Diagnostic accuracy of C-reactive protein, procalcitonin and neutrophils for the early detection of anastomotic leakage after colorectal resection: a multicentric, prospective study

Aim The aim was to determine the accuracy of C-reactive protein (CRP), procalcitonin and neutrophils in the early detection (fourth postoperative day) of anastomotic leakage (AL) after colorectal surgery. Methods We conducted a multicentre, prospective study that included a consecutive series of patients who underwent colorectal resection with anastomosis without ostomy (September 2015 to December 2017). CRP, procalcitonin and neutrophil values on the fourth postoperative day after colorectal resection along with the postoperative outcome (60-day AL, morbidity and mortality) were prospectively included in an online, anonymous database. Results The analysis ultimately included 2501 cases. The overall morbidity and mortality was 30.1% and 1.6%, respectively, and the AL rate was 8.6%. The area under the receiver operating characteristic curve values (95% CI) for detecting AL were 0.84 (0.81-0.87), 0.75 (0.72-0.79) and 0.70 (0.66-0.74) for CRP, procalcitonin and neutrophils, respectively. The best cut-off level for CRP was 119 mg/l, resulting in 70% sensitivity, 81% specificity and 97% negative predictive value. After laparoscopic resection, the accuracy for CRP and procalcitonin was increased, compared with open resection. The combination of two or three of these biomarkers did not significantly increase their accuracy. Conclusion On the fourth postoperative day, CRP was the most reliable marker for excluding AL. Its high negative predictive value, especially after laparoscopic resection, allows for safe hospital discharge on the fourth postoperative day. The routine use of procalcitonin or neutrophil counts does not seem to increase the diagnostic accuracy.

Automated summary

The study included 2501 cases, with CRP being the most reliable marker for excluding anastomotic leakage. Especially after laparoscopic resection, its high negative predictive value allows for safe hospital discharge on the fourth postoperative day. However, routine use of procalcitonin or neutrophil counts does not seem to increase diagnostic accuracy.