Perforation is an uncommon but potentially serious adverse event of endoscopic retrograde cholangiopancreatography (ERCP). Although its incidence is low, associated morbidity can be significant, particularly when diagnosis or intervention is delayed. Mechanisms of injury vary and are classified using the Stapfer system, which serves as a useful tool in determining appropriate management. Recent advancements in endoscopic closure techniques have broadened treatment options and may reduce the need for surgery. The aim of this study was to evaluate the clinical characteristics, mechanisms of injury, management strategies, and outcomes of ERCP-related perforations treated at a high-volume referral center over a 20-year period.

We retrospectively examined all ERCP procedures performed at Venizeleion General Hospital between March 2005 and June 2025. Perforation was diagnosed by: (i) direct endoscopic visualization of a mucosal defect, (ii) fluoroscopic evidence of contrast extravasation or retroperitoneal air, or (iii) post-procedural CT findings in symptomatic patients such as retroperitoneal air, fluid collections or leakage of orally administered contrast. Perforations were classified according to the Stapfer classification (Types I–IV). Data recorded included demographics, ERCP indication, perforation characteristics, therapeutic approach (conservative, endoscopic, surgical), AGREE (Adverse events in Gastrointestinal Endoscopy) grading, hospitalization duration and mortality. 

Among 6719 ERCPs, 19 perforations were identified (0.28%). Mean age of the patients was 69.4 years, and 58% were female. Choledocholithiasis was the most common indication (78.9%), followed by malignancy (15.8%). Stapfer type II perforations were the most common (63.2%, 12/19), followed by type I (26.3%, 5/19) and type III (10.5%, 2/19). Mechanisms of type II injury included large balloon dilation (6/12), standard sphincterotomy (3/12), transpancreatic sphincterotomy (2/12), and needle-knife precut sphincterotomy (1/12). Type III perforations resulted from guidewire manipulation, while type I injuries occurred in patients with altered or difficult anatomy.

Management strategies were individualized. Among type II cases, 6 patients (50%) were treated conservatively, 4 (33.3%) underwent endoscopic therapy with plastic or fully covered self-expandable metal stents, and 2 (16.7%) required surgery, with one postoperative death. Regarding type I perforations, one patient was treated conservatively, 2 underwent endoscopic closure with an over-the scope clip (OTSC), and 2 required surgery in the pre-OTSC era. One of these patients died postoperatively. Both type III cases were managed conservatively. According to the AGREE classification, perforations were graded as Grade II in 47.4%, Grade III in 42.1%, and Grade V in 10.5% (fatal cases). Overall mortality among all ERCPs was 0.03%, with both deaths occurring after surgical treatment. The mean length of hospitalization was 12.6 days.

Prompt recognition of ERCP-related perforation is essential for optimizing outcomes. With the development of advanced endoscopic closure devices, particularly OTSC, non-operative management has become a feasible and effective first-line strategy in selected cases. Multidisciplinary collaboration remains critical to achieving favorable outcomes.