Algorithm for the management of ERCP-related perforations

Background and Aims Perforation is a rare but serious adverse event of ERCP. There is no consensus to guide the clinician on the management of ERCP-related perforations, with particular controversy surrounding the immediate surgical management of postprocedurally detected duodenal perforation because of overextension of a sphincterotomy. Our aim was to assess patient outcomes using a predetermined algorithm based on managing ERCP-related duodenal perforations according to the mechanism of injury. Methods A retrospective single-center study of all consecutive patients with Stapfer type I and II perforations between 2000 and 2014 were included. Our institutional algorithm since 2000 dictated that Stapfer type I perforations (duodenal wall perforation, endoscope related) should be managed surgically unless prohibited by underlying comorbidities and Stapfer type II perforations (periampullary, sphincterotomy related) managed nonsurgically unless a deterioration in clinical status necessitated surgery. Results Sixty-one patients (mean age, 51 years; 80% women) were analyzed with Stapfer type I perforations diagnosed in 7 (11%) and type II in 54 (89%). A postprocedural diagnosis of perforation was made in 55 patients (90%). Four patients (7%) had Stapfer type II perforations that failed medical management and required surgery. The mean length of stay (LOS) in the entire cohort was 9.6 days with a low mortality rate of 3%. Systemic inflammatory response syndrome was observed in 18 patients (33%) with Stapfer type II perforations and was not associated with the need for surgery. Concurrent post-ERCP pancreatitis was diagnosed in 26 patients (43%) and was associated with an increased LOS. Conclusions Stapfer type II perforations have excellent outcomes when managed medically. We validate an algorithm for the management of ERCP-related perforations and propose that it should function as a guide.