A novel scoring system for the early detection of anastomotic leakage: bedside leak score-a pilot study.

BACKGROUND

Anastomotic leakage is a major complication in colorectal surgery, resulting in significant morbidity and mortality rates. Despite substantial progress in surgical technique, anastomotic leakage rates remain stable. An early diagnosis of anastomotic leaks was proven to reduce adverse outcomes and improve survival.

OBJECTIVE

This study aims to find a novel scoring system for detecting anastomotic leaks using inflammatory and nutritional indicators after colorectal surgery. Our purpose was to analyze the diagnostic accuracy of leak scores in predicting postoperative complications.

DESIGN

The study included colorectal cancer patients who underwent curative surgery at Koc University Hospital between 2014 and 2018. Patients were categorized into two groups depending on the presence of anastomotic leaks and compared in terms of preoperative albumin levels, CRP levels in postoperative days 1 and 3, anastomotic leakage rates, length of hospital stay, and CRP quotient, which was calculated by dividing POD 3 CRP level to POD 1 CRP level. The bedside leak score is calculated by dividing the CRP quotient by the preoperative albumin level. The predictive value of bedside leak score, CRP quotient, and preoperative albumin levels in estimating anastomotic leakage was analyzed, and a cutoff value for the leak score was calculated.

RESULTS

A total of 184 patients were included in the study. The leak score, CRP POD 3-1 ratio, and preoperative albumin levels were found to successfully detect anastomotic leakage. The area under the curve for the leak score was calculated as 0.78. The optimal cutoff value was found to be 50.3 for the bedside leak score, which shows 90.9% sensitivity and 59.3% specificity.

CONCLUSION

The leak score may represent a valuable diagnostic tool for detecting patients at risk for anastomotic leakage after colorectal surgery and planning a better strategy to reduce morbidity and mortality rates and associated costs. However, further multicenter studies with large cohorts are necessary to confirm these results.