Endoscopic versus laparoscopic drainage of pseudocyst and walled-off necrosis following acute pancreatitis: a randomized trial.

Abstract

Pancreatic fluid collections (PFC) may develop following acute pancreatitis (AP). Endoscopic and laparoscopic internal drainage are accepted modalities for drainage of PFCs but have not been compared in a randomized trial. Our objective was to compare endoscopic and laparoscopic internal drainage of pseudocyst/walled-off necrosis following AP. Patients with symptomatic pseudocysts or walled-off necrosis suitable for laparoscopic and endoscopic transmural internal drainage were randomized to either modality in a randomized controlled trial. Endoscopic drainage comprised of per-oral transluminal cystogastrostomy. Additionally, endoscopic lavage and necrosectomy were done following a step-up approach for infected collections. Surgical laparoscopic cystogastrostomy was done for drainage, lavage, and necrosectomy. Primary outcome was resolution of PFCs by the intended modality and secondary outcome was complications. Sixty patients were randomized, 30 each to laparoscopic and endoscopic drainage. Both groups were comparable for baseline characteristics. The initial success rate was 83.3% in the laparoscopic and 76.6% in the endoscopic group (p = 0.7) after the index intervention. The overall success rate of 93.3% (28/30) and 90% (27/30) in the laparoscopic and endoscopic groups respectively was also similar (p = 1.0). Two patients in the laparoscopic group required endoscopic cystogastrostomy for persistent collections. Similarly, two patients in the endoscopic group required laparoscopic drainage. Postoperative complications were comparable between the groups except for higher post-procedure infection in the endoscopic group (19 vs. 9; p = 0.01) requiring endoscopic re-intervention. Endoscopic and laparoscopic techniques have similar efficacy for internal drainage of suitable pancreatic fluid collections with < 30% debris. The choice of procedure should depend on available expertise and patient preference.