Pancreaticojejunostomy versus pancreaticogastrostomy reconstruction after pancreaticoduodenectomy for pancreatic or periampullary tumours: a multicentre randomised trial

Background Postoperative pancreatic fistula is the leading cause of death and morbidity after pancreaticoduodenectomy. However, the best reconstruction method to reduce occurrence of fistula is debated. We did a multicentre, randomised superiority trial to compare the outcomes of different reconstructive techniques in patients undergoing pancreaticoduodenectomy for pancreatic or periampullary tumours. Methods Patients aged 18-85 years with confirmed or suspected neoplasms of the pancreas, distal bile duct, ampulla vateri, duodenum, or periampullary tumours were eligible for inclusion. An internet-based platform was used to randomly assign patients to either pancreaticojejunostomy or pancreaticogastrostomy as reconstruction after pancreaticoduodenectomy, using permuted blocks with six patients per block. Within each centre the randomisation was stratified on the pancreatic duct diameter (<= 3 mm vs>3 mm) measured at the time of surgery. The primary endpoint was the occurrence of clinical postoperative pancreatic fistula (grade B or C) as defined by the International Study Group on Pancreatic Fistula. The study was not masked and analyses were done by intention to treat. Patient follow-up was closed 2 months after discharge from the hospital. This study is registered with ClinicalTrials.gov, number NCT00830778. Findings Between June, 2009, and August, 2012, we randomly allocated 167 patients to receive pancreaticojejunostomy and 162 to receive pancreaticogastrostomy. 33 (19.8%) patients in the pancreaticojejunostomy group and 13 (8.0%) in the pancreaticogastrostomy group had clinical postoperative pancreatic fistula (OR 2.86, 95% CI 1.38-6.17; p=0.002). The overall incidence of postoperative complications did not differ significantly between the groups (99 in the pancreaticojejunostomy group vs 100 in the pancreaticogastrostomy group), although more events in the pancreaticojejunostomy group were of grade >= 3a than in the pancreaticogastrostomy group (39 vs 35). Interpretation In patients undergoing pancreaticoduodenectomy for pancreatic head or periampullary tumours, pancreaticogastrostomy is more efficient than pancreaticojejunostomy in reducing the incidence of postoperative pancreatic fistula.