Introduction

Therapies of rectal cancer has improved and standardized during the last two decades. A multidisciplinary therapy including preoperative radiation combined with or without systemic chemo-therapy, followed by total mesorectal excision (TME), presents the gold standard in rectal cancer therapy [1,2,3,4,5,6]. Already in 1982, Heald et al. first described the importance of the surgical plane (in rectal cancer surgery) for an improved local recurrence rate [4]. During the last two decades, technical advances also include intersphincteric resection for ultra-low tumours, to reduce the rate of permanent stomas [1, 7, 8]. Moreover, minimal invasive surgery was introduced in the treatment for rectal cancer and showed — beside the short term benefits of laparoscopic surgery generally — similar oncological outcomes compared to conventional surgery [9,10,11,12,13].

The significant advantages of neoadjuvant radiotherapy were first observed in 1997 by the Swedish Rectal Cancer Study Group [14, 15]. The first retrospective case series about the benefits of a combined therapy were published in 1997 and 1998 [16,17,18], and the first randomized controlled trial was released in 2004 by Sauer et al. [6]. Ever since, the use of neoadjuvant therapy is recommended for all newly diagnosed rectal cancer patients with a clinical stage (c) T3 or T4, as determined with endorectal ultrasound or a pelvic magnet resonance imaging (MRI). However, anastomotic leakage (AL) poses the most relevant complication following low anterior resection (LAR) in patients with rectal cancer independent of surgical approach or technique. In previous published literature the rate of AL in rectal cancer surgery differs between 1 and 30% [19,20,21,22,23]. Furthermore, AL is associated with an increased morbidity, longer hospitalization, more costs, and even worse oncological outcome [24, 25]. Many factors have been reported for developing an AL, such as male sex, obesity, malnutrition, anaemia, weight loss, untreated stenosis before therapy, low level of anastomosis, preoperative radiation therapy, and prolonged operation time [22, 26, 27].

To prevent AL, various strategies were established after LAR — the most common procedure is a defunctioning stoma (DS) [21, 22]. It can reduce intraluminal pressure of the anastomosis through bypassing the bowel content [21]. However, variable disadvantages exist regarding a DS, such as high output stomas, renal dysfunction, bowel obstruction, stoma prolapsed, and a reduction of quality of life. Another possibility for the prevention of Al is a transanal tube placement which shows favourable effectiveness compared with non-prevention patients [21, 28]. The treatment of AL is challenging, and some patients even require reoperations, EndoVAC® therapy, or permanent stomas.

The aim of this study was to evaluate risk factors for the occurrence of AL and their influence on the oncological long-term outcome of patients who received preoperative therapy followed by low anterior rectal resection with primary anastomosis in locally advanced rectal adenocarcinoma.

Methods

Patients and eligibility

In accordance with ethical review guidelines (EK-0.04–391) and after an institutional review board approval, data were retrieved from a prospective maintained database of the Academic Teaching Hospital in Feldkirch. From January 1998 to December 2020, all locally advanced rectal cancer patients (cT3, cT4, N +) who received neoadjuvant therapy followed by TME were included in this study. Exclusion criteria were defined as surgery without primary anastomosis, patients who were inoperable, no surgery because of other reason, metastasized situation (UICC IV), and patients who were lost of follow-up or missing of perioperative details. All patients received preoperatively a colonoscopy including tissue biopsy and histological tumour examination. Tumour height was analysed with rectoscopy, and locally tumour staging was assessed with pelvic magnet resonance imaging (MRI). Additionally, in case of very low carcinomas or in question of sphincter involvement, endorectal ultrasound was used. To rule out distant metastasis, a computer tomography (CT) of the trunk was performed. The variables included in the study comparing patients with [“Leakage + ”] and without anastomotic leakage [“Leakage – “] were as followed: age, sex, body mass index (BMI), American Society of Anaesthesiologists Score (ASA) [29, 30], comorbidities, tumour location (in cm of the anal verge), preoperative chemoradiation therapy (CRT), preoperative carcinoembryonic antigen (CEA) level, clinical stages, surgical approach, protective defunctioning stoma, operative time, preoperative and postoperative C-reactive protein (CRP) values, preoperative and postoperative white blood cells (WBC) count, preoperative and postoperative Haemoglobin (Hb) value, postoperative pathological stage, duration of hospital stay, postoperative complications (according to the Clavien-Dindo-classification [31]), local recurrence, distant metastasis, cancer related death, disease-free survival (DFS), and overall survival (OS). Pathological and clinical staging was based on the 8th edition of the Union for International Cancer Control (UICC) TNM Classification of malignant tumours [32]. Local recurrence was set as any recurrence in the pelvic cavity confirmed by radiological, clinical or pathological evaluation. DFS was defined as months from the date of surgery to the date of detection of recurrency or metastasis, last follow-up, or death. OS was defined as months from the date of surgery to the date of death. Included and excluded patients are shown in Fig. 1. A total number of 163 patients were included in this study.

