BACKGROUND: This study was undertaken to assess the value of C-reactive protein (CRP) in predicting infective complications after elective upper gastrointestinal surgery. METHODS: Demographic data, clinical outcomes and serial CRPs preoperatively to post-operative day (POD) 7 were collected for patients undergoing pancreatectomy, hepatectomy and oesophago-gastrectomy between 2005 and 2016. Areas under the curve (AUC) were used to evaluate diagnostic accuracy per day of measurement. RESULTS: Of the 249 patients, 63 (25.3%) developed infective complications and 25 (10%) developed severe infective complications (≥Clavien-Dindo Grade III). Patients with infective complications trended towards higher CRP levels on POD1-POD4 and had significantly higher CRP levels on POD5 (156 versus 114 mg/dL; P = 0.03), POD6 (146 versus 93 mg/dL; P < 0.01) and POD7 (135 versus 84.6 mg/dL; P < 0.01). CRP had the best diagnostic accuracy for severe infective complications on POD6 (AUC: 0.73) and POD7 (AUC: 0.63). A CRP cut-off of 120 mg/dL on POD6 had a negative predictive value of 96.1% and a CRP cut-off of 80 mg/dL on POD7 had a negative predictive value of 94.9%. CONCLUSION: CRP measurements can be used as a negative predictive marker of infective complications in the first post-operative week after major elective upper gastrointestinal surgery.
BACKGROUND: This study was undertaken to assess the value of C-reactive protein (CRP) in predicting infective complications after elective upper gastrointestinal surgery. METHODS: Demographic data, clinical outcomes and serial CRPs preoperatively to post-operative day (POD) 7 were collected for patients undergoing pancreatectomy, hepatectomy and oesophago-gastrectomy between 2005 and 2016. Areas under the curve (AUC) were used to evaluate diagnostic accuracy per day of measurement. RESULTS: Of the 249 patients, 63 (25.3%) developed infective complications and 25 (10%) developed severe infective complications (≥Clavien-Dindo Grade III). Patients with infective complications trended towards higher CRP levels on POD1-POD4 and had significantly higher CRP levels on POD5 (156 versus 114 mg/dL; P = 0.03), POD6 (146 versus 93 mg/dL; P < 0.01) and POD7 (135 versus 84.6 mg/dL; P < 0.01). CRP had the best diagnostic accuracy for severe infective complications on POD6 (AUC: 0.73) and POD7 (AUC: 0.63). A CRP cut-off of 120 mg/dL on POD6 had a negative predictive value of 96.1% and a CRP cut-off of 80 mg/dL on POD7 had a negative predictive value of 94.9%. CONCLUSION:CRP measurements can be used as a negative predictive marker of infective complications in the first post-operative week after major elective upper gastrointestinal surgery.