Clinical course and proposed treatment strategy for ERCP-related duodenal perforation: a multicenter analysis

Abstract

Background and study aims: Endoscopic retrograde cholangiopancreatography (ERCP)-related duodenal perforation is rare but can cause high mortality. Our aim was to assess the clinical outcomes of these events. Method: A total of 59 patients who were diagnosed as having ERCP-related duodenal perforation at six institutions between 2000 and 2007 were enrolled in this multicenter retrospective study. We evaluated complications and mortality associated with ERCP-related duodenal perforation according to injury detection time (IDT), peritoneal irritation signs (PIS), systemic inflammation signs (SIS), and treatment modality in these patients. Results: Of the 59 patients, 41 (69.5 %) and 18 (30.5 %) underwent medical and surgical treatment, respectively. Duodenal perforation-related death was observed in five patients, who had received medical therapy (n = 2) and surgical therapy (n = 3). Among medically treated patients, seven patients (17.1 %) underwent endoscopic clipping immediately after the injury; surgery was not required as a salvage therapy and there were no complications or deaths among these patients. The remaining 34 patients received antibiotics combined with therapeutic fasting and intravenous hydration. Duodenal perforation-related complications depended significantly on IDT (P = 0.0001), treatment modality (P = 0.008), PIS (P = 0.003), and SIS (P = 0.010). The duodenal perforation-related mortality was significantly related to IDT (P = 0.008) and PIS (P = 0.001). Conclusions: IDT, PIS, and SIS appear to be important prognostic factors following ERCP-related duodenal perforation. Medical therapy can be suggested as an initial treatment strategy for ERCP-related duodenal perforation, and if possible, endoscopic clipping is strongly recommended. However, surgical treatment should be considered if the perforation is not expected to seal spontaneously, or if the continuing leakage causes PIS or SIS.