This network meta-analysis aimed to identify the reconstruction technique associated with lowest rates of DGE following pancreatoduodenectomy (PD) from randomised controlled trials (RCTs).
A systematic literature search of PubMed, Embase and MEDLINE databases was carried out using the PRISMA framework to identify all RCTs comparing reconstruction techniques of gastrojejunostomy after PD, with overall DGE as the primary endpoint. The primary outcome measure was overall DGE. Secondary outcomes were grade B/C DGE, duration of nasogastric tube, time to solid food intake, overall and grade B/C pancreatic fistula, bile leaks, reoperation, length of hospital stay and in-hospital mortality.
The search strategy identified eight RCTs including 761 patients. Six RCTs compared antecolic (n = 291 patients) and retrocolic Billroth II (n = 289 patients) reconstruction (n = 6 studies), and two RCTs compared antecolic Billroth II (n = 92 patients) and Roux-en-Y (n = 89 patients) reconstruction. Overall, antecolic Billroth II ranked best for overall and grade B/C DGE, bile leak, surgical site infection, length of stay and in-hospital mortality. Roux-en-Y was best for overall and grade B/C pancreatic fistula.
Antecolic Billroth II gastroenteric reconstruction is associated with the lowest rates of delayed gastric emptying after PD amongst the currently available techniques of gastrojejunostomy reconstructions.