Redo pancreaticojejunal anastomosis for late-onset complete pancreaticocutaneous fistula after pancreaticojejunostomy

Abstract

Background Pancreaticojejunal (PJ) anastomosis occasionally fails several months after pancreaticoduodenectomy (PD) with Child reconstruction and can ultimately result in a late-onset complete pancreaticocutaneous fistula (Lc-PF). Since the remnant pancreas is an isolated segment, surgical intervention is necessary to create internal drainage for the pancreatic juice; however, surgery at the previous PJ anastomosis site is technically challenging even for experienced surgeons. Here we describe a simple surgical procedure for Lc-PF, termed redo PJ anastomosis, which was developed at our facility. Methods Between January 2008 and December 2020, six consecutive patients with Lc-PF after PD underwent a redo PJ anastomosis, and the short- and long-term clinical outcomes have been evaluated. The abdominal cavity is carefully dissected through a 10-cm midline skin incision, and the PJ anastomosis site is identified using a percutaneous drain through the fistula tract as a guide, along with the main pancreatic duct (MPD) stump on the pancreatic stump. Next, the pancreatic stump is deliberately immobilized from the dorsal plane to prevent injury to the underlying major vessels. After fixing a stent tube to both the MPD and the Roux-limb using two-sided purse-string sutures, the redo PJ anastomosis is completed using single-layer interrupted sutures. Full-thickness pancreatic sutures are deliberately avoided by passing the needle through only two-thirds of the anterior side of the pancreatic stump. Results The redo PJ anastomosis was performed without any intraoperative complications in all cases. The median intraoperative bleeding and operative time were 71 (range 10–137) mL and 123 (range 56–175) min, respectively. Even though a new mild pancreatic fistula developed postoperatively in all cases, it could be conservatively treated within 3 weeks, and no other postoperative complications were recorded. During the median follow-up period of 92 (range 12–112) months, no complications at the redo PJ anastomosis site were observed. Conclusions This research shows that the redo PJ anastomosis for Lc-PF we developed is a safe, feasible, and technically no demanding procedure with acceptable short- and long-term clinical outcomes. This procedure has the potential to become the preferred treatment strategy for Lc-PF after PD.