Background : Malnutrition is not uncommon in patients undergoing pancreaticoduodenectomy (PD), leading to suboptimal outcomes following surgery, which can be further deteriorated if postoperative complications (leakage, delayed gastric emptying…) occur. Various methods of enteral feeding (nasojejunal tube, transperitoneal jejunostomy or transgastric jejunostomy) enable early postoperative enteral nutrition which can potentially decrease wound infection, promote wound healing and decrease the length of hospital stay. It is not clear whether which method is superior, especially in resource-limited setting.The goals of this study is to evaluate the safety and outcomes associated with transgastric and transperitoneal feeding jejunostomy tube placement in patients undergoing pancreaticoduodenectomy.

Methods : All patients undergoing PD from Jan 2019 to Nov 2022 at Hue University Hospital and Hue Central Hospital, being classified as high-risk using ISGPS fistula risk score and having jejunostomy tube inserted (transgastric or transperitoneal) were reviewed retrospectively. Enteral feeding was normally started on POD 2 and advanced as tolerated. Jejunostomy tubes were discontinued if adequate oral intake is resumed and removed before discharge.

Results : A total of 31 patients underwent PD with fistula risk score ≥7 having jejunostomy tube inserted during this period. There were 18 (58.1%) men and 13 (41.9%) women with the mean age of 49.1 years. Types of tumor included ampullary (16/31), distal common bile duct (5/31), pancreatic head (10/31). Delivery route of enteral nutrition was: transperitoneal jejunostomy in 21 (67.7%) and transperitoneal gastrojejunostomy in 10 (32.3%). Nasojejunal tube was not used because of nonavailability of instruments. The median time to oral intake was 6.5 days. Blockage was the most common complication (n=6, 19.4%) with 3 cases in each group. Insertion site infection was seen in 10/21 (47.6%) in transperitonal group to 3/13 (23.1%) in transgastric group (p<0.05). There were no displacement, no respiratory complications in both groups. Small bowel obstruction occurred in 2 cases (6.5%) in the percutaneous group. The mean time to resume oral intake were 5 days in most cases (n=25/31) without any differences between both groups. There was no tube related mortality. The cost of jejunostomy using this method was 10 times cheaper than nasojejunal tube.

Conclusions : Transgastric feeding jejunostomy can be an alternative to transperitoneal jejunostomy or nasojejunostomy with resonable cost and minimal complication rates.