Fig. 1
figure 1

Flow chart of excluded and included patients who underwent rectal cancer surgery followed by preoperative radiation or chemoradiation therapy at the Academic Teaching Hospital Feldkirch. Abbreviations: RTX radiotherapy, RCTX combined radio- and chemo-therapy, APR abdomino perineal resection, PME partial mesorectal excision, TME total mesorectal excision, UICC Union International Contre le Cancer

Full size image

Treatment strategy

All rectal cancer patients were preoperatively discussed in an interdisciplinary team discussion, and neoadjuvant therapy was indicated in accordance to international guidelines [2, 33]. Neoadjuvant therapy was performed in case of locally advanced (cT3-4, cN +) rectal cancer < 12 cm from the anal verge determined by CT-scan of the trunk, pelvic MRI, colonoscopy, and endorectal ultrasound. A Fluorouracil-based chemotherapy was given as standard therapy combined with radiotherapy before 2012. Based on the data from Hofheinz et al. [34], Capecitabine replaced Fluorouracil-based substances in 2012. In 11 patients, Oxaliplatin was given as a combined therapy with either Fluorouracil or Capecitabine. In case of short-course therapy, 5 Gy per day for 5 days (5 × 5 Gy) were irradiated and surgery was performed usually within 7 days.

Anastomotic leakage

Anastomotic leakage was defined as a failure of integrity of the intestinal wall at the anastomosis site that resulted in a communication of intraintestinal and extraintestinal compartments according to the statement of the International Study Group of Rectal Cancer [35]. The grading system in accordance with this publication was used to classify AL into 1 of 3 grades (A, B, or C) [36]. A was set as AL with no need of active therapeutic intervention, B as AL requiring active therapeutic intervention but manageable without re-laparotomy, and C as AL treated with re-laparotomy [35, 36].

Histopathological examination

The removed specimen was immediately fixed with formalin. The pathological examination included a macroscopic description of the specimen, a complete histopathological staging, and a scoring of regression grade according to Rödel et al. [37].

Follow-up

Follow-up was performed as described in the international guidelines of the European Society of Medical Oncology and German S3 Leitlinie [2, 33]. The patients were examined every 3 months during the first year after surgery followed by a 6-month examination interval. Each patient underwent a physical examination, complete blood cell count, serum CEA level assessment, and abdominal sonography. CT scans and pelvic MRI were performed yearly. The colonoscopy was routinely carried out 1 year after the surgery, and depending on the result, the interval was determined.

Statistical analysis

Statistical analysis was done using SPSS® (IBM™, New York). Continuous data of the patient characteristics were tested for normal distribution using the Shapiro–Wilk-Test [38]. Normally distributed data are presented as mean ± standard deviation, not-normally distributed data as median and interquartile range (IQR). Continuous data of included and excluded patients were assessed by either the t-test or Mann–Whitney-U-test. Categorical data are presented in absolute numbers (percent) and were assessed using the chi-square test or the exact Fisher test for small samples. Variables that could potentially influence the rate of AL were subjected to univariate analysis. The odds ratio (OR), its standard error, and 95% confidence interval (CI) were calculated according to Altman [39]. Parameters with a p-value ≤ 0.2, as well as age and BMI as possible independent risk factors, were entered in a multivariate logistic regression analysis. Data were collected using Excel® (Microsoft™, Seattle). Significance was set at a p-value of < 0.05.

Results

A complete follow-up was obtained from 163 patients of which 33 patients (20%) developed an AL. For five patients, no intervention was necessary (Grade A). Eleven patients required an intervention without re-laparotomy (Grade B). For 17 patients, it was necessary to proceed to a re-laparotomy for complication management (Grade C). Only in case of three patients (all Grade A), the leakage was diagnosed after discharge from hospital. Patients’ characteristics are shown in Table 1. The study group consist of 50 female (31%) and 113 male (69%) patients. AL occurred more often in male patients (22% vs. 16%, p = 0.41). Mean age at the date of surgery was 63 years (IQR: 56–72) in patients without leakage and 66 years (IQR: 60–73) in patients with AL. Median BMI and ASA-score [30] did not differ in both groups. In this study, comorbidities were significantly more often observed in patients who had a postoperative AL (38% vs. 61%, p = 0.049). The number of previous abdominal surgery and CEA level did not differ in both groups as well as the clinical stage, tumour height from the anal verge, and type of preoperative radiotherapy (short or long course).

