Introduction

Crohn’s disease (CD) is a chronic inflammatory condition of the gastrointestinal tract and terminal ileum and caecum are the most commonly affected areas. Surgery can become necessary in up to 80% of the patients, mainly for lack of response to medical management or complications such as strictures, abscesses or fistulae [1]. A limited ileocaecal resection is the most frequently performed procedure for ileocaecal CD and different anastomotic configurations and techniques have been described for restoration of the intestinal continuity, with the aim to minimise postoperative complications and clinical and surgical recurrence. The side to side anastomotic configuration has been reported to have a favourable profile in terms of postoperative CD recurrence [2] and a wide lumen stapled technique has been recommended by surgical societies [3]. Ileocaecal resection for CD has a high reported rate of postoperative complications increased by risk factors such as malnutrition, active inflammation or infection at the time of surgery [4] and immune suppression, with reported rates of intra-abdominal sepsis and anastomotic leak as high as 14–17%, respectively [5]. Moreover, repeated surgery for anastomotic related complications represent one of the main reasons for short bowel syndrome in CD [6] rather than multiple resections over time for surgical recurrence [7].

The Italian Society of Colorectal Surgery (SICCR) recently reported the results of a national multicentre study collecting benchmark data on surgical treatment of CD, highlighting significant variations in practice [8]. This study audited the different anastomotic techniques used and evaluated their influence on postoperative outcomes following ileocaecal resection for primary CD.

Methods

Study settings

The SICCR promoted the snapshot study “Current Status of Crohn’s Disease Surgery”, which is a retrospective, multicentre, observational study. A steering committee developed the study protocol following the STROBE checklist [9] and this was reviewed independently by the research board of the SICCR. Ethical approval was obtained from the promoting centres and every participating hospital had a named Principal Investigator, liaising with the local ethics committee. Obtaining informed consent from the patients was deemed not necessary by the Ethics Committees in view of the retrospective nature of the study. Participating centres were invited directly and by an open call published on the SICCR website and also disseminated during a 2 months period via the society newsletter.

Eligibility criteria

All patients (aged 16 or older) undergoing elective ileocaecal resection for primary CD from 1st June 2018 to 31st May 2019 were eligible for participation to the study. Patients undergoing proctocolectomy, proctectomy or segmental colectomy were excluded as were patients undergoing urgent or emergency surgery. Indication for surgery included limited terminal ileal disease, CD refractory to medical treatment, obstruction, internal fistulae and abscesses. All patients in whom an ileocolic anastomosis was not fashioned or it was protected by a diverting stoma were also excluded.

Study objectives

Postoperative morbidity within 30 days of surgery was the primary endpoint. Postoperative length of hospital stay (LOS) and anastomotic leak rate were the secondary outcomes.

Data collection

Collected data included: patients’ demographics, Montreal classification, preoperative medical treatment and indication for surgery, American Society of Anaesthesiologists (ASA) grade, operative details, surgical access and conversion rate, length of hospital stay, 30-day postoperative morbidity, readmissions and reoperations.

The orientation of the two bowel ends for the ileocolic anastomosis was recorded as anastomotic configuration (end to end, side to side iso-peristaltic, side to side anti-peristaltic, end to side), whilst the suturing technique was either hand-sewn or stapled.

Postoperative morbidity was defined as any complication occurring during the hospital stay or within 30 days after surgery, whilst all readmissions were recorded up to 30 days after discharge.

Statistical analysis

Categorical variables are presented as frequency and percentages, and were compared using the Chi-square test or Fisher’s exact test, as appropriate. Continuous variables are presented as mean (± standard deviation) or median (range) according to their distribution, and were compared with the use of Student’s t test or the Mann–Whitney U test in case of normal or skewed distribution, respectively. To identify variables associated with binary outcomes, uni- and multivariable logistic regression analyses were performed. Variables having a p value equal to 0.10 or less at the univariate analysis were included in the multiple regression model. The Odds ratio (ORs) with a 95% confidence interval (CI) was estimated as measure of association. All reported p values were two-tailed, and p values of less than 0.05 were considered to be statistically significant. Statistical analysis was performed using IBM SPSS Statistics for Windows, version 25.0 (IBM Corp., Armonk, NY, USA).

Results

427 patients were included following exclusion of 30 patients in whom no anastomosis was performed or it was protected by a diverting stoma. Patients’ details are reported in Table 1. There were no statistically significant differences in the preoperative BMI, phenotype of disease, perianal disease, and preoperative medical treatment amongst the groups. The side to side anastomosis was the chosen configuration in 380 patients (89%). The stapled anastomotic (n = 286; 67%), techniques were preferred to hand-sewn (n = 141; 33%).

Table 1 Anastomotic configurations in the 427 included patients
Full size table

Overall morbidity and anastomotic leak rate

Eighty seven patients had postoperative complications (20.3%), with 36 complications being Dindo-Clavien class ≥ 3 (8.4%) and 16 anastomotic leaks (3.7%).

The incidence of postoperative complications in the patients having an anti-peristaltic side to side anastomosis was 23.4% compared to 16% for the iso-peristaltic side to side anastomosis (Table 2.), but this was not statistically significant. At multivariate analyses (Table 3) postoperative morbidity was associated with an ASA grade ≥ 3.

