Critical appraisal of the techniques of pancreatic anastomosis following pancreaticoduodenectomy: A network meta-analysis

Background: Postoperative pancreatic fistula (POPF) remains a major cause of morbidity following pancreaticoduodenectomy (PD). This network meta-analysis (NMA) compared techniques of pancreatic anastomosis following PD to determine the technique with the best outcome profile. Methods: A systematic literature search was performed on the Scopus, EMBASE, Medline and Cochrane databases to identify RCTs employing the international study group of pancreatic fistula (ISGPF) definition of POPF. The primary outcome was clinically relevant POPF. Results: Five techniques of pancreatic anastomosis following PD were directly compared in 15 RCTs comprising 2428 patients. Panreatojejunostomy (PJ) end-to-side invagination vs. PJ end-to-side duct-to-mucosa was the most frequent comparison (n = 7). Overall, 971 patients underwent PJ end-to-side duct-to-mucosa, 791 patients PJ end-to-side invagination, 505 patients pancreatogastrostomy (PG) end-to-side invagination, 98 patients PG end-to-side duct-to-mucosa, and 63 patients PJ end-to-side single layer. PG duct-to-mucosa was associated with the lowest rates of clinically relevant POPF, delayed gastric emptying, infra-abdominal abscess, all postoperative morbidity and postoperative mortality, the shortest operative time and postoperative hospital stay and the lowest volume of infra-operative blood loss. Conclusion: Duct-to-mucosa pancreaticogastrostomy was associated with the lowest rates of clinically relevant POPF and had the best outcome profile among all techniques of pancreatico-anastomosis following PD.