Table 1 Patients’ characteristics, preoperative staging, and therapy
Full size table

Perioperative and postoperative results are shown in Table 2. The two frequently performed surgical methods were either an open or a laparoscopic low anterior resection with TME (open: AL − : 69%/AL + :73%, vs. laparoscopic: AL − : 31%/AL + :27%, p = 0.83). Leakage rate did not differ significantly between laparoscopic or open techniques (open: 21% vs. laparoscopic: 18%, p = 0.83). There were no differences regarding type of anastomosis (either end-end, site-end, or hand sewn colo-anal), operation time, and the rate of protective defunctioning stoma. Haemoglobin, WBC count, and CRP-value did not differ between both groups preoperatively and day one after surgery. CRP peak within 14 days after surgery was significantly higher in patients with AL (10 [IQR: 7–18] vs. 25 [IQR: 15–34], p = 0.02). Overall complication rate of the entire cohort was 36% (n = 59), whereby 20% (n = 33) of this complication was represented by the AL. Considerably more patients required surgical-, endoscopic-, or radiological re-intervention in case of complication in the AL Group, classified as Dindo-Clavien Grade III (11 [8%] vs. 31 [94%], p < 0.001). Furthermore, the duration between primary operation and stoma reversal was extensively longer (93 days [50–182] vs. 219 days [172–309], p < 0.001) and the rate of permanent stomas was doubled in patients with AL (18% vs. 36%, p = 0.03). Duration of hospital stay was significantly shorter in patients without AL (15 days [12–20] vs. 25 days [16–37], p < 0.001). Only one patient in each group died within 30 postoperative days (1% vs. 3%, p = 0.36).

Table 2 Operative results and postoperative complications
Full size table

Pathological results and follow-up data are shown in Table 3. There was no difference between both groups regarding postoperative primary tumour stage or lymph node involvement, tumour regression after neoadjuvant therapy and resection margins involvement. 73 patients (56%) without a leakage and 14 (42%) patients with AL received a postoperative adjuvant chemotherapy (p = 0.18). Median follow-up time was 153 months (95%CI: 127–179) and showing no difference between the groups. Local recurrency rate (7.7% vs. 9.1%, p = 0.73) and distant metastasis rate (25.4% vs. 18.2%, p = 0.50) did not differ in case of AL as well as the DFS (114 months, 95%CI: 56–172 vs. 67 months, 95%CI: 11–123, p = 0.26) and the OS (133 months, 95%CI: 94–172 vs. 82 months, 95%CI: 25–139, p = 0.15) compared to the patients with no AL. Figure 2 shows Kaplan–Meier survival curves of patients with and without AL. Oncological outcome showed not to be different in patients who were operated before 2010 or after 2010 (Suppl. Figure 1).

Table 3 Oncological results and follow-up data
Full size table
Fig. 2
figure 2

Kaplan–Meier survival curves of patients with or without an anastomotic leakage following neoadjuvant therapy and low anterior resection. OS overall survival, DFS disease free survival

Full size image

Univariate analyses of factors associated with an anastomotic leakage of patients who underwent preoperative therapy followed by surgery with primary anastomosis are shown in Table 4. Tumour distance lower than 6 cm from the anal verge (OR: 2.81, 95%CI: 1.08–7.29, p = 0.04) and comorbidities (OR: 2.22, 95%CI: 1.01–4.90, p = 0.049) revealed to be significant risk factors for developing an anastomotic failure after TME surgery with primary anastomosis. Age, BMI, distance lower than 6 cm from the anal verge, comorbidities, protective defunctioning stoma, and CRP > 9.2 were included into the multivariate analysis which is shown in Table 4. Comorbidities (OR: 4.41, 95%CI: 1.26–15.3, p = 0.02), BMI (OR: 0.68, 95%CI: 0.01–0.66, p = 0.02), and CRP at the first postoperative day (OR: 3.76, 95%CI: 1.11–12.70, p = 0.03) showed to be a significant risk factor for AL in the multivariate analysis. However, all evaluated risk factors had no influence on DFS or OS in the entire cohort (Table 5).

