A tailored approach to the management of perforations following endoscopic retrograde cholangiopancreatography and sphincterotomy.

Background: The management of endoscopic retrograde cholangiopancreatography (ERCP)-related perforations remains controversial. Τhe aim of the study was to determine the incidence of perforations following ERCP, their characteristics, operative and non-operative management options and clinical outcome.
Methods: A retrospective review of ERCP-related perforations, during a 21-year period, was performed. Each perforation was categorized into types I to IV according to the location, mechanism and radiographic evaluation of the injury. Comparisons were made between patients treated operatively and non-operatively.
Results: Forty-four perforations (0.4%) occurred in 9,880 procedures. They were mainly caused by the passage of the endoscope (type I) in 7 (16%) and sphincterotomy (type II) in 30 (68%) patients. The management was non-operative in 32 (72%) and operative in 12 patients. In multivariate analysis, only the type of perforation (type I: endoscope-related) was found significant for predicting operative treatment. The hospital stay was longer for patients requiring an operation (median, 24 vs 9 days). The overall mortality was 2/44 (4.5%). There was no death in the non-operative group.
Conclusions: The need for immediate operative intervention should be based on the type of injury and clinical findings. Patients with type I perforations should be treated surgically and primary repair should be tried. Patients with type II injuries may be treated initially non-operatively. Delayed operative intervention will be required in a minority of these patients.