ePoster
Talk Description
Institution: University of Auckland - Auckland, Aotearoa New Zealand
Purpose: Anastomotic leaks (AL) are a major complication following colorectal surgery, with an incidence of 2-10% and a mortality rate of 10-15%. Early detection of AL is difficult, and diagnosis is often delayed, leading to additional, avoidable harm. Several drain biomarkers have been investigated as early markers of AL, but their relative accuracy remains unclear. This review aimed to determine the diagnostic accuracy of drain fluid biomarkers for the early detection of colorectal AL.
Methodology: A systematic review of the Scopus, Embase and Medline databases was performed from inception to July 2025. Articles including colorectal surgeries with anastomoses, with measurement of drain fluid biomarkers up to and including the seventh post-operative day, were included. Details on biomarkers, timing of measurement and diagnostic performance for AL detection were extracted. Meta-analyses were performed to generate pooled sensitivity and specificity using random effects models.
Results: Searches identified 2992 studies, of which 55 were included, representing 6939 patients. A range of luminal, inflammatory and metabolic biomarkers were studied. The most studied biomarkers were amylase, pH and C-reactive protein. The most accurate biomarkers were amylase (pooled sensitivity: 99.6%, specificity: 98.5, area under the curve 0.93) and pH (pooled sensitivity: 88.1%, specificity: 94.1%, area under the curve 0.96). Inflammatory biomarkers and cytokines showed fair-moderate diagnostic performance
Conclusion: Drain fluid biomarkers have the potential to detect AL earlier than current clinical methods following colorectal surgery. Further well-powered studies are needed to assess whether timely biomarker-guided intervention can improve the early identification and management of colorectal AL.
Speakers
Contributors
Authors
Mr Samuel Lewis - , Dr William Xu - , Dr Chris Varghese - , Prof Greg O'Grady - , Prof Ian Bissett - , Dr Cameron Wells -
