Introduction

Construction of a colorectal anastomosis is frequently performed in both elective and emergency colorectal surgery. Despite considerable data available on risk factors for anastomotic leak (AL), preventing this remains a challenge with a reported incidence of 3–19% depending on leak definition and quality of data collection [1,2,3]. AL is a serious complication, which can cause peritonitis and sepsis, subsequently leading to increased mortality (0.7–15.2%) [4, 5]. Moreover, the economic burden associated with AL is significant, increasing the cost of index admission by an estimated $ 30,670 in the US ($ 48,982 vs. $ 18,312 in patients without AL; p < 0.0001) [6].

While several modifiable and nonmodifiable risk factors have been described in the literature [1, 7,8,9], ultimately the patient and operating surgeon must balance the individualised risk-to-benefit ratio of performing an anastomosis. Surgical decision making is a complex process, which although informed by protocols and evidence-based guidelines, is also largely influenced by the surgeon and patient’s personality, experience and expectations [10, 11].

The present paper evaluates the AL rates and short-term outcomes of patients undergoing colorectal surgery with an anastomosis performed by a single surgeon, using a predefined risk reduction strategy.

Methods

This study is reported using the REporting of studies Conducted using Observational Routinely collected health Data (RECORD) Statement [12]. This study was approved by the Central Adelaide Local Health Network Human Research Ethics Committee (HREC Reference number: HREC/15/RAH/186, CALHN Reference number: R20150524). All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008.

Study design and participants

The study population consisted of sequential patients who underwent elective or emergency, open, laparoscopic or robotic colorectal resections with formation of a small bowel, colonic or colorectal anastomosis for any indication by a single surgeon (TS) or trainee under his supervision. Patients without a gastrointestinal anastomosis were excluded. All patients underwent surgery from June 2017 to July 2020, at the Royal Adelaide Hospital (RAH) and St. Andrews Hospital in Adelaide. The RAH is an 800-bed public teaching hospital, and St. Andrews Hospital is a 240-bed private hospital, both with tertiary colorectal units.

Variables

Data pertaining to baseline patient characteristics, perioperative care and postoperative course were collected prospectively. Complications were classified using the Clavien–Dindo classification [13]. AL was defined as an extra luminal collection around the anastomotic site confirmed on a contrast-enhanced CT scan and/or at re-laparotomy [14]. Severity of AL was graded as A, B or C [15]—grade A: no therapeutic intervention other than antibiotics, grade B: therapeutic intervention but manageable without re-operation (this includes percutaneous drain), grade C: requires re-operation.

Anastomotic leak risk reduction strategy

All patients who were considered for anastomosis underwent this under the guidance of a predefined risk reduction strategy which was routinely used by the operating surgeon:

Risk assessment

This was done routinely using the validated www.anastomoticleak.com risk calculator for elective colon cancer resection (Fig. 1; [5]). In patients who had operations for other indications, risk was estimated based on published independent risk factors [16, 17]. This estimated leak risk was discussed with the patient preoperatively as part of the consent process and to develop a risk profile for an anastomosis [18].

Fig. 1
figure 1

Screenshot of the anastomotic leak risk calculator. www.anastomoticleak.com

Full size image

Medication optimisation

Immunosuppressants, anti-inflammatories (including NSAIDs) and anticoagulants (including fish oil) were stopped if at all possible prior to surgery with enough time for effects to have worn off completely (depending on drug type) [19]. These medications were not restarted for 2 weeks after surgery if possible. The risk:benefit of this aggressive strategy was discussed with the patient prior to surgery. All patients were given intravenous ceftriaxone and metronidazole on induction, with mechanical bowel prep only used in left-sided cases, with a fleet enema in right-sided cases. All patients received laxatives from day 1 postoperatively.

