Randomized clinical trial of duct-to-mucosa versus invagination pancreaticojejunostomy after pancreatoduodenectomy

Abstract

Abstract Background The postoperative pancreatic fistula (POPF) rate for duct-to-mucosa and invagination anastomosis after pancreatoduodenectomy is still debated. The aim of this RCT was to investigate the POPF rate for duct-to-mucosa versus invagination pancreaticojejunostomy. Methods Patients were stratified by pancreatic texture and diameter of the main pancreatic duct and randomized to the duct-to-mucosa or invagination group. The primary endpoint was the rate of clinically relevant POPF (defined as grade B or C). Secondary endpoints were suture material cost for pancreaticojejunostomy, drain insertion duration and duration of postoperative hospital stay. Results Some 120 patients undergoing pancreatoduodenectomy were included following consent. Clinically relevant POPF developed in six of 59 patients (10 per cent) in the invagination group and in 14 of 61 patients (23 per cent) in the duct-to-mucosa group (P = 0·077). Duration of drain insertion (6 versus 7 days respectively; P = 0·027) and postoperative hospital stay (19 versus 24 days; P = 0·015) were shorter in the invagination group. Subgroup analysis for 61 patients with a soft pancreas revealed a lower rate of clinically relevant POPF in the invagination group (10 per cent versus 42 per cent in the duct-to-mucosa group; P = 0·010). Among 20 patients with a clinically relevant POPF, the six patients in the invagination group had a shorter duration of drain insertion (38·5 days versus 49 days for 14 patients in the duct-to-mucosa group; P = 0·028) and postoperative hospital stay (42 versus 54·5 days respectively; P = 0·028). Conclusion This study did not demonstrate a superiority of invagination over duct-to-mucosa pancreaticojejunostomy in the risk of POPF. However, in high-risk patients with a soft pancreas, invagination may reduce the risk of clinically relevant POPF compared with duct-to-mucosa. Registration number: UMIN000005890 (http://www.umin.ac.jp).