Right hemicolectomy anastomotic leak study: a review of right hemicolectomy in the binational clinical outcomes registry (BCOR)

Introduction

Colorectal cancer (CRC) is a common malignancy in Australia.¹ Surgical management remains the mainstay of treatment for right-sided colon cancer. Anastomotic leak (AL) is a serious complication of colectomy and is responsible for significant morbidity and mortality. In particular, AL is associated with a delay to commencing adjuvant chemotherapy and poorer cancer-specific and overall survival.^2-4

In the Australian and New Zealand context, individual colorectal unit and multicentre state databases have enabled the examination of the AL rate in colorectal surgery, including following a right hemicolectomy for colon cancer.^5-8 However, risk factors contributing to AL in right hemicolectomy have not been extensively studied at the national or binational level, particularly with respect to the impact of different anastomotic techniques which may contribute to AL.

The Bowel Cancer Outcomes Registry (BCOR) is a prospectively collected clinical quality registry established in 2007 by members of the Colorectal Surgical Society of Australia and New Zealand (CSSANZ). It is a prospective audit of patients presenting with CRC, including those who have undergone colectomy with or without anastomosis for colon cancer. The BCOR currently includes patients from 175 hospitals, with participation mandated for colorectal training units. This study aimed to examine the AL rate from the BCOR patient information collated from 2007 to 2021 to identify potential factors that may contribute to AL.⁹

Materials and methods

Results

Discussion

This study examined the Australian and New Zealand experiences of AL in patients who had a right hemicolectomy intended for colon cancer. This is the first study to provide a comprehensive review of the clinicopathological features of patients who underwent a primary anastomosis together with risk factors associated with an anastomotic leak recorded in the binational registry. An AL rate of 2.0% was identified using BCOR data. Risk factors associated with AL were age greater than or equal to 75, male gender, public health insurance status, open surgery and emergency surgery.

A wide range of AL rates has been reported in the literature. Globally, national, international, and local studies have reported an incidence between 1.3% and 8.4%.^{2, 10-13} It is noteworthy that the AL rate in this study is lower than that in similar national and international studies with mandatory data collection, likely reflecting the difference between voluntary and involuntary data collection and analysis. Additionally, the wide-ranging reported leak rate reflects the continuing heterogeneity in the definition of AL.¹⁴ In the BCOR, AL is broadly defined as ‘clinical and/or radiological evidence of leak from an anastomosis’.⁹ The lack of an international consensus definition impedes standardized reporting and comparability between databases.^{15, 16} The inherent subjectivity in the clinical and radiological assessment of AL presents a significant barrier to achieving such a consensus, an issue noted in multiple Delphi studies.^{15, 17, 18} We strongly advocate for a uniform definition to enable meaningful comparisons between different datasets and to more accurately document the incidence of AL.

In this analysis, higher ASA status, open surgery, emergency operation, public health insurance status, and male sex significantly contributed to the AL rate. These risk factors are consistent with those identified in previous studies.^{2, 10, 12, 19} Public health insurance status likely reflects that the majority of open and emergency cases with higher ASA status are undertaken in a public hospital setting. This is supported by the strong association found in this study between public health insurance status, increasing ASA status, open surgery, and emergency surgery. Previous studies have demonstrated an association between AL and additional risk factors including poor nutritional status, diabetes mellitus and smoking.^{10, 12, 20-22} Unfortunately, this degree of data granularity was not available in the BCOR database, which is a limitation of this study. Specific reference to pre-existing patient comorbidities and the likelihood of developing an AL would likely improve the quality of information and assist in better risk stratification for AL.

In the 13 147 patients examined, the anastomotic technique was not a risk factor for AL (n = 7827 stapled anastomoses versus n = 1306 handsewn anastomoses). The majority of patients in the BCOR had a stapled anastomosis, reflecting the clinical practice of Australian and New Zealand surgeons. The clinical significance of this finding, however, is difficult to determine given that this data point was not recorded in close to one third of patients.