Standardised technique

Excellent arterial inflow was aggressively pursued for both proximal and distal ends of all anastomoses by erring on the side of the blood supply. High volume pulsatile flow needed to be demonstrated at the marginal inflow for colonic ends, and at the divided mesenteric and mucosal edges in ileal, colonic, and rectal ends. If possible, all ileo-ileal, ileocolic, and ileorectal anastomosis were performed with a stapled side–side (functional end–to–end) anastomosis with a liner cutting stapler and transverse noncutting stapler [20]. At least 2 crotch sutures were performed to completely relieve the new join of tension, and the transverse staple line buried with interrupted 3‑0 Polydioxanone Suture (PDS). All left sided colorectal anastomoses were performed using a double-stapled technique with a contour 35 mm stapler to divide the rectum, and a circular stapler to complete the join. The anvil was brought out anterior to the transverse staple line (as this achieves relatively better vascularity since the stapler is taking relatively less mesorectum and more rectal wall) (Fig. 2). Laparoscopic/robotic staplers were not used to divide the rectum due to inferior angle of approach. Covering loop ileostomy was used in low and ultralow joins if one or more of the following risks were present: male, obese, neoadjuvant radiotherapy.

Fig. 2
figure 2

End-to-end anastomosis (EEA) using either 29 mm ETHICONTM Circular Stapler (Johnson & Johnson, NSW, Australia) or EEATM Circular Stapler with Tri-StapleTM Technology (MedTronic, SA, Australia)

Full size image

Statistical analysis

Data were analysed using IBM SPSS Statistics for Windows Version 26.0 (IBM Corp., Armonk, NY, USA). Continuous variable parametricity was tested using the Shapiro–Wilk test, and results are presented as mean (standard deviation) for parametric and median (range) for nonparametric data. Categorical data are displayed as frequency (percentage).

Results

A total of 191 colorectal resections were performed. Of these, 46 resections received an end stoma, and in 145 resections an anastomosis was formed. Patient demographics and surgical procedures are presented in Table 1. A defunctioning stoma was created in 22 (15.2%) cases and was most commonly performed after ultra-low/low anterior resections (15 of 51 low rectal joins) and modified exenterative procedures (4 of 51 low rectal joins).

Table 1 Patient demographics and surgical procedures
Full size table

Perioperative and postoperative data are shown in Table 2. AL occurred in 2 of 145 anastomoses (1.3%), and both were grade A leaks managed with antibiotics alone. There were no AL that required invasive intervention. Major complications (Clavien–Dindo ≥ grade 3) occurred in 13 (9.0%) patients. There were 2 (1.4%) postoperative deaths (Clavien–Dindo grade V); 1 patient died following a myocardial infarction at home after being discharged from hospital, and another patient died after aspiration pneumonia.

Table 2 Perioperative characteristics and postoperative complications after anastomosis
Full size table

Discussion

The results of this prospective analysis of a single-surgeon’s experience using a predefined risk reduction strategy demonstrates a low AL rate of 1.3% (n = 2) in 134 patients, and no patient requiring invasive intervention. This leak rate and severity is much lower than expected from other prospective data sets with accurate leak definitions [4, 21, 22].

The aetiology of anastomotic nonhealing in the modern era is postulated by the authors to relate more to physiological factors than surgical technique. In this series, pre-operative decision as to whether to restore continuity or defunction was heavily influenced by published nonmodifiable risk factors, rather than technical considerations [1, 7, 8]. Modifiable risk factors were aggressively optimised including using ensuring excellent intra-operative blood supply (recently been shown to be an independent predictor of AL) [9]. In addition, medication optimisation, including very limited perioperative use of anti-inflammatories and anticoagulants may have also contributed [9, 23].

There are significant limitations in this study. The small sample size and absence of a direct comparison precludes any definitive conclusions. In addition, it is unclear what contribution one or more of the risk reduction strategies had in isolation, whether there was an additive effect, or whether heightened vigilance in general had a role to play. In the future, comparative studies would be required to confirm the role of an active risk reduction strategy and incorporate investigation of newer concepts that may reduce the risk further such as modification of the microbiome, and immunofluorescence to further assess blood supply [24, 25].

Conclusion

A predefined risk reduction strategy may help improve patient selection and reduce anastomotic leak rates. A comparative study is warranted.

Main novel aspects

  • The authors show that using a predefined risk reduction strategy can help reduce anastomotic leaks in colorectal surgery.

  • Pre-operative decision whether to restore continuity or defunction should rely more on nonmodifiable risk factors rather than on surgical technique.

  • These findings provide the basis for a more comprehensive comparative study.