Umeå University's logo

Introduction:

Anastomotic leakage is one of the most serious complications following anterior resection for rectal cancer. Despite improvements in perioperative care and surgical techniques, leakage still occurs in 10–20% of patients. Identifying modifiable risk factors remains crucial, as the mechanisms predisposing to leakage are multifactorial and incompletely understood. Patient-related factors, tumour characteristics, surgical technique, and perioperative pharmacological exposures may all influence the risk for anastomotic leakage. In particular, the potential protective effect of preoperative β-blocker therapy has recently gained attention, but the evidence is so far limited. Beyond an increase in mortality, anastomotic leakage may have far-reaching consequences. Leakage has been suggested to impair oncological outcomes, and it is increasingly recognised as a potential cause of bowel dysfunction and reduced quality of life. As survival rates after rectal cancer continue to improve, long-term functional outcomes and quality of life have become increasingly important. However, the mechanisms linking anastomotic leakage to recurrence, bowel function, and health-related quality of life are not fully understood. To address these gaps, four consecutive multicentre studies were conducted, establishing the empirical foundation for this thesis. These studies investigated preoperative medication use in relation to anastomotic leakage, whether the inflammatory response after surgery mediates the relationship between leakage and recurrence, the long-term impact of leakage on bowel dysfunction, and the influence of leakage on long-term health-related quality of life after anterior resection for rectal cancer.

Aim and research questions:

The overall aim of this thesis was to improve the understanding of anastomotic leakage after anterior resection for rectal cancer by exploring how preoperative medication use may influence leakage risk, how postoperative inflammatory responses relate to subsequent voncological outcomes, and how leakage affects long-term patient-reported bowel function and quality of life. The specific research questions were:

1. Is preoperative use of β-blockers and other commonly prescribed drugs related to an increased risk of anastomotic leakage after anterior resection for rectal cancer (Study I)?

2. Does anastomotic leakage after anterior resection for rectal cancer influence recurrence-free survival, and is any such effect mediated by the postoperative inflammatory response, as measured by CRP (Study II)?

3.Does anastomotic leakage after anterior resection for rectal cancer influence bowel function (Study III) and health-related quality of life (Study IV)?

Methods:

This thesis is based on four multicentre cohort studies investigating anastomotic leakage after anterior resection for rectal cancer. The studies combine chart-reviewed data from 11 Swedish hospitals and population-based data from the Swedish Colorectal Cancer Registry (SCRCR), a nationwide quality registry with >99% coverage of rectal cancer surgery.

Studies I and II were retrospective multicentre studies, including 1,126 patients who underwent anterior resection between 2014 and 2018. Clinical and perioperative variables were retrieved from medical records and entered into an electronic database and subsequently linked with SCRCR data.

In Study I, the main exposure was preoperative medication use, with β-blockers as the primary variable of interest. The outcome was any anastomotic leakage within 12 months, classified according to the International Study Group of Rectal Cancer definition. Logistic and Poisson regression, supplemented by targeted maximum likelihood estimation, were applied to estimate associations between medicationviuse and leakage risk. The primary aim was to evaluate whether β-blocker use was predictive of anastomotic leakage.

Study II used the same cohort to examine recurrence-free and overall survival in relation to anastomotic leakage and postoperative systemic inflammation, measured as the highest C-reactive protein (CRP) concentration within 14 days. Linear and Cox regression models were used to assess associations, and a mediation analysis decomposed total, direct, and indirect effects of leakage via CRP. The main aim was to investigate whether anastomotic leakage after anterior resection for rectal cancer was associated with recurrence-free survival, and whether this could be explained in part by the postoperative inflammatory response, measured by CRP.

Studies III and IV were nationwide, population-based studies using SCRCR data for patients operated between 2015 and 2017. Survivors three years after surgery received validated patient-reported outcome questionnaires by mail.

Study III assessed bowel function using the Low Anterior Resection Syndrome (LARS) questionnaire. The primary outcome was major LARS among patients without a permanent stoma, with a secondary composite outcome of major LARS or permanent stoma. Propensity-score weighting was used to adjust for confounding. The primary aim was to assess whether anastomotic leakage influenced the risk of major LARS in the long-term.

Study IV examined health-related quality of life using the EORTC QLQ (European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire) C30 and CR29 instruments in the same cohort. The primary outcome was the C30 summary score, and secondary outcomes were the remaining functional and symptom scales. Analyses were stratified by stoma status, and targeted maximum likelihood estimation was used to adjust for confounding. The aim was to evaluate the impact of leakage on long-term quality of life.