Table 4 Univariate and multivariate analysis of factors associated with an anastomotic leakage of patients who underwent preoperative chemo- (radiation-) therapy followed by surgery
Full size table
Table 5 Long-term outcome of risk factors associated with AL in the multivariate analysis
Full size table

Discussion

Anastomotic leakage is one of the most serious complication after low anterior resections and is associated with increased costs, longer hospital stay, increased postoperative morbidity and mortality, and higher rate of local recurrency [23, 40, 41]. In this study, oncological long-term data were not affected in case of AL. Low tumours and comorbidities revealed to be essential risk factors for the occurrence of AL but did not show any association with worse oncological outcome.

The incidence of AL has been reported to occur in 1–30% [23, 25, 42]. In this study, an anastomotic leakage rate of 20% was observed. In relation to previous published literature, the relatively high AL rate could be explained by a precise documentation of this prospective database. Moreover, all patients received follow-up and AL were well documented after discharge of the surgical department. However, it is already reported that rectal cancer in obese, diabetic male patients who smoke have a higher risk for anastomotic failure in rectal cancer surgery [43]. In this cohort, more male patients with AL were detected (22% vs. 16%), which was not statically significant but consistently to previous published literature [44, 45]. The rate of comorbidities was substantially higher in patients with AL (38% vs. 61%, p = 0.049) which was also revealed to be an essential risk factor for AL in the univariate and multivariate analyses. Underlying pulmonary- and pre-existent vascular disease are both reported to be independent risk factors for an AL in rectal cancer surgery in a meta-analysis published by McDermott et al. in 2015 [45]. Furthermore, patients with an ASA grade higher than II was revealed to have a higher risk for a leakage in colorectal surgery as well [45]. Controversially, our data did not show a difference regarding ASA-score. However, ASA-score is not fully concordant with pre-existent comorbidities. Moreover, the ASA-score is strongly depending on the experience of the anaesthesiologist who is preforming the evaluation before surgery.

Nevertheless, preoperative therapy for downsizing or downstaging of rectal tumours reduces local recurrency rates. The influence of preoperative therapy on AL is still discussed controversially [45]. In this study — especially short course radiotherapy — was associated with a high rate of AL (32%). Sparreboom et al. could clearly demonstrate that the rate of AL was high if the interval between radiotherapy and surgery was too short [46]. The high rate of AL after short course therapy in this study may partially be explained because of the influence of radiotherapy on the wound healing of the primary anastomosis.

In previous literature, several strategies to reduce the rate of AL were described. Proximal faecal diversion via loop ileostomy or colostomy is a common strategy to reduce the risk for AL. In the past, the role of a diverting stoma was discussed controversially. The first randomized controlled trials could not demonstrate a significant benefit [47, 48]. Anyhow, newer data clearly demonstrated a reduction of AL − and reoperation rate without increasing morbidity in case of performing a diverting stoma [49]. On the one hand, it can be assumed that faecal bypassing may reduce the severity of grade B or grade C leakages and may offer a chance for a radiological or endoscopic intervention. On the other hand, grade A leakages may not be detected at all and may heal before stoma closure. In this study, only 61% of the patients who developed AL received a protective stoma. This could partially be explained that the standard operation procedure regarding the protective stoma changed during the study period. All patients after 2014 received a protective stoma independent of pretended risk factors. The role of transanal tubes was also discussed controversially in the past. A recently published randomized trial by Zhao et al. demonstrated no significant difference regarding the AL rate [23]. Early detection of leakage and begin of therapy seems to be crucial for the outcome in case of this complication. In this study, only three patients with AL were diagnosed after discharge from hospital with grade A leakages. All other patients with AL were diagnosed during hospitalization. High CRP levels on postoperative day 4 and 7 should indicate the need for a more careful patient evaluation [50]. The data in this study showed significantly increased CRP peaks within the first 14 postoperative days in case of AL. On the first postoperative day, CRP was an individual risk factor in the multivariate analysis and underlines that CRP still seems to be an important marker for the detection of infectious complications after colorectal surgery.