Several studies have focused on whether a handsewn versus stapled ileocolic anastomotic technique is the more likely contributor to AL rates, with conflicting findings. A 2011 Cochrane review pooling 835 patients who had undergone a right hemicolectomy for cancer from four studies, reported that a stapled technique was associated with a lower AL rate compared with a handsewn technique of re-establishing intestinal continuity (OR 0.28; 95% CI 0.10–0.75).²³ Similarly, in a multicentre study of 957 colon cancer patients who had a right hemicolectomy by Puleo et al. (2013),²⁴ the AL rate was 2.5% for stapled anastomoses and 4.9% for handsewn anastomoses (P < 0.05). Conversely, a number of subsequent, large national and multicentre European studies have found in favour of a handsewn technique, demonstrating up to double the risk of AL in stapled ileocolic anastomoses.^{11, 12, 19, 25} A Danish retrospective cohort study (n = 1414) demonstrated a leak rate of 5.4% (21/391) in the stapled group compared with 2.4% (24/1023) in the handsewn group.¹⁹ Similarly, an international prospective observational study (n = 2515) demonstrated a significantly increased risk of leak associated with a stapled anastomosis (OR 1.5; 95% CI 1.1–2.3).²⁶ A large-scale randomized control trial assessing different anastomotic techniques would be ideal. However, the ethical considerations of such a study are limited by underlying surgeon preference and the requirement for situational intraoperative decision-making regarding the choice of anastomotic technique. An alternative interpretation of this finding is that anastomotic technique is of minor importance. Rather, other considerations, including careful patient selection, surgeon judgement regarding the intraoperative findings, patient physiological response and adherence to sound surgical principles when constructing an anastomosis, are of equal, if not greater, significance.

Comparing the characteristics of patients who had a primary anastomosis constructed with those who did not have intestinal continuity re-established at the index operation, enabled the examination of factors contributing to the intraoperative decision-making regarding primary anastomosis formation. Patients with a higher ASA status, advanced tumour stage and an urgent operation were less likely to have a primary anastomosis formed. Interestingly, only an urgent operation was associated with an increased risk of AL. Not surprisingly, patients with a higher ASA status were more than twice as likely not to have a primary anastomosis formed. This suggests that this is either over-estimated as a risk factor for AL in intraoperative decision making, or more likely, that these patients are simply under-represented in the primary anastomosis group. Patients with private health insurance status and those who had minimally invasive surgery were more likely to have an anastomosis formed. This is consistent with the finding that these factors were associated with a decreased risk of AL. Male sex, which was identified as a risk factor for AL, was not significantly different between the anastomosis and no anastomosis groups. The underlying reason for the role of sex as a risk factor for AL remains unclear.

This study had several limitations. A retrospective review, albeit of prospectively collected data, presents inherent limitations in interpreting causality between the identified significant risk factors and AL. Additionally, a subgroup analysis to clarify the impact of emergency compared with elective surgery on AL would be ideal. This is however impacted by the low event rate of AL which limits the ability of the multivariate model to account for confounding factors, impacting the reliability of such an analysis. Furthermore, comparing outcomes of stapled or handsewn anastomoses, the majority of patients had a stapled anastomosis, introducing the possibility of selection bias. Interestingly, 353 patients are recorded as having Stage 0 disease. It is unclear from the recorded information whether these patients had an endoscopically treated lesion and a subsequent staged resection. The availability of these data points would aid interpretation of data analysis. Although advantageous with respect to providing macroscopic binational data, a large multicentre database has limitations with regards to the completeness of the data and the ability to collect detailed clinicopathological data points (e.g., patient-related risk factors such as smoking and diabetes mellitus). The inclusion of these data points could potentially aid in identification of further causative factors leading to AL.

This study provides a comprehensive overview of AL in patients treated by right hemicolectomy for colon cancer in Australian and New Zealand. Clinically significant patient-related and operative risk factors were identified. Anastomotic technique was not associated with AL incidence. Ultimately, a consensus definition of AL must be implemented in addition to a more comprehensive documentation of the many additional patient, clinical and institutional factors that contribute to AL. Knowledge of these variables is of critical importance in identifying patients at risk of AL and in the overall tailored management of patients who develop an AL.

Author contributions

Sireesha Koneru: Data curation; formal analysis; investigation; methodology; writing – original draft; writing – review and editing. Mifanwy M. Reece: Conceptualization; data curation; investigation; methodology; supervision; writing – review and editing. Dulani Goonawardhana: Conceptualization; writing – review and editing. Pierre H. Chapuis: Investigation; supervision; validation; writing – review and editing. Krishanth Naidu: Investigation; methodology; writing – review and editing. Kheng-Seong Ng: Data curation; formal analysis; investigation; methodology; supervision; validation; writing – review and editing. Matthew J. F. X. Rickard: Conceptualization; investigation; methodology; project administration; supervision; validation; writing – review and editing.

Conflict of interest statement

None declared.