PS02.213: PREDICTING ANASTOMOTIC LEAKAGE AFTER ESOPHAGECTOMY: THE VALUE OF C-REACTIVE PROTEIN. A NATIONWIDE REGISTER STUDY

Abstract

Abstract Background Anastomotic leakage (AL) after oesophageal surgery is a serious complication, leading to increased postoperative mortality. Early detection and treatment is essential. The aim of this study was to investigate if inflammatory serum parameters in the early postoperative course after oesophageal cancer surgery, could predict development of an AL. Methods All patients in Sweden with AL after thoraco-abdominal surgery for oesophageal or gastroesophageal junction cancer from 2006–01-01 to 2017–01-31 were collected from the Swedish National Registry for esophageal and gastric cancer (NREV). Two matched controls for each AL-patient were also selected from the registry. Pre-, per- and postoperative variables were collected from NREV. The following parameters were collected from patient charts: number of days from surgery to AL, diagnostic modality, method of treatment, serum levels of C-reactive protein (CRP), White blood cell count (WBC) and serum-Albumin taken on postoperative day 1 to 7 (POD1–7). Thoracic- and cervical anastomosis were analysed separately. Perioperative data and inflammatory markers were compared between controls and AL-patients. Receiver Operating Characteristic (ROC) curves was constructed to identify optimal cut-off values and a multivariate analysis was performed to identify independent risk factors. Results A total of 155 patients with AL and 307 controls (n = 462) were analysed. In patients with thoracic anastomosis median levels of CRP and WBC were significantly different on POD 2–7 and on POD 4–7 comparing AL-patients to controls. In patients with cervical anastomosis there was a significant difference in WBC levels on POD 1–7. ROC-curve analysis identified CRP on POD3 (cut-off 221) and POD4 (cut-off 203) as the best candidates for predicting AL in patients with a thoracic anastomosis, displaying an area under curve of 0.754 (Sensitivity 59%, Specificity 83%) and 0.731 (Sensitivity: 57%, Specificity 82%). Conclusion In this nation-wide register study we found that a model combining S-CRP levels on POD 3 and POD 4 after a thoraco-abdominal esophagectomy, could identify patients with increased risk of anastomotic leakage several days before it was diagnosed. However, since the sensitivity of this test is relatively low for predicting AL, it must be assessed in a larger clinical context. Disclosure All authors have declared no conflicts of interest.