Preoperative biomarkers related to inflammation may identify high-risk anastomoses in colorectal cancer surgery: explorative study

Abstract

Background: Colorectal anastomotic leakage can be considered a process of failed wound healing, for which related biomarkers might be a promising research area to decrease leak rates.

Methods: Patients who had elective surgery with a primary anastomosis for non-metastatic colorectal cancer, at two university hospitals between 1 January 2010 and 31 December 2015 were included. Patients with an anastomotic leak were identified and matched (1:1) to complication-free controls on the basis of sex, age, tumour stage, tumour location, and operating hospital. Preoperative blood samples were analysed by use of protein panels associated with systemic or enteric inflammation by proteomics, and enzyme-linked immunosorbent assays. Multivariable projection methods were used in the statistical analyses and adjusted for multiple comparisons to reduce false positivity. Rectal cancer tissue samples were evaluated with immunohistochemistry to determine local expression of biomarkers that differed significantly between cases and controls.

Results: Out of 726 patients undergoing resection, 41 patients with anastomotic leakage were matched to 41 controls. Patients with rectal cancer with leakage displayed significantly elevated serum levels of 15 proteins related to inflammation. After controlling for a false discovery rate, levels of C-X-C motif chemokine 6 (CXCL6) and C-C motif chemokine 11 (CCL11) remained significant. In patients with colonic cancer with leakage, levels of high-sensitivity C-reactive protein (hs-CRP) were increased before surgery. Local expression of CXCL6 and CCL11, and their receptors, were similar in rectal tissues between cases and controls.

Conclusion: Patients with anastomotic leakage could have an upregulated inflammatory response before surgery, as expressed by elevated serological levels of CXCL6 and CCL11 for rectal cancer and hs-CRP levels in patients with colonic cancer respectively.

Fig. 1

Study flow chart of patients reported to the Swedish Colorectal Cancer Registry to have undergone elective resection with a primary anastomosis for non-metastatic colorectal cancer at either Uppsala or Umeå university hospitals between 1 January 2010 and 31 December 2015 Anastomotic leakage was determined by review of medical records using an international consensus definition (see main text for further details), while potential controls also underwent chart review. *Complication-free corresponding to a Clavien–Dindo score 0 postoperative course.

Fig. 2

Preoperative distributions (tertiles) and a grouped discriminant analysis of proteins CXCL6 and CCL11 that were significantly higher in patients with anastomotic leakage after rectal cancer resection (34 cases and controls matched 1:1) Receiver operating characteristic curves and an estimation of area under the curve (AUC) denoting prediction accuracy. QDA, quadratic discriminant analysis.

Fig. 3

Representative images of immunohistochemical staining of normal tissue at the resection margin from rectal cancer patients defined as cases (anastomotic leakage) and controls (complication-free postoperative course) using antibodies against CXCL6 and CCL11 a,b CXCL6 is expressed in the mucosal epithelium and by some stromal cells. c,d CCL11 is mainly expressed by the epithelium. As illustrated, there were no signs of differences in expression pattern or intensity in patients with anastomotic leakage, and those with a complication-free postoperative course.