Published online Jun 28, 2022. doi: 10.3748/wjg.v28.i24.2758
Peer-review started: January 13, 2022
First decision: March 8, 2022
Revised: March 22, 2022
Accepted: May 14, 2022
Article in press: May 14, 2022
Published online: June 28, 2022
Processing time: 161 Days and 13.8 Hours
Colorectal anastomotic leakage (CAL) is one of the most dreaded complications after colorectal surgery, with an incidence that can be as high as 27%. This event is associated with increased morbidity and mortality; therefore, its early diagnosis is crucial to reduce clinical consequences and costs. Some biomarkers have been suggested as laboratory tools for the diagnosis of CAL.
To assess the usefulness of plasma C-reactive protein (CRP) and calprotectin (CLP) as early predictors of CAL.
A prospective monocentric observational study was conducted including patients who underwent colorectal resection with anastomosis, from March 2017 to August 2019. Patients were divided into three groups: G1 – no complications; G2 – complications not related to CAL; and G3 – CAL. Five biomarkers were measured and analyzed in the first 5 postoperative days (PODs), namely white blood cell (WBC) count, eosinophil cell count (ECC), CRP, CLP, and procalcitonin (PCT). Clinical criteria, such as abdominal pain and clinical condition, were also assessed. The correlation between biomarkers and CAL was evaluated. Receiver operating characteristic (ROC) curve analysis was used to compare the accuracy of these biomarkers as predictors of CAL, and the area under the ROC curve (AUROC), specificity, sensitivity, positive predictive value, and negative predictive value (NPV) during this period were estimated.
In total, 25 of 396 patients developed CAL (6.3%), and the mean time for this diagnosis was 9.0 ± 6.8 d. Some operative characteristics, such as surgical approach, blood loss, intraoperative complications, and duration of the procedure, were notably related to the development of CAL. The length of hospital stay was markedly higher in the group that developed CAL compared with the group with complications other than CAL and the group with no complications (median of 21 d vs 13 d and 7 d respectively; P < 0.001). For abdominal pain, the best predictive performance was on POD4 and POD5, with the largest AUROC of 0.84 on POD4. Worsening of the clinical condition was associated with the diagnosis of CAL, presenting a higher predictive effect on POD5, with an AUROC of 0.9. WBC and ECC showed better predictive effects on POD5 (AUROC = 0.62 and 0.7, respectively). Those markers also presented a high NPV (94%-98%). PCT had the best predictive effect on POD5 (AUROC = 0.61), although it presented low accuracy. However, this biomarker revealed a high NPV on POD3, POD4, and POD5 (96%, 95%, and 96%, respectively). The mean CRP value on POD5 was significantly higher in the group that developed CAL compared with the group without complications (195.5 ± 139.9 mg/L vs 59.5 ± 43.4 mg/L; P < 0.00001). On POD5, CRP had a NPV of 98%. The mean CLP value on POD3 was significantly higher in G3 compared with G1 (5.26 ± 3.58 μg/mL vs 11.52 ± 6.81 μg/mL; P < 0.00005). On POD3, the combination of CLP and CRP values showed a high diagnostic accuracy (AUROC = 0.82), providing a 5.2 d reduction in the time to CAL diagnosis.
CRP and CLP are moderate predictors of CAL. However, the combination of these biomarkers presents an increased diagnostic accuracy, potentially decreasing the time to CAL diagnosis.
Core Tip: Colorectal anastomotic leakage (CAL) remains a serious postoperative complication. It is associated with high morbidity rates, affecting overall costs and patients’ quality of life. Clinical criteria, imaging studies, and biomarkers have been considered to increase diagnostic accuracy. Plasma C-reactive protein, calprotectin, procalcitonin, white blood cell count, and eosinophil cell count have been proposed as predictors of anastomotic leakage. The combination of C-reactive protein and calprotectin after a minimal clinical suspicion of CAL has shown good diagnostic accuracy, allowing clinicians to reduce the time to CAL detection. Regression models can facilitate building a decision model, as the score proposed for the early detection of CAL.