Bell et al. [43] later simplified Parks’ anatomical classification into “simple” or “complex”. Bell’s classification system is easier to apply and may guide surgical treatment. However, it should be used with caution when planning surgery. In particular, anal-vaginal fistulas represent a “grey area”. Ano- and recto-vaginal fistulas are all regarded as “complex” according to Bell’s classification, but low, anal-introital fistulas should be considered “simple” and treated as such (see Item 17).

Along with evaluation of disease activity (e.g. by means of perianal disease activity index (PDAI), Table 2) [44], one of the various anatomical classifications should be used to plan surgery and to assess disease evolution.

Table 2 Perianal Crohn’s disease activity index, proposed by Irvine et al. [44]
Full size table

Preoperative assessment

In experienced hands, accuracy is high, which allows concomitant drainage of abscesses. In patients with perianal or perineal pain, an abscess is almost always found. When an abscess is suspected, upfront EUA should be performed to control sepsis.

Pelvic-perineal MRI is the method of choice to assess perianal CD and can be used to guide treatment [45]. Endo-anal US, especially with three-dimensional (3D) reconstruction, and MRI have similar sensitivities at detecting perianal fistulas, but the latter has higher specificity [46, 47]. The ideal management is a combination of EUA, US, and MRI.

Since endoscopic findings (e.g. rectal disease) can alter managements, endoscopy is recommended routinely in perianal CD [46, 48].

Treatment

  • Perianal abscess

It is important that patients in whom the suspicion of perianal abscess is high are offered timely EUA with abscess drainage, in order to avoid spread of the local sepsis with predictably unfavourable outcomes of subsequent treatment attempts. Perianal abscess must not be overlooked.

  • Simple fistula

If the patient has no discomfort directly related to the fistula, and if no abscesses are found, the fistula can be simply managed with watchful waiting [49]. Antibiotics with fistulotomy or loose-seton placement are effective options for simple fistulas with accompanying symptoms [50]. The same principles apply to low anal-introital fistulas.

  • Complex fistula

It is important to rule out and treat concomitant rectal involvement [48]. Antibiotics are administrated routinely in clinical practice. Healing rates of complex perianal fistulas could be improved by adding an immunosuppressant to loose-seton placement [50]. Recent evidence suggests that combined treatment with surgery followed by infusions of biologics is associated with a higher healing rate, shorter time until healing, and longer disease-free intervals than treatment with biologics or surgery alone [49, 51]. The use of adipose-derived stem cells injected along the tract has been recently investigated in complex perianal CD, but definitive data from trials are not yet available [52].

Patients with severe active perianal CD who fail conventional or rescue treatments [53] may be candidates to diverting ileostomy, but the rate of conversion is very low and decreases over time [54]. Proctectomy is still needed in 10–18 % of patients, due to disabling symptoms and concerns about cancer [52].

Site-specific considerations

Localized ileocaecal disease

Most studies agree that primary, stricturing CD localized exclusively at the terminal ileum and associated with only minimal inflammation is best treated by early surgery, to avoid exposing the patient to medications [55, 56]. Minimally invasive surgery is recommended in centres with expertise, allowing early recovery and reducing adhesions [57–59].

Localized colonic disease

Although segmental resections increase the risk of post-operative clinical recurrence, the benefits of avoiding proctocolectomy overwhelm the risk of relapses, and this is the most widely accepted procedure for colonic CD with limited involvement [60–62].

Multisegment colonic disease

When both ends of the colon are involved, segmental resections could still be an option, but a meta-analysis pointed out that colectomy with ileorectal anastomosis ensures longer disease-free intervals, with 4.4 more years to relapse than segmental colectomy [62].

Additional considerations

Dilation of strictures

A study from the Cleveland Clinic suggested that with endoscopic dilation both primary and recurrent strictures have similar outcomes [31]. Indications and drawbacks of endoscopic dilation have been discussed in the supporting text of Item 9 [30]. Patients should be aware that repeated procedures may be needed, and that the risk of surgery after dilation of anastomotic strictures is high, so that some patients could benefit from early surgery [63]. Due to the high risk of cancer in strictures of the large bowel, and the potential risk of perforation with disease dissemination, the panel recommend resection of colonic strictures.

Strictureplasty

Bowel sparing techniques are associated with higher recurrence rates than resection, but are extremely useful in patients in whom massive resection and several, repeated surgical procedures have been performed within a short time, due to the risk of short bowel syndrome [30]. Patients’ nutritional status should be good; otherwise, septic complications are unavoidable after stricturoplasty, due to impaired healing.