It is well known that AL has a high impact on morbidity and mortality. It could be confirmed a longer hospital stay (15 vs. 25, p < 0.001), longer time between first operation and stoma reversal (93 vs. 219, p < 0.001), and a double high rate of permanent stoma (18% vs. 36%, p = 0.03) in this study. Operative technique or type of anastomosis had no influence on the AL rate in our examination. The benefits of minimal invasive surgery regarding short-term outcomes are well known in colorectal surgery [13]. Also, oncological findings do not decrease with the use of laparoscopic techniques [11, 51]. After all, independent of technique rectal cancer surgery has a high morbidity. Complication management strongly depends on surgeons’ individual experience and has a high influence on the patients’ outcome.

The influence of AL on long-term oncological survival outcomes is still under debate and two main reasons were previously described: Firstly, the immunological effect of an inflammation caused by a leakage. In a murine model, abdominal infection suppresses the number and activity of natural killer cells and promotes distant metastasis [20, 52]. Secondly, the delay or cancellation of oncological adjuvant therapy caused by an AL may influence a poorer outcome in rectal cancer patients [53]. Those assumptions are promoted by various previous published literature: In a recently published randomized control trial by Koedam et al., it was shown that oncological outcomes are negatively influenced by the occurrence of AL after rectal cancer surgery [54]. Similar results were shown by Yang et al. in a previous published meta-analysis of 44 698 patients [55]. The delay of adjuvant therapy and the influence of immunological effects on long-term data are serious. Also, in this study, grade C leakages showed the worst long-term outcome compared to grade B — or A leakages — even if not significant due to low number of patients (Fig. 3). Contrarily to previous studies and our data, Wang et al. showed no significant effect on distant recurrency in case of AL, whereas local recurrence rate was increased and overall survival was decreased [56]. However, other authors are reported that AL has no influence on local recurrence, disease free survival, or overall survival at all [1, 20]. In this study, DFS (114 months, 95%CI: 56–172 vs. 67 months, 95%CI: 11–123, p = 0.26) and OS (133 months, 95%CI: 94–172 vs. 82 months, 95%CI: 25–139, p = 0.15) showed no difference between patients with or without AL. Also, risk factors for AL had no influence on long-term survival data in this cohort. However, survival curves were better in patients without AL, even if not significant. The interpretation of those results is limited because of a relatively low number of patients included into this study.

Fig. 3
figure 3

Kaplan–Meier survival curves of type of AL (Grade A/B/C) and effect on oncological long-term data

Full size image

The distance between tumour and anal verge is widely accepted as surrogate for the height of the anastomosis and was proved to be a relevant risk factor for AL in the past [57]. In this study, tumour distance under 6 cm was revealed to be a significant risk factor for anastomotic failure (2.81 [1.08–7.29] p = 0.04). Local tumour control and sphincter preservation is an important consideration for patients with low tumour localization. In the context of comparable oncological findings of differing tumour heights, it may be functional aspects and quality of life which dictates the surgical strategy [58]. Nevertheless, AL can also decrease survival rate and quality of life after low anterior resection, which makes the decision between sphincter preservation or APR surgery even more complex.

This study has several limitations to be mentioned. Firstly, the study was designed as a retrospective non-randomized trial, which may cause a selection bias. A relatively high number of patients had to be excluded because of lost to follow up or missing of perioperative details (n = 71). Secondly, the quality of mesorectal excision is lacking. Long-term effects of the quality of surgery are well known and first described by Heald et al. in 1982 [4]. Macroscopic evaluation of the specimen by the pathologist was introduced in 2015 at our hospital, and details had to be excluded for this study. Finally, the relatively low number of patients might indicate a statistical bias.

However, based on our data, AL does not influence oncological long-term outcome of patients with locally advanced rectal cancer. AL is still a serious problem in rectal cancer surgery and is associated with a significant higher morbidity, longer hospital stay and an increased permanent stoma rate. It could be clearly demonstrated that low tumour localization and comorbidities are crucial risk factors for developing AL, whereas DFS and OS were not influenced by several risk factors. Independent of our data, morbidity is increased seriously in presence of AL. Preoperative patients’ selection and careful evaluation plays a decisive role in rectal cancer surgery to avoid AL. Regardless of the significance, a protective defunctioning stoma can be recommended for all patients who underwent preoperative therapy followed by LAR with primary anastomosis.

Conclusion

Anastomotic leakage remains a serious problem in rectal cancer surgery. AL was associated with a significant longer hospital stay, more permanent stomas, and a higher morbidity. Especially distance of tumour from the anal verge and comorbidities could be detected as independent risk factors for AL. Oncological findings and long-term outcome were not influenced by this specific complication.