Across all studies, confounders were selected based on causal diagrams, and missing data were handled using multiple imputation by chained equations. Statistical analyses were performed using R (version 4.4.0) and STATA (version 17).

Results:

Study I

A total of 1,126 consecutive patients who underwent anterior resection for rectal cancer at 11 Swedish centres between 2014 and 2018 were included in the study cohort. Anastomotic leakage within 12 months occurred in 232 patients (20.6%), comprising 67% staple-line insufficiencies, 28.8% pelvic abscesses, and 4.2% fistulas. Most leaks occurred within 30 days (69.7%), with a median onset on postoperative day 16. Preoperative β-blocker use was recorded in 255 patients (22.6%). Leakage rates were similar between users and non-users (20.8% vs 20.5%). In adjusted logistic and Poisson regression analyses, no association was observed between preoperative β-blocker use and anastomotic leakage (OR 0.94, 95% CI 0.64–1.38; IRR 0.95, 95% CI 0.68–1.33). Causally oriented analyses using targeted maximum likelihood estimation yielded similar findings (RR 0.98, 95% CI 0.73–1.31). Sensitivity analyses for early and late leakage, as well as adjustment for specific cardiovascular comorbidities, yielded similar results.

Study II

Using the same multicentre cohort after application of additional exclusion criteria, 1,036 patients remained for analysis. Anastomotic leakage occurred in 21.0% of patients (14.2% grade A, 56.4% grade B, and 29.4% grade C). Median maximum postoperative CRP levels were higher in patients with leakage (218 mg/L) than in those without (108 mg/L). Leakage was more frequent among men, smokers, and patients with higher American Society of Anesthesiologists’ (ASA) grade, as well as in those who received neoadjuvant therapy and total mesorectal excision. At a median follow-up of 61 months, mortality was 11.5% in patients with leakage and 12.6% in those without; recurrence-free survival was 82.6% and 77.8%, respectively. Overall survival was comparable, although grade C leaks showed slightly poorer outcomes. Mediation analysis demonstrated that leakage increased postoperative CRP by approximately 99 mg/L, but CRP itself had no measurable impact on recurrence or death (HR 1.00, 95% CI 1.00–1.00). The total effect of leakage on recurrence or death was estimated with a HR 0.66 (95% CI 0.43–0.94), driven by a direct effect (HR 0.59, 95% CI 0.38–0.86). Sensitivity analyses restricted to severe (grade C) leakage showed no clear association with recurrence or mortality.

Study III

After identification using the nationwide SCRCR, 1,778 patients whoviiiunderwent anterior resection for rectal cancer between 2015 and 2017 were contacted, of whom 1,178 responded (response rate 66.2%). Among 1,033 patients without a permanent stoma, 52 (5.0%) had a registered anastomotic leakage. Major LARS was reported by 69.2% of patients with leakage and 52.9% of those without. After adjustment for confounding using propensity score weighting, anastomotic leakage significantly increased the risk of major LARS (OR 2.09, 95% CI 1.13–3.87), corresponding to an adjusted absolute risk difference of 17.5%. The association was stronger among patients requiring reoperation (OR 2.78, 95% CI 0.87–8.91) and when including permanent stoma in the composite outcome (OR 3.90, 95% CI 2.20–6.91).

Study IV

The same nationwide cohort as in Study III was used to assess health-related quality of life three years after surgery. Among 1,178 responders, 104 (8.8%) had experienced anastomotic leakage. Patients with leakage reported lower EORTC QLQ-C30 summary scores compared with those without (80 vs 86, p < 0.01), a small but statistically significant difference (-4 points after adjustment, p < 0.01). In addition, lower scores were observed for global health status, role and emotional functioning, and higher scores for fatigue, pain, dyspnoea, and insomnia. In the colorectal cancer-specific QLQ-CR29 module, patients with leakage reported moderately worse body image and depending on stoma status, more perianal skin irritation (without stoma) or increased stool leakage from the stoma bag (with stoma). No clinically meaningful differences were observed in overall health-related quality of life beyond these domains.

Conclusion:

Anastomotic leakage after anterior resection for rectal cancer remains a frequent complication. Preoperative β-blocker use was not predictive of leakage, and postoperative inflammation did not mediate an increased risk of recurrence or death. At long-term follow-up, leakage was related to worse bowel function, reflected by a higher risk of major LARS, and when permanent stoma was included in a composite outcome, the association was even stronger. The overall impact on health-related quality of life was small and mainly limited to specific domains such as body image and stoma-related symptoms. Overall, these findings indicate that the main long-term burden of leakage is functional rather than oncological.