Early C-reactive protein after colorectal surgery is not predictive of anastomotic leak: a retrospective cohort study.

Abstract

Prior studies suggest postoperative C-reactive protein (CRP) trends are sensitive for predicting anastomotic leak (AL) after elective colorectal surgery. However, in the setting of enhanced recovery pathways, multi-day CRP trends may not be feasible. This study aimed to assess the realistic and clinical utility of CRP in prediction of AL. A retrospective review of patients who underwent elective colectomy or proctectomy from January 2019 to October 2020 at a single institution was performed. Comorbidities, operative characteristics, and perioperative outcomes were recorded. CRP was checked routinely on POD1 and on a clinical basis subsequently. The association between 10-point change in CRP-POD1 and AL was evaluated using multivariable logistic regression. The relationships between CRP-POD3, CRP-POD1, and AL were assessed using exploratory analyses. Among 332 patients, 23 (6.9%) developed AL, of which 9 cases (39%) were diagnosed upon readmission. AL was not associated with mortality. Median length of stay was 3 days (IQR 2-5). Median days to AL diagnosis was 7 (IQR 4-15). Adjusting for diverting stoma, steroid use, diagnosis, and open surgery, each 10-point increase in CRP was associated with increased odds of AL (OR 1.12, 95% CI 1.03-1.21, p=0.008). CRP-POD1 had poor discriminant utility for detecting AL (AUC 0.62, 95% CI = 0.494-0.746; p=0.061). CRP on POD1 is not a reliable method to predict a leak, and trending CRP may not be practical with decreasing lengths of stay in colorectal surgery.