Table 2 Postoperative outcomes according to anastomotic configuration
Full size table
Table 3 Univariate and multivariate analyses for postoperative morbidity
Full size table

Table 4 reports the morbidity profile of the anti-peristaltic and iso-peristaltic side to side anastomosis according to stapled vs hand-sewn techniques. At multivariate analysis anastomotic leak (Table 5) was independent of the type of anastomosis performed, while was associated with an ASA grade ≥ 3, presence of perianal disease and ileocolonic localisation of disease (Montreal L = 3).

Table 4 Postoperative outcomes of the 379 patients who received a side to side anastomoses (stapled vs. hand-sewn techniques)
Full size table
Table 5 Univariate and multivariate analyses for anastomotic leak
Full size table

Other anastomotic techniques

The Kono-S anastomosis was performed in ten patients (2.3%) for ileocolic reconstruction. Only 1 patient (10%) in this group had a penetrating phenotype of disease compared to 29.3% in all the included patients. No patients in the Kono-S group had perianal disease, compared to 13.3% amongst all the included patients. Similarly a small proportion of patients in the Kono- S group were on steroidal treatment at the time of the surgery (10%) compared to approximately one third (32.1%) of the overall patient population.

There was a high rate of postoperative morbidity (25%) and anastomotic leak (10%) in the group of patients receiving an end to side ileocolic anastomosis, but this did not reach statistical significance.

LOS, readmissions and reoperations

Median LOS was 7 days (range 3–95). There were 17 reoperations (4%) and 21 readmissions (4.9%), with details provided in Table 6. Four predictors of LOS were identified after multivariate analysis. The laparoscopic approach was the only associated with a reduced LOS (p = 0.017), while age, ASA grade ≥ 3 or administration of preoperative TPN were associated with increased LOS.

Table 6 Readmissions and reoperations according to anastomotic configuration
Full size table

Discussion

The side to side was the most commonly used anastomotic configuration for intestinal reconstruction following ileocaecal resection for CD (89%) and stapled techniques (67%) were preferred over hand-sewn. We also found that both iso-peristaltic and anti-peristaltic orientations were commonly used, adding to the findings of the European Society of Coloproctology (ESCP) snapshot study on postoperative outcomes in CD, which did not differentiate between these two anastomotic orientations [10], whilst our study reported a slightly higher rate of postoperative morbidity with the anti-peristaltic configuration, even if this did not reach statistical significance. Moreover, compared to the ESCP snapshot study, our results included a larger cohort of patients, in spite of focusing on primary CD resection only, which was decided in view of the known higher postoperative complications rate following recurrent CD surgery [11]. Our study reported a postoperative morbidity rate of 20.3%, with 8.7% of the patients having Dindo–Clavien class 3 or higher complications, which calls attention to understanding the factors associated with a worse postoperative outcome.

Despite the possible associations with a higher rate of postoperative complications for the end to end ileocolic anastomosis [12] it has been previously reported that patients having an end to end anastomosis for intestinal reconstruction following CD surgery have better postoperative quality of life and reduced readmission to hospital at 2 years from surgery compared to the side to side anastomosis [13]. Moreover, the widely adopted denomination of the side to side anti-peristaltic anastomosis as “functional end to end anastomosis” has also been questioned by experimental studies evaluating peristalsis and bacteria at the anastomotic site [14]. In our cohort of patients, an end to end anastomosis was used in only 17 cases (4%). Additionally, a very small proportion of patients had a Kono-S ileocolic anastomosis formation (2.3%), which seemed to be preferred in patients at low risk for postoperative complications and recurrence, as no patient suffered from perianal disease and only one patient was affected by a penetrating phenotype of CD.

The high number of patients recruited and the multicentre design are the main strengths of our study, together with the focus on key performance indicators of CD surgery, outcomes which are often not so thoroughly audited as in cancer surgery. This national snapshot study also allowed collection of data on the use of laparoscopic surgery, which was the chosen approach in 72.8% of the patients with a conversion rate of 7.7%, similar to the experience reported in two specialist centres [15] including 538 patients with a 71.2% use of laparoscopy and 12.3% conversion rate. We reported an anastomotic leak in 3.7% of the patients, which compares favourably with published rates of 4.5% [16] when the ileocolic anastomosis is not protected by a diverting stoma. We found that anastomotic leak was associated with ASA grade ≥ 3, a penetrating phenotype of disease and ileo-colonic distribution of CD [17], which could guide the decision-making on when to fashion an ileostomy rather than a primary anastomosis.

Our study did not evaluate the extent of the mesenteric resection. Coffey et al. showed a significantly decreased surgical recurrence rate when incorporating a substantial portion of mesentery in the resected specimen [18]. While the mesentery is likely to play a pathogenic role in CD, it is also crucial for intestinal perfusion, and extensive removal may compromise bowel tissue with concerns also regarding haemorrhagic dangers associated with division of the mesentery in patients with CD and potential need for increased length of resected bowel if larger mesenteric segments are excised [19]. Another limitation of our study is the self-reporting nature of the data collection and the retrospective design. Whilst the lack of statistically significant differences for the primary and secondary outcomes according to anastomotic technique or configuration might be due to the limited power of the study, it certainly highlights the need for multidisciplinary led decision making and patient centred approach, moving the focus from a single causative factor (i.e. anastomotic technique, preoperative treatment) to a multifactorial approach.

Conclusions

The side to side was the most commonly used anastomotic configuration for ileocolic reconstruction following primary CD resection. There was no difference in postoperative morbidity according to anastomotic technique and configuration. Anastomotic leak was associated with ASA grade ≥ 3, a penetrating phenotype of disease and ileo-colonic distribution of CD.