Conventional techniques can be used for strictures shorter than 20 cm. These mainly consist of the Heineke–Mikulicz technique (suitable for 5- to 10-cm strictures) and the Finney technique (suitable for 10- to 25-cm strictures). These are easier to perform and are associated with a lower risk of recurrences [30, 64]. Unconventional strictureplasties can be suitable for longer strictures or for many short strictures in close proximity. These techniques consist of variants of the Heineke–Mikulicz and Finney procedures [65, 66] as well as more advanced procedures [67] which should be used only in special situations due to their complexity and the higher risk of recurrence [30, 68]. There is debate about the role of ileocaecal strictureplasty, although some centres report good results [69]. Colonic strictureplasties are not recommended due to concerns about cancer.

Anastomotic technique

Evidences suggest that no difference can be expected between hand-sewn and stapled ileocolic anastomosis, on condition that an adequate width of the lumen is achieved [70–72]. Stapled side-to-side anastomosis with a wide lumen is easier to perform, is less time-consuming, and results in functional end-to-end. It is the most effective choice [72].

Laparoscopy

Minimally invasive surgery is recommended in abdominal CD when performed by experienced teams. Short-term benefits consist of early bowel function restoration, whereas better cosmesis and body image are observed in the long term. The risk of incisional hernia may be reduced [73, 74]. The best results are obtained in patients with non-penetrating, small bowel disease. However, laparoscopy involves longer operative times and higher costs.

Complex or recurrent CD can be treated laparoscopically by surgeons with adequate skills [73, 74].

Post-operative treatment and prophylaxis

Smoking is the strongest predictor of recurrence after surgical treatment of CD, and smokers have a 2.5-fold increase in surgical recurrence compared with non-smokers [75, 76]. Previous surgery, perianal CD, massive small bowel resections, as well as avoidance of early post-operative medical treatment are associated with an increased risk of recurrence [75, 77].

A meta-analysis compared the effectiveness of thiopurines versus 5-aminosalicylic acid (5-ASA) derivatives and concluded that the former are more effective in preventing relapses but are more frequently associated with adverse events [78]. Since results are even more apparent at 1-year follow-up, prolonged prophylaxis is recommended [79]. Tursi et al. [80] recently tested adalimumab and infliximab as prophylactic treatment after resection of CD, but ideal candidates for such an approach need to be identified in larger trials.

Section II: Considerations on items with poor agreement

Overall, the agreement was high. Disagreement (“Do not agree”) did not exceed 25 % in all items. Several items require further comment.

Interestingly, items based on everyday practice and good sense (“expert opinion”) rather than strict evidence-based criteria achieved high agreement.

Great variability among respondents was observed regarding CD of the colon (items 19 and 20), with <50 % agreement, and 20–25 % of item rejection (“Do not agree”). Decisions about resection or continued medical treatment should be patient-tailored and take into account disease duration, extent, and activity. Colonic disease raises concerns about cancer, and the role of strictureplasty has been questioned. At the same time, these strictureplasties are easy to follow-up by endoscopy. Robust data are lacking on colonic CD.

The effect of perioperative medical treatment on post-operative complications was another matter of debate (items 4 and 27). The adverse effects of prolonged corticosteroid treatment are well known, and the safety of thiopurines is supported by consistent data. Though data on biologics are still conflicting, concerns about infection are justified when evaluating specifically post-operative abdominal complications. Post-operative prophylaxis is warranted in high-risk patients. There were also discrepancies in the combined management of complex perianal disease (item 17). It is likely that some surgeons would give patients biologics as first-line treatment to avoid immunosuppressant drugs.

Serological markers are accurate and probably not used enough in clinical practice (item 2). At the same time, there is a need of further, reliable, non-invasive markers to be used as an adjunct to clinical parameters.

Classification of perianal disease by means of anatomical and functional scales is useful to monitor disease evolution and treatment results, but was not accepted by all respondents (item 12). A validated, easy-to-use score taking both dimensions into account is desirable. Some respondents questioned considering ano-introital fistulas as simple, leaving asymptomatic simple fistulas untreated, or adding antibiotics to surgery in patients with simple fistulas (item 16). Decisions are based on patient preference, but such conduct is justifiable.

Some participants did not agree completely that surgery should be performed in children (item 10), but if available treatments are delivered correctly and no benefits are observed, the harm of delaying surgery should be considered, when operative management can remove inflammation.

Strong data support the use of laparoscopy in CD, but long-term benefits should not be overestimated and surgeons must have advanced skills, explaining suboptimal agreement (item 24). Dilation of small bowel strictures is being popularized, but access to it is limited in some hospitals (item 21). There are associated complications, suggesting that candidates for the procedure should be referred to high-volume centres or treated with alternative approaches. Respondents felt that further data concerning the ideal type of anastomosis after ileocolic resection would be useful (item 23).

Colonic and perianal CD are the areas that should be further investigated with collaborative studies, and the effects of medical treatment on surgery should be reported in detail in trials and observational studies. Technical aspects and newer techniques also need to be implemented, representing another field to be further explored.