Serial postoperative C-reactive protein for early prediction of anastomotic leakage after colorectal surgery: a prospective observational study

Background: Anastomotic leakage is a potentially life-threatening complication after colorectal surgery. Because clinical manifestations may be delayed or nonspecific, an accessible biochemical marker may facilitate earlier recognition and investigation. This study evaluated the association between serial postoperative serum C-reactive protein (CRP) levels and anastomotic leakage after colorectal surgery. Methods: This prospective observational study was conducted in the Department of Colorectal Surgery, Shaheed Suhrawardy Medical College and Hospital, Dhaka, from July 2023 to June 2024. Forty patients undergoing colorectal surgery with anastomosis were included. Serum CRP was measured serially during the postoperative period, and patients were monitored clinically for anastomotic leakage, confirmed by clinical, radiological, endoscopic, or operative findings. Continuous variables were compared using the unpaired Student t test, and diagnostic performance was assessed using receiver operating characteristic (ROC) analysis. Results: The mean age was 43.1±12.1 years, and 26 (65.0%) patients were male. Diabetes mellitus and hypertension were present in 10 (25.0%) and 5 (12.5%) patients, respectively. Anastomotic leakage occurred in 3 (7.5%) patients. CRP did not differ significantly between leakage and non-leakage groups on postoperative day (POD) 1 but was significantly higher in the leakage group from POD 2 through POD 8. On POD 2, a CRP cutoff of 255.3 mg/L yielded an area under the ROC curve (AUC) of 0.934, with 100% sensitivity and 92.1% specificity. On POD 3, the AUC was 0.974, with 100% sensitivity and 94.7% specificity. Conclusions: Serial postoperative CRP was associated with anastomotic leakage, with discriminatory performance evident from POD 2. CRP may complement clinical assessment in identifying patients requiring further evaluation. Larger multicentre studies are needed to validate these findings and establish generalizable thresholds.