October’s Paper of the Month is a multicentric, prospective study looking at detection one of the most feared complications of colorectal surgery: anastomotic leakage.
Diagnostic accuracy of C-reactive protein, procalcitonin and neutrophils for the early detection of anastomotic leakage after colorectal resection: a multicentric, prospective study. DOI: 10.1111/codi.15845. Colorectal Disease 2021;00:1–8.
What is known about this subject?
Anastomotic leak (AL) is one of the most feared complications of colorectal surgery. The consequences of AL include chronic sepsis, permanent stoma, delays in oncological treatment alongside substantial healthcare costs. Early diagnosis of AL can limit such sequelae and may reduce its associated morbidity and mortality. Therefore, the potential to identify patients with AL, through use of a biomarker, could be of significant value to the treating clinician. Biomarkers of interest include C-reactive protein (CRP), neutrophils and procalcitonin
Design
Observational, prospective, multicentre cohort study
Population
- Consecutive patients ≥18yrs undergoing elective or emergency surgery requiring colonic or rectal resection with a single anastomosis without a stoma.
- Patients receiving intraoperative intra-peritoneal chemotherapy were excluded.
Method and outcomes
- All participants were tested for CRP, procalcitonin and neutrophil levels on the 4th postoperative day.
- AL was defined as occurring up to 60 days postoperatively diagnosed either radiologically or clinically.
- The authors used receiver operating characteristic curves to analyse the reliability of each biomarker as a predictor for AL. They used the area under the curve (AUC) and Youdens statistic (a statistic that captures the performance of a dichotomous diagnostic test) to determine the cut-off of each biomarker at 90% sensitivity.
Primary outcome
Reliability of CRP, procalcitonin and neutrophils to predict AL on the 4th postoperative day
Secondary outcomes
- To determine the cut-off for each biomarker to detect the presence of AL
- To determine whether combining these biomarkers offered additional accuracy in the detection of AL
- To assess whether surgical access (open vs laparoscopic) affected the reliability of these biomarkers in detecting AL
Results
- Data from 2501 patients was analysed across 34 centres.
- The anastomotic leak rate was 8.6% (n=216)
- 60 day morbidity and mortality were 30.1% and 1.6% respectively.
- CRP was found to be the most reliable biomarker with an AUC of 0.84.
- Procalcitonin and neutrophils had an AUC of 0.75 and 0.64 respectively.
- All 3 biomarkers saw their AUC increase when patients undergoing laparoscopic surgery were analysed.
- A CRP cut-off of 119.5mg/L had a sensitivity of 70%, a specificity of 81%, a negative predictive value of 97% and a positive predictive value of 26% for AL
- The minimum cut-off for CRP for a sensitivity of 90% was 66.5mg/L
- Regression modelling showed that combining all 3 biomarkers did not increase diagnostic accuracy.
Implications for colorectal practice?
This paper reports the largest prospective multicentre study examining the accuracy of 3 biomarkers for predicting AL. The data suggest that interpretation of CRP measurement on the 4th postoperative day may be the optimal timing for this. Specifically, a level of 119.5mg/L in this cohort had a high negative predictive value of 97% (the probability that individuals with a CRP <119.5mg/L don’t have AL). The authors argue that this CRP threshold may help guide the safe discharge of patients. Furthermore, additional investigation may be prompted if such levels are exceed, perhaps with the benefit of identifying AL earlier in the clinical course. All blood indices should be interpreted within the clinical context.