Anastomotic leaks in stage IV colorectal cancer

Suat Chin Ng, Douglas Stupart, David Bartolo, David Watters

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)


Background: The purpose of this study was to determine the anastomotic leak rate for colorectal cancer resections in patients with metastases (compared to those without), and to determine the impact of anastomotic leaks on survival. Methods: This is a retrospective analysis of all patients who underwent resection and primary anastomosis for colorectal adenocarcinoma at a single institution between January 2002 and December 2014. Results: A total of 843 patients underwent a resection and primary anastomosis for colorectal adenocarcinoma (661 colon and 182 rectal). Of these, 135 (16%) had metastases and 708 (84%) did not. Anastomotic leaks occurred in 17 of 135 (13%) patients with metastases, and in 37 of 798 (5.2%) patients without metastases (P = 0.003). Peri-operative mortality occurred in 13 of 135 (9.6%) patients with metastases, compared with 19 of 708 (2.7%) patients without metastases (P = 0.0003). Anastomotic leak was associated with a reduction in overall survival (median survival 121 months without anastomotic leak versus 66 months in patients who had an anastomotic leak (P = 0.02)). If the patients who died peri-operatively are excluded from this analysis, however, long-term mortality was similar (125 months versus 101 months; P = 0.70). Conclusion: Metastatic disease was associated with an increased risk of anastomotic leak and a higher peri-operative mortality rate after colorectal resections for cancer. Patients with anastomotic leaks had a higher peri-operative mortality rate, but long-term survival was unaffected beyond the peri-operative phase.

Original languageEnglish
Pages (from-to)E649-E653
JournalANZ Journal of Surgery
Issue number9
Publication statusPublished - 1 Sep 2018


Dive into the research topics of 'Anastomotic leaks in stage IV colorectal cancer'. Together they form a unique fingerprint.

Cite this